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/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
27 EC_KEY_precompute_mult() These functions are not widely used and applications
28 should instead switch to named curves which OpenSSL has hardcoded lookup
33 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
34 should instead use the L<EC_POINT_mul(3)> function.
38 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
39 that are not applicable to the new provider model. Applications should
40 instead use EVP_default_properties_is_fips_enabled() and
41 EVP_default_properties_enable_fips().
45 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
46 is set, an unexpected EOF is ignored, it pretends a close notify was received
47 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
51 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
52 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
53 used and applications should instead use the
54 L<EC_POINT_set_affine_coordinates(3)> and
55 L<EC_POINT_get_affine_coordinates(3)> functions.
59 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
60 arrays to be more easily constructed via a series of utility functions.
61 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
62 the various push functions and finally convert to a passable OSSL_PARAM
63 array using OSSL_PARAM_BLD_to_param().
67 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
68 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
69 internal keys, if they correspond to one of those built in types.
73 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
74 contain a provider side internal key.
78 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
79 They are old functions that we don't use, and that you could disable with
80 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
84 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
85 have been converted to Markdown with the goal to produce documents
86 which not only look pretty when viewed online in the browser, but
87 remain well readable inside a plain text editor.
89 To achieve this goal, a 'minimalistic' Markdown style has been applied
90 which avoids formatting elements that interfere too much with the
91 reading flow in the text file. For example, it
93 * avoids [ATX headings][] and uses [setext headings][] instead
94 (which works for `<h1>` and `<h2>` headings only).
95 * avoids [inline links][] and uses [reference links][] instead.
96 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
98 [ATX headings]: https://github.github.com/gfm/#atx-headings
99 [setext headings]: https://github.github.com/gfm/#setext-headings
100 [inline links]: https://github.github.com/gfm/#inline-link
101 [reference links]: https://github.github.com/gfm/#reference-link
102 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
103 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
105 *Matthias St. Pierre*
107 * The test suite is changed to preserve results of each test recipe.
108 A new directory test-runs/ with subdirectories named like the
109 test recipes are created in the build tree for this purpose.
113 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
114 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
115 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
117 *David von Oheimb, Martin Peylo*
119 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
120 The legacy OCSP-focused and only partly documented API is retained.
121 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
125 * All of the low level RSA functions have been deprecated including:
127 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
128 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
129 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
130 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
131 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
132 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
133 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
134 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
135 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
136 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
137 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
138 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
139 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
140 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
141 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
142 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
143 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
144 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
145 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
146 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
147 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
148 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
149 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
150 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
151 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
152 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
153 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
154 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
156 Use of these low level functions has been informally discouraged for a long
157 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
158 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
159 L<EVP_PKEY_decrypt(3)>.
163 * X509 certificates signed using SHA1 are no longer allowed at security
165 In TLS/SSL the default security level is 1. It can be set either
166 using the cipher string with @SECLEVEL, or calling
167 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
168 a call to SSL_CTX_use_certificate() will fail if the security level is not
170 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
171 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
176 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
177 modified to use PKEY APIs. These commands are now in maintenance mode
178 and no new features will be added to them.
182 * The command line utility rsautl has been deprecated.
183 Instead use the pkeyutl program.
187 * The command line utilities genrsa and rsa have been modified to use PKEY
188 APIs These commands are now in maintenance mode and no new features will
193 * All of the low level DH functions have been deprecated including:
195 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
196 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
197 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
198 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
199 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
200 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
201 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
202 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
203 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
204 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
205 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
206 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
207 DH_meth_get_generate_params and DH_meth_set_generate_params.
209 Use of these low level functions has been informally discouraged for a long
210 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
211 and L<EVP_PKEY_derive(3)>.
215 * All of the low level DSA functions have been deprecated including:
217 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
218 DSA_get_default_method, DSA_set_method, DSA_get_method,
219 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
220 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
221 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
222 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
223 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
224 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
225 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
226 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
227 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
228 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
229 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
232 Use of these low level functions has been informally discouraged for a long
233 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
234 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
238 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
239 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
240 This means that applications don't have to look at the curve NID and
241 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
242 However, they still can, that EVP_PKEY_set_alias_type() call acts as
243 a no-op when the EVP_PKEY is already of the given type.
245 Parameter and key generation is also reworked to make it possible
246 to generate EVP_PKEY_SM2 parameters and keys without having to go
247 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
248 However, code that does the latter will still work as before.
252 * Deprecated low level ECDH and ECDSA functions. These include:
254 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
255 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
258 Use of these low level functions has been informally discouraged for a long
259 time. Instead applications should use the EVP_PKEY_derive(3),
260 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
264 * Deprecated the EC_KEY_METHOD functions. These include:
266 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
267 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
268 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
269 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
270 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
271 EC_KEY_METHOD_get_verify.
273 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
277 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
278 and EVP_PKEY_decrypt() instead.
279 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
280 and EVP_PKEY_encrypt() instead.
284 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
285 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
286 a new formulation to include all the things it can be used for,
287 as well as words of caution.
291 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
292 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
296 * All of the low level HMAC functions have been deprecated including:
298 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
299 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
302 Use of these low level functions has been informally discouraged for a long
303 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
304 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
305 and L<EVP_MAC_final(3)>.
309 * Over two thousand fixes were made to the documentation, including:
310 - Common options (such as -rand/-writerand, TLS version control, etc)
311 were refactored and point to newly-enhanced descriptions in openssl.pod.
312 - Added style conformance for all options (with help from Richard Levitte),
313 documented all reported missing options, added a CI build to check
314 that all options are documented and that no unimplemented options
316 - Documented some internals, such as all use of environment variables.
317 - Addressed all internal broken L<> references.
321 * All of the low level CMAC functions have been deprecated including:
323 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
324 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
326 Use of these low level functions has been informally discouraged for a long
327 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
328 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
329 and L<EVP_MAC_final(3)>.
333 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
334 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
337 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
338 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
339 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
340 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
341 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
342 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
343 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
344 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
345 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
346 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
348 Use of these low level functions has been informally discouraged
349 for a long time. Applications should use the EVP_DigestInit_ex(3),
350 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
354 * Corrected the documentation of the return values from the `EVP_DigestSign*`
355 set of functions. The documentation mentioned negative values for some
356 errors, but this was never the case, so the mention of negative values
359 Code that followed the documentation and thereby check with something
360 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
364 * All of the low level cipher functions have been deprecated including:
366 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
367 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
368 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
369 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
370 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
371 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
372 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
373 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
374 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
375 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
376 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
377 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
378 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
379 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
380 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
381 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
382 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
383 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
384 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
385 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
386 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
387 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
388 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
389 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
390 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
391 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
392 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
393 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
394 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
396 Use of these low level functions has been informally discouraged for
397 a long time. Applications should use the high level EVP APIs, e.g.
398 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
399 equivalently named decrypt functions instead.
401 *Matt Caswell and Paul Dale*
403 * Removed include/openssl/opensslconf.h.in and replaced it with
404 include/openssl/configuration.h.in, which differs in not including
405 <openssl/macros.h>. A short header include/openssl/opensslconf.h
406 was added to include both.
408 This allows internal hacks where one might need to modify the set
409 of configured macros, for example this if deprecated symbols are
410 still supposed to be available internally:
412 #include <openssl/configuration.h>
414 #undef OPENSSL_NO_DEPRECATED
415 #define OPENSSL_SUPPRESS_DEPRECATED
417 #include <openssl/macros.h>
419 This should not be used by applications that use the exported
420 symbols, as that will lead to linking errors.
424 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
425 used in exponentiation with 512-bit moduli. No EC algorithms are
426 affected. Analysis suggests that attacks against 2-prime RSA1024,
427 3-prime RSA1536, and DSA1024 as a result of this defect would be very
428 difficult to perform and are not believed likely. Attacks against DH512
429 are considered just feasible. However, for an attack the target would
430 have to re-use the DH512 private key, which is not recommended anyway.
431 Also applications directly using the low level API BN_mod_exp may be
432 affected if they use BN_FLG_CONSTTIME.
437 * Most memory-debug features have been deprecated, and the functionality
438 replaced with no-ops.
442 * Added documentation for the STACK API. OpenSSL only defines the STACK
443 functions where they are used.
447 * Introduced a new method type and API, OSSL_SERIALIZER, to
448 represent generic serializers. An implementation is expected to
449 be able to serialize an object associated with a given name (such
450 as an algorithm name for an asymmetric key) into forms given by
451 implementation properties.
453 Serializers are primarily used from inside libcrypto, through
454 calls to functions like EVP_PKEY_print_private(),
455 PEM_write_bio_PrivateKey() and similar.
457 Serializers are specified in such a way that they can be made to
458 directly handle the provider side portion of an object, if this
459 provider side part comes from the same provider as the serializer
460 itself, but can also be made to handle objects in parametrized
461 form (as an OSSL_PARAM array of data). This allows a provider to
462 offer generic serializers as a service for any other provider.
466 * Added a .pragma directive to the syntax of configuration files, to
467 allow varying behavior in a supported and predictable manner.
468 Currently added pragma:
472 This allows dollar signs to be a keyword character unless it's
473 followed by a opening brace or parenthesis. This is useful for
474 platforms where dollar signs are commonly used in names, such as
475 volume names and system directory names on VMS.
479 * Added functionality to create an EVP_PKEY from user data. This
480 is effectively the same as creating a RSA, DH or DSA object and
481 then assigning them to an EVP_PKEY, but directly using algorithm
482 agnostic EVP functions. A benefit is that this should be future
483 proof for public key algorithms to come.
487 * Change the interpretation of the '--api' configuration option to
488 mean that this is a desired API compatibility level with no
489 further meaning. The previous interpretation, that this would
490 also mean to remove all deprecated symbols up to and including
491 the given version, no requires that 'no-deprecated' is also used
492 in the configuration.
494 When building applications, the desired API compatibility level
495 can be set with the OPENSSL_API_COMPAT macro like before. For
496 API compatibility version below 3.0, the old style numerical
497 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
498 For version 3.0 and on, the value is expected to be the decimal
499 value calculated from the major and minor version like this:
501 MAJOR * 10000 + MINOR * 100
505 -DOPENSSL_API_COMPAT=30000 For 3.0
506 -DOPENSSL_API_COMPAT=30200 For 3.2
508 To hide declarations that are deprecated up to and including the
509 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
510 given when building the application as well.
514 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
515 access to certificate and CRL stores via URIs and OSSL_STORE
518 This adds the following functions:
520 - X509_LOOKUP_store()
521 - X509_STORE_load_file()
522 - X509_STORE_load_path()
523 - X509_STORE_load_store()
524 - SSL_add_store_cert_subjects_to_stack()
525 - SSL_CTX_set_default_verify_store()
526 - SSL_CTX_load_verify_file()
527 - SSL_CTX_load_verify_dir()
528 - SSL_CTX_load_verify_store()
530 Also, the following functions are now deprecated:
532 - X509_STORE_load_locations() (use X509_STORE_load_file(),
533 X509_STORE_load_path() or X509_STORE_load_store() instead)
534 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
535 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
539 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
540 The presence of this system service is determined at run-time.
544 * Added functionality to create an EVP_PKEY context based on data
545 for methods from providers. This takes an algorithm name and a
546 property query string and simply stores them, with the intent
547 that any operation that uses this context will use those strings
548 to fetch the needed methods implicitly, thereby making the port
549 of application written for pre-3.0 OpenSSL easier.
553 * The undocumented function NCONF_WIN32() has been deprecated; for
554 conversion details see the HISTORY section of doc/man5/config.pod
558 * Introduced the new functions EVP_DigestSignInit_ex() and
559 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
560 EVP_DigestVerifyUpdate() have been converted to functions. See the man
561 pages for further details.
565 * Over two thousand fixes were made to the documentation, including:
566 adding missing command flags, better style conformance, documentation
569 *Rich Salz, Richard Levitte*
571 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
572 X25519, X448, Ed25519 and Ed448.
576 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
581 * Deprecated the public definition of ERR_STATE as well as the function
582 ERR_get_state(). This is done in preparation of making ERR_STATE an
587 * Added ERR functionality to give callers access to the stored function
588 names that have replaced the older function code based functions.
590 New functions are ERR_get_error_func(), ERR_peek_error_func(),
591 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
592 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
593 and ERR_peek_last_error_all().
595 These functions have become deprecated: ERR_get_error_line_data(),
596 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
597 ERR_func_error_string().
601 * Extended testing to be verbose for failing tests only. The make variables
602 VERBOSE_FAILURE or VF can be used to enable this:
604 $ make VF=1 test # Unix
605 $ mms /macro=(VF=1) test ! OpenVMS
606 $ nmake VF=1 test # Windows
610 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
611 used even when parsing explicit parameters, when loading a serialized key
612 or calling `EC_GROUP_new_from_ecpkparameters()`/
613 `EC_GROUP_new_from_ecparameters()`.
614 This prevents bypass of security hardening and performance gains,
615 especially for curves with specialized EC_METHODs.
616 By default, if a key encoded with explicit parameters is loaded and later
617 serialized, the output is still encoded with explicit parameters, even if
618 internally a "named" EC_GROUP is used for computation.
622 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
623 this change, EC_GROUP_set_generator would accept order and/or cofactor as
624 NULL. After this change, only the cofactor parameter can be NULL. It also
625 does some minimal sanity checks on the passed order.
630 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
631 An attack is simple, if the first CMS_recipientInfo is valid but the
632 second CMS_recipientInfo is chosen ciphertext. If the second
633 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
634 encryption key will be replaced by garbage, and the message cannot be
635 decoded, but if the RSA decryption fails, the correct encryption key is
636 used and the recipient will not notice the attack.
637 As a work around for this potential attack the length of the decrypted
638 key must be equal to the cipher default key length, in case the
639 certifiate is not given and all recipientInfo are tried out.
640 The old behaviour can be re-enabled in the CMS code by setting the
641 CMS_DEBUG_DECRYPT flag.
645 * Early start up entropy quality from the DEVRANDOM seed source has been
646 improved for older Linux systems. The RAND subsystem will wait for
647 /dev/random to be producing output before seeding from /dev/urandom.
648 The seeded state is stored for future library initialisations using
649 a system global shared memory segment. The shared memory identifier
650 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
651 the desired value. The default identifier is 114.
655 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
656 when primes for RSA keys are computed.
657 Since we previously always generated primes == 2 (mod 3) for RSA keys,
658 the 2-prime and 3-prime RSA modules were easy to distinguish, since
659 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
660 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
661 This avoids possible fingerprinting of newly generated RSA modules.
665 * Correct the extended master secret constant on EBCDIC systems. Without this
666 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
667 negotiate EMS will fail. Unfortunately this also means that TLS connections
668 between EBCDIC systems with this fix, and EBCDIC systems without this
669 fix will fail if they negotiate EMS.
673 * Changed the library initialisation so that the config file is now loaded
674 by default. This was already the case for libssl. It now occurs for both
675 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
676 OPENSSL_init_crypto() to suppress automatic loading of a config file.
680 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
681 where the former acts as a replacement for ERR_put_error(), and the
682 latter replaces the combination ERR_put_error()+ERR_add_error_data().
683 ERR_raise_data() adds more flexibility by taking a format string and
684 an arbitrary number of arguments following it, to be processed with
689 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
690 to check if a named provider is loaded and available. When called, it
691 will also activate all fallback providers if such are still present.
695 * Enforce a minimum DH modulus size of 512 bits.
699 * Changed DH parameters to generate the order q subgroup instead of 2q.
700 Previously generated DH parameters are still accepted by DH_check
701 but DH_generate_key works around that by clearing bit 0 of the
702 private key for those. This avoids leaking bit 0 of the private key.
706 * Significantly reduce secure memory usage by the randomness pools.
710 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
715 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
716 algorithms. An implementation of a key exchange algorithm can be obtained
717 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
718 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
719 the older EVP_PKEY_derive_init() function. See the man pages for the new
720 functions for further details.
724 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
728 * Removed the function names from error messages and deprecated the
731 * Removed NextStep support and the macro OPENSSL_UNISTD
735 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
736 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
737 Also removed "export var as function" capability; we do not export
738 variables, only functions.
742 * RC5_32_set_key has been changed to return an int type, with 0 indicating
743 an error and 1 indicating success. In previous versions of OpenSSL this
744 was a void type. If a key was set longer than the maximum possible this
749 * Support SM2 signing and verification schemes with X509 certificate.
753 * Use SHA256 as the default digest for TS query in the ts app.
757 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
758 This checks that the salt length is at least 128 bits, the derived key
759 length is at least 112 bits, and that the iteration count is at least 1000.
760 For backwards compatibility these checks are disabled by default in the
761 default provider, but are enabled by default in the fips provider.
762 To enable or disable these checks use the control
763 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
767 * Default cipher lists/suites are now available via a function, the
768 #defines are deprecated.
772 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
773 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
774 for Windows Store apps easier. Also, the "no-uplink" option has been added.
778 * Join the directories crypto/x509 and crypto/x509v3
782 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
783 This changes the size when using the genpkey app when no size is given. It
784 fixes an omission in earlier changes that changed all RSA, DSA and DH
785 generation apps to use 2048 bits by default.
789 * Added command 'openssl kdf' that uses the EVP_KDF API.
793 * Added command 'openssl mac' that uses the EVP_MAC API.
797 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
798 as default directories. Also added the command 'openssl info'
799 for scripting purposes.
803 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
804 deprecated. These undocumented functions were never integrated into the EVP
805 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
806 Bi-directional IGE mode. These modes were never formally standardised and
807 usage of these functions is believed to be very small. In particular
808 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
809 is ever used. The security implications are believed to be minimal, but
810 this issue was never fixed for backwards compatibility reasons. New code
811 should not use these modes.
815 * Add prediction resistance to the DRBG reseeding process.
819 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
820 mandated by IEEE Std 1619-2018.
824 * Added newline escaping functionality to a filename when using openssl dgst.
825 This output format is to replicate the output format found in the `*sum`
826 checksum programs. This aims to preserve backward compatibility.
828 *Matt Eaton, Richard Levitte, and Paul Dale*
830 * Removed the heartbeat message in DTLS feature, as it has very
831 little usage and doesn't seem to fulfill a valuable purpose.
832 The configuration option is now deprecated.
836 * Changed the output of 'openssl {digestname} < file' to display the
837 digest name in its output.
841 * Added a new generic trace API which provides support for enabling
842 instrumentation through trace output. This feature is mainly intended
843 as an aid for developers and is disabled by default. To utilize it,
844 OpenSSL needs to be configured with the `enable-trace` option.
846 If the tracing API is enabled, the application can activate trace output
847 by registering BIOs as trace channels for a number of tracing and debugging
850 The 'openssl' application has been expanded to enable any of the types
851 available via environment variables defined by the user, and serves as
852 one possible example on how to use this functionality.
854 *Richard Levitte & Matthias St. Pierre*
856 * Added build tests for C++. These are generated files that only do one
857 thing, to include one public OpenSSL head file each. This tests that
858 the public header files can be usefully included in a C++ application.
860 This test isn't enabled by default. It can be enabled with the option
861 'enable-buildtest-c++'.
865 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
869 * Add KMAC to EVP_MAC.
873 * Added property based algorithm implementation selection framework to
878 * Added SCA hardening for modular field inversion in EC_GROUP through
879 a new dedicated field_inv() pointer in EC_METHOD.
880 This also addresses a leakage affecting conversions from projective
881 to affine coordinates.
883 *Billy Bob Brumley, Nicola Tuveri*
885 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
886 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
887 those algorithms that were already supported through the EVP_PKEY API
888 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
889 and scrypt are now wrappers that call EVP_KDF.
893 * Build devcrypto engine as a dynamic engine.
897 * Add keyed BLAKE2 to EVP_MAC.
901 * Fix a bug in the computation of the endpoint-pair shared secret used
902 by DTLS over SCTP. This breaks interoperability with older versions
903 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
904 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
905 interoperability with such broken implementations. However, enabling
906 this switch breaks interoperability with correct implementations.
908 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
909 re-used X509_PUBKEY object if the second PUBKEY is malformed.
913 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
917 * Change the license to the Apache License v2.0.
921 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
923 - Major releases (indicated by incrementing the MAJOR release number)
924 may introduce incompatible API/ABI changes.
925 - Minor releases (indicated by incrementing the MINOR release number)
926 may introduce new features but retain API/ABI compatibility.
927 - Patch releases (indicated by incrementing the PATCH number)
928 are intended for bug fixes and other improvements of existing
929 features only (like improving performance or adding documentation)
930 and retain API/ABI compatibility.
934 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
938 * Remove the 'dist' target and add a tarball building script. The
939 'dist' target has fallen out of use, and it shouldn't be
940 necessary to configure just to create a source distribution.
944 * Recreate the OS390-Unix config target. It no longer relies on a
945 special script like it did for OpenSSL pre-1.1.0.
949 * Instead of having the source directories listed in Configure, add
950 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
955 * Add GMAC to EVP_MAC.
959 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
963 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
964 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
965 to facilitate the continued use of MACs through raw private keys in
966 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
970 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
971 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
975 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
976 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
977 are retained for backwards compatibility.
981 * AES-XTS mode now enforces that its two keys are different to mitigate
982 the attacked described in "Efficient Instantiations of Tweakable
983 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
984 Details of this attack can be obtained from:
985 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
989 * Rename the object files, i.e. give them other names than in previous
990 versions. Their names now include the name of the final product, as
991 well as its type mnemonic (bin, lib, shlib).
995 * Added new option for 'openssl list', '-objects', which will display the
996 list of built in objects, i.e. OIDs with names.
1000 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1001 improves application performance by removing data copies and providing
1002 applications with zero-copy system calls such as sendfile and splice.
1009 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1011 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1013 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1014 while reading in libssl then we would report an error back to the
1015 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1016 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1017 therefore give a hint as to what went wrong.
1021 * Check that ed25519 and ed448 are allowed by the security level. Previously
1022 signature algorithms not using an MD were not being checked that they were
1023 allowed by the security level.
1027 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1028 was not quite right. The behaviour was not consistent between resumption
1029 and normal handshakes, and also not quite consistent with historical
1030 behaviour. The behaviour in various scenarios has been clarified and
1031 it has been updated to make it match historical behaviour as closely as
1036 * *[VMS only]* The header files that the VMS compilers include automatically,
1037 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1038 that the C++ compiler doesn't understand. This is a shortcoming in the
1039 compiler, but can be worked around with `__cplusplus` guards.
1041 C++ applications that use OpenSSL libraries must be compiled using the
1042 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1043 functions. Otherwise, only functions with symbols of less than 31
1044 characters can be used, as the linker will not be able to successfully
1045 resolve symbols with longer names.
1049 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1050 The presence of this system service is determined at run-time.
1054 * Added newline escaping functionality to a filename when using openssl dgst.
1055 This output format is to replicate the output format found in the `*sum`
1056 checksum programs. This aims to preserve backward compatibility.
1058 *Matt Eaton, Richard Levitte, and Paul Dale*
1060 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1065 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1067 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1068 number generator (RNG). This was intended to include protection in the
1069 event of a fork() system call in order to ensure that the parent and child
1070 processes did not share the same RNG state. However this protection was not
1071 being used in the default case.
1073 A partial mitigation for this issue is that the output from a high
1074 precision timer is mixed into the RNG state so the likelihood of a parent
1075 and child process sharing state is significantly reduced.
1077 If an application already calls OPENSSL_init_crypto() explicitly using
1078 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1081 *Matthias St. Pierre*
1083 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1084 used even when parsing explicit parameters, when loading a serialized key
1085 or calling `EC_GROUP_new_from_ecpkparameters()`/
1086 `EC_GROUP_new_from_ecparameters()`.
1087 This prevents bypass of security hardening and performance gains,
1088 especially for curves with specialized EC_METHODs.
1089 By default, if a key encoded with explicit parameters is loaded and later
1090 serialized, the output is still encoded with explicit parameters, even if
1091 internally a "named" EC_GROUP is used for computation.
1095 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1096 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1097 NULL. After this change, only the cofactor parameter can be NULL. It also
1098 does some minimal sanity checks on the passed order.
1103 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1104 An attack is simple, if the first CMS_recipientInfo is valid but the
1105 second CMS_recipientInfo is chosen ciphertext. If the second
1106 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1107 encryption key will be replaced by garbage, and the message cannot be
1108 decoded, but if the RSA decryption fails, the correct encryption key is
1109 used and the recipient will not notice the attack.
1110 As a work around for this potential attack the length of the decrypted
1111 key must be equal to the cipher default key length, in case the
1112 certifiate is not given and all recipientInfo are tried out.
1113 The old behaviour can be re-enabled in the CMS code by setting the
1114 CMS_DEBUG_DECRYPT flag.
1119 * Early start up entropy quality from the DEVRANDOM seed source has been
1120 improved for older Linux systems. The RAND subsystem will wait for
1121 /dev/random to be producing output before seeding from /dev/urandom.
1122 The seeded state is stored for future library initialisations using
1123 a system global shared memory segment. The shared memory identifier
1124 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1125 the desired value. The default identifier is 114.
1129 * Correct the extended master secret constant on EBCDIC systems. Without this
1130 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1131 negotiate EMS will fail. Unfortunately this also means that TLS connections
1132 between EBCDIC systems with this fix, and EBCDIC systems without this
1133 fix will fail if they negotiate EMS.
1137 * Use Windows installation paths in the mingw builds
1139 Mingw isn't a POSIX environment per se, which means that Windows
1140 paths should be used for installation.
1145 * Changed DH_check to accept parameters with order q and 2q subgroups.
1146 With order 2q subgroups the bit 0 of the private key is not secret
1147 but DH_generate_key works around that by clearing bit 0 of the
1148 private key for those. This avoids leaking bit 0 of the private key.
1152 * Significantly reduce secure memory usage by the randomness pools.
1156 * Revert the DEVRANDOM_WAIT feature for Linux systems
1158 The DEVRANDOM_WAIT feature added a select() call to wait for the
1159 /dev/random device to become readable before reading from the
1160 /dev/urandom device.
1162 It turned out that this change had negative side effects on
1163 performance which were not acceptable. After some discussion it
1164 was decided to revert this feature and leave it up to the OS
1165 resp. the platform maintainer to ensure a proper initialization
1166 during early boot time.
1168 *Matthias St. Pierre*
1170 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1172 * Add build tests for C++. These are generated files that only do one
1173 thing, to include one public OpenSSL head file each. This tests that
1174 the public header files can be usefully included in a C++ application.
1176 This test isn't enabled by default. It can be enabled with the option
1177 'enable-buildtest-c++'.
1181 * Enable SHA3 pre-hashing for ECDSA and DSA.
1185 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1186 This changes the size when using the genpkey app when no size is given. It
1187 fixes an omission in earlier changes that changed all RSA, DSA and DH
1188 generation apps to use 2048 bits by default.
1192 * Reorganize the manual pages to consistently have RETURN VALUES,
1193 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1194 util/fix-doc-nits accordingly.
1196 *Paul Yang, Joshua Lock*
1198 * Add the missing accessor EVP_PKEY_get0_engine()
1202 * Have apps like 's_client' and 's_server' output the signature scheme
1203 along with other cipher suite parameters when debugging.
1207 * Make OPENSSL_config() error agnostic again.
1211 * Do the error handling in RSA decryption constant time.
1215 * Prevent over long nonces in ChaCha20-Poly1305.
1217 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1218 for every encryption operation. RFC 7539 specifies that the nonce value
1219 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1220 and front pads the nonce with 0 bytes if it is less than 12
1221 bytes. However it also incorrectly allows a nonce to be set of up to 16
1222 bytes. In this case only the last 12 bytes are significant and any
1223 additional leading bytes are ignored.
1225 It is a requirement of using this cipher that nonce values are
1226 unique. Messages encrypted using a reused nonce value are susceptible to
1227 serious confidentiality and integrity attacks. If an application changes
1228 the default nonce length to be longer than 12 bytes and then makes a
1229 change to the leading bytes of the nonce expecting the new value to be a
1230 new unique nonce then such an application could inadvertently encrypt
1231 messages with a reused nonce.
1233 Additionally the ignored bytes in a long nonce are not covered by the
1234 integrity guarantee of this cipher. Any application that relies on the
1235 integrity of these ignored leading bytes of a long nonce may be further
1236 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1237 is safe because no such use sets such a long nonce value. However user
1238 applications that use this cipher directly and set a non-default nonce
1239 length to be longer than 12 bytes may be vulnerable.
1241 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1247 * Add DEVRANDOM_WAIT feature for Linux systems
1249 On older Linux systems where the getrandom() system call is not available,
1250 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1251 Contrary to getrandom(), the /dev/urandom device will not block during
1252 early boot when the kernel CSPRNG has not been seeded yet.
1254 To mitigate this known weakness, use select() to wait for /dev/random to
1255 become readable before reading from /dev/urandom.
1257 * Ensure that SM2 only uses SM3 as digest algorithm
1261 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1263 * Change the info callback signals for the start and end of a post-handshake
1264 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1265 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1266 confused by this and assume that a TLSv1.2 renegotiation has started. This
1267 can break KeyUpdate handling. Instead we no longer signal the start and end
1268 of a post handshake message exchange (although the messages themselves are
1269 still signalled). This could break some applications that were expecting
1270 the old signals. However without this KeyUpdate is not usable for many
1275 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1277 * Timing vulnerability in DSA signature generation
1279 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1280 timing side channel attack. An attacker could use variations in the signing
1281 algorithm to recover the private key.
1283 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1288 * Timing vulnerability in ECDSA signature generation
1290 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1291 timing side channel attack. An attacker could use variations in the signing
1292 algorithm to recover the private key.
1294 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1299 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1300 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1301 of two gigabytes and the error handling improved.
1303 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1304 categorized as a normal bug, not a security issue, because the DRBG reseeds
1305 automatically and is fully functional even without additional randomness
1306 provided by the application.
1308 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1310 * Add a new ClientHello callback. Provides a callback interface that gives
1311 the application the ability to adjust the nascent SSL object at the
1312 earliest stage of ClientHello processing, immediately after extensions have
1313 been collected but before they have been processed. In particular, this
1314 callback can adjust the supported TLS versions in response to the contents
1319 * Add SM2 base algorithm support.
1323 * s390x assembly pack: add (improved) hardware-support for the following
1324 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1325 aes-cfb/cfb8, aes-ecb.
1329 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1330 parameter is no longer accepted, as it leads to a corrupt table. NULL
1331 pem_str is reserved for alias entries only.
1335 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1336 step for prime curves. The new implementation is based on formulae from
1337 differential addition-and-doubling in homogeneous projective coordinates
1338 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1339 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1340 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1341 to work in projective coordinates.
1343 *Billy Bob Brumley, Nicola Tuveri*
1345 * Change generating and checking of primes so that the error rate of not
1346 being prime depends on the intended use based on the size of the input.
1347 For larger primes this will result in more rounds of Miller-Rabin.
1348 The maximal error rate for primes with more than 1080 bits is lowered
1351 *Kurt Roeckx, Annie Yousar*
1353 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1357 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1358 moving between systems, and to avoid confusion when a Windows build is
1359 done with mingw vs with MSVC. For POSIX installs, there's still a
1360 symlink or copy named 'tsget' to avoid that confusion as well.
1364 * Revert blinding in ECDSA sign and instead make problematic addition
1365 length-invariant. Switch even to fixed-length Montgomery multiplication.
1369 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1370 step for binary curves. The new implementation is based on formulae from
1371 differential addition-and-doubling in mixed Lopez-Dahab projective
1372 coordinates, modified to independently blind the operands.
1374 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1376 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1377 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1378 EC_METHODs to implement their own specialized "ladder step", to take
1379 advantage of more favorable coordinate systems or more efficient
1380 differential addition-and-doubling algorithms.
1382 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1384 * Modified the random device based seed sources to keep the relevant
1385 file descriptors open rather than reopening them on each access.
1386 This allows such sources to operate in a chroot() jail without
1387 the associated device nodes being available. This behaviour can be
1388 controlled using RAND_keep_random_devices_open().
1392 * Numerous side-channel attack mitigations have been applied. This may have
1393 performance impacts for some algorithms for the benefit of improved
1394 security. Specific changes are noted in this change log by their respective
1399 * AIX shared library support overhaul. Switch to AIX "natural" way of
1400 handling shared libraries, which means collecting shared objects of
1401 different versions and bitnesses in one common archive. This allows to
1402 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1403 doesn't affect the way 3rd party applications are linked, only how
1404 multi-version installation is managed.
1408 * Make ec_group_do_inverse_ord() more robust and available to other
1409 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1410 mitigations are applied to the fallback BN_mod_inverse().
1411 When using this function rather than BN_mod_inverse() directly, new
1412 EC cryptosystem implementations are then safer-by-default.
1416 * Add coordinate blinding for EC_POINT and implement projective
1417 coordinate blinding for generic prime curves as a countermeasure to
1418 chosen point SCA attacks.
1420 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1422 * Add blinding to ECDSA and DSA signatures to protect against side channel
1423 attacks discovered by Keegan Ryan (NCC Group).
1427 * Enforce checking in the pkeyutl command line app to ensure that the input
1428 length does not exceed the maximum supported digest length when performing
1429 a sign, verify or verifyrecover operation.
1433 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1434 I/O in combination with something like select() or poll() will hang. This
1435 can be turned off again using SSL_CTX_clear_mode().
1436 Many applications do not properly handle non-application data records, and
1437 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1438 around the problems in those applications, but can also break some.
1439 It's recommended to read the manpages about SSL_read(), SSL_write(),
1440 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1441 SSL_CTX_set_read_ahead() again.
1445 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1446 now allow empty (zero character) pass phrases.
1450 * Apply blinding to binary field modular inversion and remove patent
1451 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1455 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1456 binary and prime elliptic curves.
1460 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1461 constant time fixed point multiplication.
1465 * Revise elliptic curve scalar multiplication with timing attack
1466 defenses: ec_wNAF_mul redirects to a constant time implementation
1467 when computing fixed point and variable point multiplication (which
1468 in OpenSSL are mostly used with secret scalars in keygen, sign,
1469 ECDH derive operations).
1470 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1473 * Updated CONTRIBUTING
1477 * Updated DRBG / RAND to request nonce and additional low entropy
1478 randomness from the system.
1480 *Matthias St. Pierre*
1482 * Updated 'openssl rehash' to use OpenSSL consistent default.
1486 * Moved the load of the ssl_conf module to libcrypto, which helps
1487 loading engines that libssl uses before libssl is initialised.
1491 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1495 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1497 *Ingo Schwarze, Rich Salz*
1499 * Added output of accepting IP address and port for 'openssl s_server'
1503 * Added a new API for TLSv1.3 ciphersuites:
1504 SSL_CTX_set_ciphersuites()
1505 SSL_set_ciphersuites()
1509 * Memory allocation failures consistently add an error to the error
1514 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1515 in libcrypto when run as setuid/setgid.
1519 * Load any config file by default when libssl is used.
1523 * Added new public header file <openssl/rand_drbg.h> and documentation
1524 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1526 *Matthias St. Pierre*
1528 * QNX support removed (cannot find contributors to get their approval
1529 for the license change).
1533 * TLSv1.3 replay protection for early data has been implemented. See the
1534 SSL_read_early_data() man page for further details.
1538 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1539 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1540 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1541 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1542 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1543 configuration has been separated out. See the ciphers man page or the
1544 SSL_CTX_set_ciphersuites() man page for more information.
1548 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1549 in responder mode now supports the new "-multi" option, which
1550 spawns the specified number of child processes to handle OCSP
1551 requests. The "-timeout" option now also limits the OCSP
1552 responder's patience to wait to receive the full client request
1553 on a newly accepted connection. Child processes are respawned
1554 as needed, and the CA index file is automatically reloaded
1555 when changed. This makes it possible to run the "ocsp" responder
1556 as a long-running service, making the OpenSSL CA somewhat more
1557 feature-complete. In this mode, most diagnostic messages logged
1558 after entering the event loop are logged via syslog(3) rather than
1563 * Added support for X448 and Ed448. Heavily based on original work by
1568 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1569 objects loaded. This adds the functions OSSL_STORE_expect() and
1570 OSSL_STORE_find() as well as needed tools to construct searches and
1571 get the search data out of them.
1575 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1576 version of OpenSSL should review their configuration settings to ensure
1577 that they are still appropriate for TLSv1.3. For further information see:
1578 <https://wiki.openssl.org/index.php/TLS1.3>
1582 * Grand redesign of the OpenSSL random generator
1584 The default RAND method now utilizes an AES-CTR DRBG according to
1585 NIST standard SP 800-90Ar1. The new random generator is essentially
1586 a port of the default random generator from the OpenSSL FIPS 2.0
1587 object module. It is a hybrid deterministic random bit generator
1588 using an AES-CTR bit stream and which seeds and reseeds itself
1589 automatically using trusted system entropy sources.
1591 Some of its new features are:
1592 - Support for multiple DRBG instances with seed chaining.
1593 - The default RAND method makes use of a DRBG.
1594 - There is a public and private DRBG instance.
1595 - The DRBG instances are fork-safe.
1596 - Keep all global DRBG instances on the secure heap if it is enabled.
1597 - The public and private DRBG instance are per thread for lock free
1600 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1602 * Changed Configure so it only says what it does and doesn't dump
1603 so much data. Instead, ./configdata.pm should be used as a script
1604 to display all sorts of configuration data.
1608 * Added processing of "make variables" to Configure.
1612 * Added SHA512/224 and SHA512/256 algorithm support.
1616 * The last traces of Netware support, first removed in 1.1.0, have
1621 * Get rid of Makefile.shared, and in the process, make the processing
1622 of certain files (rc.obj, or the .def/.map/.opt files produced from
1623 the ordinal files) more visible and hopefully easier to trace and
1624 debug (or make silent).
1628 * Make it possible to have environment variable assignments as
1629 arguments to config / Configure.
1633 * Add multi-prime RSA (RFC 8017) support.
1637 * Add SM3 implemented according to GB/T 32905-2016
1638 * Jack Lloyd <jack.lloyd@ribose.com>,
1639 Ronald Tse <ronald.tse@ribose.com>,
1640 Erick Borsboom <erick.borsboom@ribose.com> *
1642 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1643 as documented in RFC6066.
1644 Based on a patch from Tomasz Moń
1646 *Filipe Raimundo da Silva*
1648 * Add SM4 implemented according to GB/T 32907-2016.
1649 * Jack Lloyd <jack.lloyd@ribose.com>,
1650 Ronald Tse <ronald.tse@ribose.com>,
1651 Erick Borsboom <erick.borsboom@ribose.com> *
1653 * Reimplement -newreq-nodes and ERR_error_string_n; the
1654 original author does not agree with the license change.
1658 * Add ARIA AEAD TLS support.
1662 * Some macro definitions to support VS6 have been removed. Visual
1663 Studio 6 has not worked since 1.1.0
1667 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1668 without clearing the errors.
1672 * Add "atfork" functions. If building on a system that without
1673 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1674 requirements. The RAND facility now uses/requires this.
1682 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1683 not possible to disable entirely. However, it's still possible to
1684 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1687 To disable, configure with 'no-ui-console'. 'no-ui' is still
1688 possible to use as an alias. Check at compile time with the
1689 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1690 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1694 * Add a STORE module, which implements a uniform and URI based reader of
1695 stores that can contain keys, certificates, CRLs and numerous other
1696 objects. The main API is loosely based on a few stdio functions,
1697 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1698 OSSL_STORE_error and OSSL_STORE_close.
1699 The implementation uses backends called "loaders" to implement arbitrary
1700 URI schemes. There is one built in "loader" for the 'file' scheme.
1704 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1705 then adjusted to work on FreeBSD 8.4 as well.
1706 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1707 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1711 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1712 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1713 error code calls like this:
1715 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1717 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1718 that can be encoded in C. For the foreseeable future, this will only
1721 *Richard Levitte and Tim Hudson*
1723 * Removed BSD cryptodev engine.
1727 * Add a build target 'build_all_generated', to build all generated files
1728 and only that. This can be used to prepare everything that requires
1729 things like perl for a system that lacks perl and then move everything
1730 to that system and do the rest of the build there.
1734 * In the UI interface, make it possible to duplicate the user data. This
1735 can be used by engines that need to retain the data for a longer time
1736 than just the call where this user data is passed.
1740 * Ignore the '-named_curve auto' value for compatibility of applications
1743 *Tomas Mraz <tmraz@fedoraproject.org>*
1745 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1746 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1747 alerts across multiple records (some of which could be empty). In practice
1748 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1749 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1750 support this at all. Supporting it adds significant complexity to the
1751 record layer, and its removal is unlikely to cause interoperability
1756 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1757 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1758 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1763 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1764 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1766 *Richard Levitte, Andy Polyakov*
1768 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1773 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1774 platform rather than 'mingw'.
1778 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1779 success if they are asked to add an object which already exists
1780 in the store. This change cascades to other functions which load
1781 certificates and CRLs.
1785 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1786 facilitate stack unwinding even from assembly subroutines.
1790 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1791 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1795 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1796 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1797 which is the minimum version we support.
1801 * Certificate time validation (X509_cmp_time) enforces stricter
1802 compliance with RFC 5280. Fractional seconds and timezone offsets
1803 are no longer allowed.
1807 * Add support for ARIA
1811 * s_client will now send the Server Name Indication (SNI) extension by
1812 default unless the new "-noservername" option is used. The server name is
1813 based on the host provided to the "-connect" option unless overridden by
1814 using "-servername".
1818 * Add support for SipHash
1822 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1823 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1824 prevent issues where no progress is being made and the peer continually
1825 sends unrecognised record types, using up resources processing them.
1829 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1830 using the algorithm defined in
1831 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1835 * Heartbeat support has been removed; the ABI is changed for now.
1837 *Richard Levitte, Rich Salz*
1839 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1843 * The RSA "null" method, which was partially supported to avoid patent
1844 issues, has been replaced to always returns NULL.
1851 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1853 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1854 used even when parsing explicit parameters, when loading a serialized key
1855 or calling `EC_GROUP_new_from_ecpkparameters()`/
1856 `EC_GROUP_new_from_ecparameters()`.
1857 This prevents bypass of security hardening and performance gains,
1858 especially for curves with specialized EC_METHODs.
1859 By default, if a key encoded with explicit parameters is loaded and later
1860 serialized, the output is still encoded with explicit parameters, even if
1861 internally a "named" EC_GROUP is used for computation.
1865 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1866 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1867 NULL. After this change, only the cofactor parameter can be NULL. It also
1868 does some minimal sanity checks on the passed order.
1873 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1874 An attack is simple, if the first CMS_recipientInfo is valid but the
1875 second CMS_recipientInfo is chosen ciphertext. If the second
1876 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1877 encryption key will be replaced by garbage, and the message cannot be
1878 decoded, but if the RSA decryption fails, the correct encryption key is
1879 used and the recipient will not notice the attack.
1880 As a work around for this potential attack the length of the decrypted
1881 key must be equal to the cipher default key length, in case the
1882 certifiate is not given and all recipientInfo are tried out.
1883 The old behaviour can be re-enabled in the CMS code by setting the
1884 CMS_DEBUG_DECRYPT flag.
1889 * Use Windows installation paths in the mingw builds
1891 Mingw isn't a POSIX environment per se, which means that Windows
1892 paths should be used for installation.
1897 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
1899 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1900 This changes the size when using the genpkey app when no size is given. It
1901 fixes an omission in earlier changes that changed all RSA, DSA and DH
1902 generation apps to use 2048 bits by default.
1906 * Prevent over long nonces in ChaCha20-Poly1305.
1908 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1909 for every encryption operation. RFC 7539 specifies that the nonce value
1910 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1911 and front pads the nonce with 0 bytes if it is less than 12
1912 bytes. However it also incorrectly allows a nonce to be set of up to 16
1913 bytes. In this case only the last 12 bytes are significant and any
1914 additional leading bytes are ignored.
1916 It is a requirement of using this cipher that nonce values are
1917 unique. Messages encrypted using a reused nonce value are susceptible to
1918 serious confidentiality and integrity attacks. If an application changes
1919 the default nonce length to be longer than 12 bytes and then makes a
1920 change to the leading bytes of the nonce expecting the new value to be a
1921 new unique nonce then such an application could inadvertently encrypt
1922 messages with a reused nonce.
1924 Additionally the ignored bytes in a long nonce are not covered by the
1925 integrity guarantee of this cipher. Any application that relies on the
1926 integrity of these ignored leading bytes of a long nonce may be further
1927 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1928 is safe because no such use sets such a long nonce value. However user
1929 applications that use this cipher directly and set a non-default nonce
1930 length to be longer than 12 bytes may be vulnerable.
1932 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1938 * Added SCA hardening for modular field inversion in EC_GROUP through
1939 a new dedicated field_inv() pointer in EC_METHOD.
1940 This also addresses a leakage affecting conversions from projective
1941 to affine coordinates.
1943 *Billy Bob Brumley, Nicola Tuveri*
1945 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1946 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1950 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1954 * Remove the 'dist' target and add a tarball building script. The
1955 'dist' target has fallen out of use, and it shouldn't be
1956 necessary to configure just to create a source distribution.
1960 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
1962 * Timing vulnerability in DSA signature generation
1964 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1965 timing side channel attack. An attacker could use variations in the signing
1966 algorithm to recover the private key.
1968 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1973 * Timing vulnerability in ECDSA signature generation
1975 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1976 timing side channel attack. An attacker could use variations in the signing
1977 algorithm to recover the private key.
1979 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1984 * Add coordinate blinding for EC_POINT and implement projective
1985 coordinate blinding for generic prime curves as a countermeasure to
1986 chosen point SCA attacks.
1988 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1990 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
1992 * Client DoS due to large DH parameter
1994 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1995 malicious server can send a very large prime value to the client. This will
1996 cause the client to spend an unreasonably long period of time generating a
1997 key for this prime resulting in a hang until the client has finished. This
1998 could be exploited in a Denial Of Service attack.
2000 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2005 * Cache timing vulnerability in RSA Key Generation
2007 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2008 a cache timing side channel attack. An attacker with sufficient access to
2009 mount cache timing attacks during the RSA key generation process could
2010 recover the private key.
2012 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2013 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2018 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2019 parameter is no longer accepted, as it leads to a corrupt table. NULL
2020 pem_str is reserved for alias entries only.
2024 * Revert blinding in ECDSA sign and instead make problematic addition
2025 length-invariant. Switch even to fixed-length Montgomery multiplication.
2029 * Change generating and checking of primes so that the error rate of not
2030 being prime depends on the intended use based on the size of the input.
2031 For larger primes this will result in more rounds of Miller-Rabin.
2032 The maximal error rate for primes with more than 1080 bits is lowered
2035 *Kurt Roeckx, Annie Yousar*
2037 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2041 * Add blinding to ECDSA and DSA signatures to protect against side channel
2042 attacks discovered by Keegan Ryan (NCC Group).
2046 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2047 now allow empty (zero character) pass phrases.
2051 * Certificate time validation (X509_cmp_time) enforces stricter
2052 compliance with RFC 5280. Fractional seconds and timezone offsets
2053 are no longer allowed.
2057 * Fixed a text canonicalisation bug in CMS
2059 Where a CMS detached signature is used with text content the text goes
2060 through a canonicalisation process first prior to signing or verifying a
2061 signature. This process strips trailing space at the end of lines, converts
2062 line terminators to CRLF and removes additional trailing line terminators
2063 at the end of a file. A bug in the canonicalisation process meant that
2064 some characters, such as form-feed, were incorrectly treated as whitespace
2065 and removed. This is contrary to the specification (RFC5485). This fix
2066 could mean that detached text data signed with an earlier version of
2067 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2068 signed with a fixed OpenSSL may fail to verify with an earlier version of
2069 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2070 and use the "-binary" flag (for the "cms" command line application) or set
2071 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2075 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2077 * Constructed ASN.1 types with a recursive definition could exceed the stack
2079 Constructed ASN.1 types with a recursive definition (such as can be found
2080 in PKCS7) could eventually exceed the stack given malicious input with
2081 excessive recursion. This could result in a Denial Of Service attack. There
2082 are no such structures used within SSL/TLS that come from untrusted sources
2083 so this is considered safe.
2085 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2091 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2093 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2094 effectively reduced to only comparing the least significant bit of each
2095 byte. This allows an attacker to forge messages that would be considered as
2096 authenticated in an amount of tries lower than that guaranteed by the
2097 security claims of the scheme. The module can only be compiled by the
2098 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2100 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2106 * Add a build target 'build_all_generated', to build all generated files
2107 and only that. This can be used to prepare everything that requires
2108 things like perl for a system that lacks perl and then move everything
2109 to that system and do the rest of the build there.
2113 * Backport SSL_OP_NO_RENGOTIATION
2115 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2116 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2117 changes this is no longer possible in 1.1.0. Therefore the new
2118 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2119 1.1.0 to provide equivalent functionality.
2121 Note that if an application built against 1.1.0h headers (or above) is run
2122 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2123 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2127 * Removed the OS390-Unix config target. It relied on a script that doesn't
2132 * rsaz_1024_mul_avx2 overflow bug on x86_64
2134 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2135 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2136 Analysis suggests that attacks against RSA and DSA as a result of this
2137 defect would be very difficult to perform and are not believed likely.
2138 Attacks against DH1024 are considered just feasible, because most of the
2139 work necessary to deduce information about a private key may be performed
2140 offline. The amount of resources required for such an attack would be
2141 significant. However, for an attack on TLS to be meaningful, the server
2142 would have to share the DH1024 private key among multiple clients, which is
2143 no longer an option since CVE-2016-0701.
2145 This only affects processors that support the AVX2 but not ADX extensions
2146 like Intel Haswell (4th generation).
2148 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2149 was originally found via the OSS-Fuzz project.
2154 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2156 * bn_sqrx8x_internal carry bug on x86_64
2158 There is a carry propagating bug in the x86_64 Montgomery squaring
2159 procedure. No EC algorithms are affected. Analysis suggests that attacks
2160 against RSA and DSA as a result of this defect would be very difficult to
2161 perform and are not believed likely. Attacks against DH are considered just
2162 feasible (although very difficult) because most of the work necessary to
2163 deduce information about a private key may be performed offline. The amount
2164 of resources required for such an attack would be very significant and
2165 likely only accessible to a limited number of attackers. An attacker would
2166 additionally need online access to an unpatched system using the target
2167 private key in a scenario with persistent DH parameters and a private
2168 key that is shared between multiple clients.
2170 This only affects processors that support the BMI1, BMI2 and ADX extensions
2171 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2173 This issue was reported to OpenSSL by the OSS-Fuzz project.
2178 * Malformed X.509 IPAddressFamily could cause OOB read
2180 If an X.509 certificate has a malformed IPAddressFamily extension,
2181 OpenSSL could do a one-byte buffer overread. The most likely result
2182 would be an erroneous display of the certificate in text format.
2184 This issue was reported to OpenSSL by the OSS-Fuzz project.
2189 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2191 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2192 platform rather than 'mingw'.
2196 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2197 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2198 which is the minimum version we support.
2202 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2204 * Encrypt-Then-Mac renegotiation crash
2206 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2207 negotiated where it was not in the original handshake (or vice-versa) then
2208 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2209 and servers are affected.
2211 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2216 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2218 * Truncated packet could crash via OOB read
2220 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2221 cipher is being used, then a truncated packet can cause that host to
2222 perform an out-of-bounds read, usually resulting in a crash.
2224 This issue was reported to OpenSSL by Robert Święcki of Google.
2229 * Bad (EC)DHE parameters cause a client crash
2231 If a malicious server supplies bad parameters for a DHE or ECDHE key
2232 exchange then this can result in the client attempting to dereference a
2233 NULL pointer leading to a client crash. This could be exploited in a Denial
2236 This issue was reported to OpenSSL by Guido Vranken.
2241 * BN_mod_exp may produce incorrect results on x86_64
2243 There is a carry propagating bug in the x86_64 Montgomery squaring
2244 procedure. No EC algorithms are affected. Analysis suggests that attacks
2245 against RSA and DSA as a result of this defect would be very difficult to
2246 perform and are not believed likely. Attacks against DH are considered just
2247 feasible (although very difficult) because most of the work necessary to
2248 deduce information about a private key may be performed offline. The amount
2249 of resources required for such an attack would be very significant and
2250 likely only accessible to a limited number of attackers. An attacker would
2251 additionally need online access to an unpatched system using the target
2252 private key in a scenario with persistent DH parameters and a private
2253 key that is shared between multiple clients. For example this can occur by
2254 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2255 similar to CVE-2015-3193 but must be treated as a separate problem.
2257 This issue was reported to OpenSSL by the OSS-Fuzz project.
2262 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2264 * ChaCha20/Poly1305 heap-buffer-overflow
2266 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2267 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2268 crash. This issue is not considered to be exploitable beyond a DoS.
2270 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2275 * CMS Null dereference
2277 Applications parsing invalid CMS structures can crash with a NULL pointer
2278 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2279 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2280 structure callback if an attempt is made to free certain invalid encodings.
2281 Only CHOICE structures using a callback which do not handle NULL value are
2284 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2289 * Montgomery multiplication may produce incorrect results
2291 There is a carry propagating bug in the Broadwell-specific Montgomery
2292 multiplication procedure that handles input lengths divisible by, but
2293 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2294 and DH private keys are impossible. This is because the subroutine in
2295 question is not used in operations with the private key itself and an input
2296 of the attacker's direct choice. Otherwise the bug can manifest itself as
2297 transient authentication and key negotiation failures or reproducible
2298 erroneous outcome of public-key operations with specially crafted input.
2299 Among EC algorithms only Brainpool P-512 curves are affected and one
2300 presumably can attack ECDH key negotiation. Impact was not analyzed in
2301 detail, because pre-requisites for attack are considered unlikely. Namely
2302 multiple clients have to choose the curve in question and the server has to
2303 share the private key among them, neither of which is default behaviour.
2304 Even then only clients that chose the curve will be affected.
2306 This issue was publicly reported as transient failures and was not
2307 initially recognized as a security issue. Thanks to Richard Morgan for
2308 providing reproducible case.
2313 * Removed automatic addition of RPATH in shared libraries and executables,
2314 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2318 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2320 * Fix Use After Free for large message sizes
2322 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2323 message larger than approx 16k is received then the underlying buffer to
2324 store the incoming message is reallocated and moved. Unfortunately a
2325 dangling pointer to the old location is left which results in an attempt to
2326 write to the previously freed location. This is likely to result in a
2327 crash, however it could potentially lead to execution of arbitrary code.
2329 This issue only affects OpenSSL 1.1.0a.
2331 This issue was reported to OpenSSL by Robert Święcki.
2336 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2338 * OCSP Status Request extension unbounded memory growth
2340 A malicious client can send an excessively large OCSP Status Request
2341 extension. If that client continually requests renegotiation, sending a
2342 large OCSP Status Request extension each time, then there will be unbounded
2343 memory growth on the server. This will eventually lead to a Denial Of
2344 Service attack through memory exhaustion. Servers with a default
2345 configuration are vulnerable even if they do not support OCSP. Builds using
2346 the "no-ocsp" build time option are not affected.
2348 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2353 * SSL_peek() hang on empty record
2355 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2356 sends an empty record. This could be exploited by a malicious peer in a
2357 Denial Of Service attack.
2359 This issue was reported to OpenSSL by Alex Gaynor.
2364 * Excessive allocation of memory in tls_get_message_header() and
2365 dtls1_preprocess_fragment()
2367 A (D)TLS message includes 3 bytes for its length in the header for the
2368 message. This would allow for messages up to 16Mb in length. Messages of
2369 this length are excessive and OpenSSL includes a check to ensure that a
2370 peer is sending reasonably sized messages in order to avoid too much memory
2371 being consumed to service a connection. A flaw in the logic of version
2372 1.1.0 means that memory for the message is allocated too early, prior to
2373 the excessive message length check. Due to way memory is allocated in
2374 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2375 to service a connection. This could lead to a Denial of Service through
2376 memory exhaustion. However, the excessive message length check still takes
2377 place, and this would cause the connection to immediately fail. Assuming
2378 that the application calls SSL_free() on the failed connection in a timely
2379 manner then the 21Mb of allocated memory will then be immediately freed
2380 again. Therefore the excessive memory allocation will be transitory in
2381 nature. This then means that there is only a security impact if:
2383 1) The application does not call SSL_free() in a timely manner in the event
2384 that the connection fails
2386 2) The application is working in a constrained environment where there is
2387 very little free memory
2389 3) The attacker initiates multiple connection attempts such that there are
2390 multiple connections in a state where memory has been allocated for the
2391 connection; SSL_free() has not yet been called; and there is insufficient
2392 memory to service the multiple requests.
2394 Except in the instance of (1) above any Denial Of Service is likely to be
2395 transitory because as soon as the connection fails the memory is
2396 subsequently freed again in the SSL_free() call. However there is an
2397 increased risk during this period of application crashes due to the lack of
2398 memory - which would then mean a more serious Denial of Service.
2400 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2401 (CVE-2016-6307 and CVE-2016-6308)
2405 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2406 had to be removed. Primary reason is that vendor assembler can't
2407 assemble our modules with -KPIC flag. As result it, assembly
2408 support, was not even available as option. But its lack means
2409 lack of side-channel resistant code, which is incompatible with
2410 security by todays standards. Fortunately gcc is readily available
2411 prepackaged option, which we firmly point at...
2415 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2417 * Windows command-line tool supports UTF-8 opt-in option for arguments
2418 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2419 (to any value) allows Windows user to access PKCS#12 file generated
2420 with Windows CryptoAPI and protected with non-ASCII password, as well
2421 as files generated under UTF-8 locale on Linux also protected with
2426 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2427 have been disabled by default and removed from DEFAULT, just like RC4.
2428 See the RC4 item below to re-enable both.
2432 * The method for finding the storage location for the Windows RAND seed file
2433 has changed. First we check %RANDFILE%. If that is not set then we check
2434 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2435 all else fails we fall back to C:\.
2439 * The EVP_EncryptUpdate() function has had its return type changed from void
2440 to int. A return of 0 indicates and error while a return of 1 indicates
2445 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2446 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2447 off the constant time implementation for RSA, DSA and DH have been made
2448 no-ops and deprecated.
2452 * Windows RAND implementation was simplified to only get entropy by
2453 calling CryptGenRandom(). Various other RAND-related tickets
2456 *Joseph Wylie Yandle, Rich Salz*
2458 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2459 and `OPENSSL_LH_`, respectively. The old names are available
2460 with API compatibility. They new names are now completely documented.
2464 * Unify TYPE_up_ref(obj) methods signature.
2465 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2466 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2467 int (instead of void) like all others TYPE_up_ref() methods.
2468 So now these methods also check the return value of CRYPTO_atomic_add(),
2469 and the validity of object reference counter.
2471 *fdasilvayy@gmail.com*
2473 * With Windows Visual Studio builds, the .pdb files are installed
2474 alongside the installed libraries and executables. For a static
2475 library installation, ossl_static.pdb is the associate compiler
2476 generated .pdb file to be used when linking programs.
2480 * Remove openssl.spec. Packaging files belong with the packagers.
2484 * Automatic Darwin/OSX configuration has had a refresh, it will now
2485 recognise x86_64 architectures automatically. You can still decide
2486 to build for a different bitness with the environment variable
2487 KERNEL_BITS (can be 32 or 64), for example:
2489 KERNEL_BITS=32 ./config
2493 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2494 256 bit AES and HMAC with SHA256.
2498 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2502 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2506 * To enable users to have their own config files and build file templates,
2507 Configure looks in the directory indicated by the environment variable
2508 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2509 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2510 name and is used as is.
2514 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2515 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2516 X509_CERT_FILE_CTX was removed.
2520 * "shared" builds are now the default. To create only static libraries use
2521 the "no-shared" Configure option.
2525 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2526 All of these option have not worked for some while and are fundamental
2531 * Make various cleanup routines no-ops and mark them as deprecated. Most
2532 global cleanup functions are no longer required because they are handled
2533 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2534 Explicitly de-initing can cause problems (e.g. where a library that uses
2535 OpenSSL de-inits, but an application is still using it). The affected
2536 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2537 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2538 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2539 COMP_zlib_cleanup().
2543 * --strict-warnings no longer enables runtime debugging options
2544 such as REF_DEBUG. Instead, debug options are automatically
2545 enabled with '--debug' builds.
2547 *Andy Polyakov, Emilia Käsper*
2549 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2550 have been moved out of the public header files. New functions for managing
2551 these have been added.
2555 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2556 objects have been moved out of the public header files. New
2557 functions for managing these have been added.
2561 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2562 have been moved out of the public header files. New functions for managing
2563 these have been added.
2567 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2568 moved out of the public header files. New functions for managing these
2573 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2577 * Removed the mk1mf build scripts.
2581 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2582 it is always safe to #include a header now.
2586 * Removed the aged BC-32 config and all its supporting scripts
2590 * Removed support for Ultrix, Netware, and OS/2.
2594 * Add support for HKDF.
2596 *Alessandro Ghedini*
2598 * Add support for blake2b and blake2s
2602 * Added support for "pipelining". Ciphers that have the
2603 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2604 encryptions/decryptions simultaneously. There are currently no built-in
2605 ciphers with this property but the expectation is that engines will be able
2606 to offer it to significantly improve throughput. Support has been extended
2607 into libssl so that multiple records for a single connection can be
2608 processed in one go (for >=TLS 1.1).
2612 * Added the AFALG engine. This is an async capable engine which is able to
2613 offload work to the Linux kernel. In this initial version it only supports
2614 AES128-CBC. The kernel must be version 4.1.0 or greater.
2618 * OpenSSL now uses a new threading API. It is no longer necessary to
2619 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2620 are two supported threading models: pthreads and windows threads. It is
2621 also possible to configure OpenSSL at compile time for "no-threads". The
2622 old threading API should no longer be used. The functions have been
2623 replaced with "no-op" compatibility macros.
2625 *Alessandro Ghedini, Matt Caswell*
2627 * Modify behavior of ALPN to invoke callback after SNI/servername
2628 callback, such that updates to the SSL_CTX affect ALPN.
2632 * Add SSL_CIPHER queries for authentication and key-exchange.
2636 * Changes to the DEFAULT cipherlist:
2637 - Prefer (EC)DHE handshakes over plain RSA.
2638 - Prefer AEAD ciphers over legacy ciphers.
2639 - Prefer ECDSA over RSA when both certificates are available.
2640 - Prefer TLSv1.2 ciphers/PRF.
2641 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2646 * Change the ECC default curve list to be this, in order: x25519,
2647 secp256r1, secp521r1, secp384r1.
2651 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2652 disabled by default. They can be re-enabled using the
2653 enable-weak-ssl-ciphers option to Configure.
2657 * If the server has ALPN configured, but supports no protocols that the
2658 client advertises, send a fatal "no_application_protocol" alert.
2659 This behaviour is SHALL in RFC 7301, though it isn't universally
2660 implemented by other servers.
2664 * Add X25519 support.
2665 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2666 for public and private key encoding using the format documented in
2667 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2668 key generation and key derivation.
2670 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2675 * Deprecate SRP_VBASE_get_by_user.
2676 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2677 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2678 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2679 seed, even if the seed is configured.
2681 Users should use SRP_VBASE_get1_by_user instead. Note that in
2682 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2683 also that even though configuring the SRP seed attempts to hide
2684 invalid usernames by continuing the handshake with fake
2685 credentials, this behaviour is not constant time and no strong
2686 guarantees are made that the handshake is indistinguishable from
2687 that of a valid user.
2691 * Configuration change; it's now possible to build dynamic engines
2692 without having to build shared libraries and vice versa. This
2693 only applies to the engines in engines/, those in crypto/engine/
2694 will always be built into libcrypto (i.e. "static").
2696 Building dynamic engines is enabled by default; to disable, use
2697 the configuration option "disable-dynamic-engine".
2699 The only requirements for building dynamic engines are the
2700 presence of the DSO module and building with position independent
2701 code, so they will also automatically be disabled if configuring
2702 with "disable-dso" or "disable-pic".
2704 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2705 are also taken away from openssl/opensslconf.h, as they are
2710 * Configuration change; if there is a known flag to compile
2711 position independent code, it will always be applied on the
2712 libcrypto and libssl object files, and never on the application
2713 object files. This means other libraries that use routines from
2714 libcrypto / libssl can be made into shared libraries regardless
2715 of how OpenSSL was configured.
2717 If this isn't desirable, the configuration options "disable-pic"
2718 or "no-pic" can be used to disable the use of PIC. This will
2719 also disable building shared libraries and dynamic engines.
2723 * Removed JPAKE code. It was experimental and has no wide use.
2727 * The INSTALL_PREFIX Makefile variable has been renamed to
2728 DESTDIR. That makes for less confusion on what this variable
2729 is for. Also, the configuration option --install_prefix is
2734 * Heartbeat for TLS has been removed and is disabled by default
2735 for DTLS; configure with enable-heartbeats. Code that uses the
2736 old #define's might need to be updated.
2738 *Emilia Käsper, Rich Salz*
2740 * Rename REF_CHECK to REF_DEBUG.
2744 * New "unified" build system
2746 The "unified" build system is aimed to be a common system for all
2747 platforms we support. With it comes new support for VMS.
2749 This system builds supports building in a different directory tree
2750 than the source tree. It produces one Makefile (for unix family
2751 or lookalikes), or one descrip.mms (for VMS).
2753 The source of information to make the Makefile / descrip.mms is
2754 small files called 'build.info', holding the necessary
2755 information for each directory with source to compile, and a
2756 template in Configurations, like unix-Makefile.tmpl or
2759 With this change, the library names were also renamed on Windows
2760 and on VMS. They now have names that are closer to the standard
2761 on Unix, and include the major version number, and in certain
2762 cases, the architecture they are built for. See "Notes on shared
2763 libraries" in INSTALL.
2765 We rely heavily on the perl module Text::Template.
2769 * Added support for auto-initialisation and de-initialisation of the library.
2770 OpenSSL no longer requires explicit init or deinit routines to be called,
2771 except in certain circumstances. See the OPENSSL_init_crypto() and
2772 OPENSSL_init_ssl() man pages for further information.
2776 * The arguments to the DTLSv1_listen function have changed. Specifically the
2777 "peer" argument is now expected to be a BIO_ADDR object.
2779 * Rewrite of BIO networking library. The BIO library lacked consistent
2780 support of IPv6, and adding it required some more extensive
2781 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2782 which hold all types of addresses and chains of address information.
2783 It also introduces a new API, with functions like BIO_socket,
2784 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2785 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2786 have been adapted accordingly.
2790 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2795 * CRIME protection: disable compression by default, even if OpenSSL is
2796 compiled with zlib enabled. Applications can still enable compression
2797 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2798 using the SSL_CONF library to configure compression.
2802 * The signature of the session callback configured with
2803 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2804 was explicitly marked as `const unsigned char*` instead of
2809 * Always DPURIFY. Remove the use of uninitialized memory in the
2810 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2814 * Removed many obsolete configuration items, including
2815 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2816 MD2_CHAR, MD2_INT, MD2_LONG
2818 IDEA_SHORT, IDEA_LONG
2819 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2821 *Rich Salz, with advice from Andy Polyakov*
2823 * Many BN internals have been moved to an internal header file.
2825 *Rich Salz with help from Andy Polyakov*
2827 * Configuration and writing out the results from it has changed.
2828 Files such as Makefile include/openssl/opensslconf.h and are now
2829 produced through general templates, such as Makefile.in and
2830 crypto/opensslconf.h.in and some help from the perl module
2833 Also, the center of configuration information is no longer
2834 Makefile. Instead, Configure produces a perl module in
2835 configdata.pm which holds most of the config data (in the hash
2836 table %config), the target data that comes from the target
2837 configuration in one of the `Configurations/*.conf~ files (in
2842 * To clarify their intended purposes, the Configure options
2843 --prefix and --openssldir change their semantics, and become more
2844 straightforward and less interdependent.
2846 --prefix shall be used exclusively to give the location INSTALLTOP
2847 where programs, scripts, libraries, include files and manuals are
2848 going to be installed. The default is now /usr/local.
2850 --openssldir shall be used exclusively to give the default
2851 location OPENSSLDIR where certificates, private keys, CRLs are
2852 managed. This is also where the default openssl.cnf gets
2854 If the directory given with this option is a relative path, the
2855 values of both the --prefix value and the --openssldir value will
2856 be combined to become OPENSSLDIR.
2857 The default for --openssldir is INSTALLTOP/ssl.
2859 Anyone who uses --openssldir to specify where OpenSSL is to be
2860 installed MUST change to use --prefix instead.
2864 * The GOST engine was out of date and therefore it has been removed. An up
2865 to date GOST engine is now being maintained in an external repository.
2866 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2867 support for GOST ciphersuites (these are only activated if a GOST engine
2872 * EGD is no longer supported by default; use enable-egd when
2875 *Ben Kaduk and Rich Salz*
2877 * The distribution now has Makefile.in files, which are used to
2878 create Makefile's when Configure is run. *Configure must be run
2879 before trying to build now.*
2883 * The return value for SSL_CIPHER_description() for error conditions
2888 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2890 Obtaining and performing DNSSEC validation of TLSA records is
2891 the application's responsibility. The application provides
2892 the TLSA records of its choice to OpenSSL, and these are then
2893 used to authenticate the peer.
2895 The TLSA records need not even come from DNS. They can, for
2896 example, be used to implement local end-entity certificate or
2897 trust-anchor "pinning", where the "pin" data takes the form
2898 of TLSA records, which can augment or replace verification
2899 based on the usual WebPKI public certification authorities.
2903 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2904 continues to support deprecated interfaces in default builds.
2905 However, applications are strongly advised to compile their
2906 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2907 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2908 or the 1.1.0 releases.
2910 In environments in which all applications have been ported to
2911 not use any deprecated interfaces OpenSSL's Configure script
2912 should be used with the --api=1.1.0 option to entirely remove
2913 support for the deprecated features from the library and
2914 unconditionally disable them in the installed headers.
2915 Essentially the same effect can be achieved with the "no-deprecated"
2916 argument to Configure, except that this will always restrict
2917 the build to just the latest API, rather than a fixed API
2920 As applications are ported to future revisions of the API,
2921 they should update their compile-time OPENSSL_API_COMPAT define
2922 accordingly, but in most cases should be able to continue to
2923 compile with later releases.
2925 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2926 0x10000000L and 0x00908000L, respectively. However those
2927 versions did not support the OPENSSL_API_COMPAT feature, and
2928 so applications are not typically tested for explicit support
2929 of just the undeprecated features of either release.
2933 * Add support for setting the minimum and maximum supported protocol.
2934 It can bet set via the SSL_set_min_proto_version() and
2935 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2936 MaxProtocol. It's recommended to use the new APIs to disable
2937 protocols instead of disabling individual protocols using
2938 SSL_set_options() or SSL_CONF's Protocol. This change also
2939 removes support for disabling TLS 1.2 in the OpenSSL TLS
2940 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2944 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2948 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2949 and integrates ECDSA and ECDH functionality into EC. Implementations can
2950 now redirect key generation and no longer need to convert to or from
2953 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2954 include the ec.h header file instead.
2958 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2959 ciphers who are no longer supported and drops support the ephemeral RSA key
2960 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2964 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2965 opaque. For HMAC_CTX, the following constructors and destructors
2968 HMAC_CTX *HMAC_CTX_new(void);
2969 void HMAC_CTX_free(HMAC_CTX *ctx);
2971 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2972 destroy such methods has been added. See EVP_MD_meth_new(3) and
2973 EVP_CIPHER_meth_new(3) for documentation.
2976 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2977 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2978 EVP_MD_CTX_reset() should be called instead to reinitialise
2979 an already created structure.
2980 2) For consistency with the majority of our object creators and
2981 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2982 EVP_MD_CTX_(new|free). The old names are retained as macros
2983 for deprecated builds.
2987 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2988 cryptographic operations to be performed asynchronously as long as an
2989 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2990 further details. Libssl has also had this capability integrated with the
2991 introduction of the new mode SSL_MODE_ASYNC and associated error
2992 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2993 pages. This work was developed in partnership with Intel Corp.
2997 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2998 always enabled now. If you want to disable the support you should
2999 exclude it using the list of supported ciphers. This also means that the
3000 "-no_ecdhe" option has been removed from s_server.
3004 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3005 SSL_{CTX_}set1_curves() which can set a list.
3009 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3010 curve you want to support using SSL_{CTX_}set1_curves().
3014 * State machine rewrite. The state machine code has been significantly
3015 refactored in order to remove much duplication of code and solve issues
3016 with the old code (see ssl/statem/README for further details). This change
3017 does have some associated API changes. Notably the SSL_state() function
3018 has been removed and replaced by SSL_get_state which now returns an
3019 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
3020 altogether. The previous handshake states defined in ssl.h and ssl3.h have
3025 * All instances of the string "ssleay" in the public API were replaced
3026 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3027 Some error codes related to internal RSA_eay API's were renamed.
3031 * The demo files in crypto/threads were moved to demo/threads.
3035 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3038 *Matt Caswell, Rich Salz*
3040 * New ASN.1 embed macro.
3042 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3043 structure is not allocated: it is part of the parent. That is instead of
3051 This reduces memory fragmentation and make it impossible to accidentally
3052 set a mandatory field to NULL.
3054 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3055 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3056 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3061 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3065 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3066 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3067 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3068 DES and RC4 ciphersuites.
3072 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3073 This changes the decoding behaviour for some invalid messages,
3074 though the change is mostly in the more lenient direction, and
3075 legacy behaviour is preserved as much as possible.
3079 * Fix no-stdio build.
3080 * David Woodhouse <David.Woodhouse@intel.com> and also
3081 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3083 * New testing framework
3084 The testing framework has been largely rewritten and is now using
3085 perl and the perl modules Test::Harness and an extended variant of
3086 Test::More called OpenSSL::Test to do its work. All test scripts in
3087 test/ have been rewritten into test recipes, and all direct calls to
3088 executables in test/Makefile have become individual recipes using the
3089 simplified testing OpenSSL::Test::Simple.
3091 For documentation on our testing modules, do:
3093 perldoc test/testlib/OpenSSL/Test/Simple.pm
3094 perldoc test/testlib/OpenSSL/Test.pm
3098 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3099 are used; the latter aborts on memory leaks (usually checked on exit).
3100 Some undocumented "set malloc, etc., hooks" functions were removed
3101 and others were changed. All are now documented.
3105 * In DSA_generate_parameters_ex, if the provided seed is too short,
3108 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3110 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3111 from RFC4279, RFC4785, RFC5487, RFC5489.
3113 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3114 original RSA_PSK patch.
3118 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3119 era flag was never set throughout the codebase (only read). Also removed
3120 SSL3_FLAGS_POP_BUFFER which was only used if
3121 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3125 * Changed the default name options in the "ca", "crl", "req" and "x509"
3126 to be "oneline" instead of "compat".
3130 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3131 not aware of clients that still exhibit this bug, and the workaround
3132 hasn't been working properly for a while.
3136 * The return type of BIO_number_read() and BIO_number_written() as well as
3137 the corresponding num_read and num_write members in the BIO structure has
3138 changed from unsigned long to uint64_t. On platforms where an unsigned
3139 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3144 * Given the pervasive nature of TLS extensions it is inadvisable to run
3145 OpenSSL without support for them. It also means that maintaining
3146 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3147 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3151 * Removed support for the two export grade static DH ciphersuites
3152 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3153 were newly added (along with a number of other static DH ciphersuites) to
3154 1.0.2. However the two export ones have *never* worked since they were
3155 introduced. It seems strange in any case to be adding new export
3156 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3160 * Version negotiation has been rewritten. In particular SSLv23_method(),
3161 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3162 and turned into macros which simply call the new preferred function names
3163 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3164 should use the new names instead. Also as part of this change the ssl23.h
3165 header file has been removed.
3169 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3170 code and the associated standard is no longer considered fit-for-purpose.
3174 * RT2547 was closed. When generating a private key, try to make the
3175 output file readable only by the owner. This behavior change might
3176 be noticeable when interacting with other software.
3178 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3183 * Added HTTP GET support to the ocsp command.
3187 * Changed default digest for the dgst and enc commands from MD5 to
3192 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3196 * Added support for TLS extended master secret from
3197 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3198 initial patch which was a great help during development.
3202 * All libssl internal structures have been removed from the public header
3203 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3204 now redundant). Users should not attempt to access internal structures
3205 directly. Instead they should use the provided API functions.
3209 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3210 Access to deprecated functions can be re-enabled by running config with
3211 "enable-deprecated". In addition applications wishing to use deprecated
3212 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3213 will, by default, disable some transitive includes that previously existed
3214 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3218 * Added support for OCB mode. OpenSSL has been granted a patent license
3219 compatible with the OpenSSL license for use of OCB. Details are available
3220 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3221 for OCB can be removed by calling config with no-ocb.
3225 * SSLv2 support has been removed. It still supports receiving a SSLv2
3226 compatible client hello.
3230 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3231 done while fixing the error code for the key-too-small case.
3233 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3235 * CA.sh has been removed; use CA.pl instead.
3239 * Removed old DES API.
3243 * Remove various unsupported platforms:
3249 Sinix/ReliantUNIX RM400
3254 16-bit platforms such as WIN16
3258 * Clean up OPENSSL_NO_xxx #define's
3259 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3260 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3261 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3262 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3263 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3264 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3265 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3266 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3267 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3268 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3272 * Cleaned up dead code
3273 Remove all but one '#ifdef undef' which is to be looked at.
3277 * Clean up calling of xxx_free routines.
3278 Just like free(), fix most of the xxx_free routines to accept
3279 NULL. Remove the non-null checks from callers. Save much code.
3283 * Add secure heap for storage of private keys (when possible).
3284 Add BIO_s_secmem(), CBIGNUM, etc.
3285 Contributed by Akamai Technologies under our Corporate CLA.
3289 * Experimental support for a new, fast, unbiased prime candidate generator,
3290 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3292 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3294 * New output format NSS in the sess_id command line tool. This allows
3295 exporting the session id and the master key in NSS keylog format.
3297 *Martin Kaiser <martin@kaiser.cx>*
3299 * Harmonize version and its documentation. -f flag is used to display
3302 *mancha <mancha1@zoho.com>*
3304 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3305 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3307 *mancha <mancha1@zoho.com>*
3309 * Fix some double frees. These are not thought to be exploitable.
3311 *mancha <mancha1@zoho.com>*
3313 * A missing bounds check in the handling of the TLS heartbeat extension
3314 can be used to reveal up to 64k of memory to a connected client or
3317 Thanks for Neel Mehta of Google Security for discovering this bug and to
3318 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3319 preparing the fix [CVE-2014-0160][]
3321 *Adam Langley, Bodo Moeller*
3323 * Fix for the attack described in the paper "Recovering OpenSSL
3324 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3325 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3326 <http://eprint.iacr.org/2014/140>
3328 Thanks to Yuval Yarom and Naomi Benger for discovering this
3329 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3331 *Yuval Yarom and Naomi Benger*
3333 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3334 this fixes a limitation in previous versions of OpenSSL.
3338 * Experimental encrypt-then-mac support.
3340 Experimental support for encrypt then mac from
3341 draft-gutmann-tls-encrypt-then-mac-02.txt
3343 To enable it set the appropriate extension number (0x42 for the test
3344 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3346 For non-compliant peers (i.e. just about everything) this should have no
3349 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3353 * Add EVP support for key wrapping algorithms, to avoid problems with
3354 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3355 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3356 algorithms and include tests cases.
3360 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3365 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3366 MGF1 digest and OAEP label.
3370 * Make openssl verify return errors.
3372 *Chris Palmer <palmer@google.com> and Ben Laurie*
3374 * New function ASN1_TIME_diff to calculate the difference between two
3375 ASN1_TIME structures or one structure and the current time.
3379 * Update fips_test_suite to support multiple command line options. New
3380 test to induce all self test errors in sequence and check expected
3385 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3386 sign or verify all in one operation.
3390 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3391 test programs and fips_test_suite. Includes functionality to parse
3392 the minimal script output of fipsalgest.pl directly.
3396 * Add authorisation parameter to FIPS_module_mode_set().
3400 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3404 * Use separate DRBG fields for internal and external flags. New function
3405 FIPS_drbg_health_check() to perform on demand health checking. Add
3406 generation tests to fips_test_suite with reduced health check interval to
3407 demonstrate periodic health checking. Add "nodh" option to
3408 fips_test_suite to skip very slow DH test.
3412 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3417 * More extensive health check for DRBG checking many more failure modes.
3418 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3419 combination: call this in fips_test_suite.
3423 * Add support for canonical generation of DSA parameter 'g'. See
3426 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3427 POST to handle HMAC cases.
3431 * Add functions FIPS_module_version() and FIPS_module_version_text()
3432 to return numerical and string versions of the FIPS module number.
3436 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3437 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3438 outside the validated module in the FIPS capable OpenSSL.
3442 * Minor change to DRBG entropy callback semantics. In some cases
3443 there is no multiple of the block length between min_len and
3444 max_len. Allow the callback to return more than max_len bytes
3445 of entropy but discard any extra: it is the callback's responsibility
3446 to ensure that the extra data discarded does not impact the
3447 requested amount of entropy.
3451 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3452 information in FIPS186-3, SP800-57 and SP800-131A.
3456 * CCM support via EVP. Interface is very similar to GCM case except we
3457 must supply all data in one chunk (i.e. no update, final) and the
3458 message length must be supplied if AAD is used. Add algorithm test
3463 * Initial version of POST overhaul. Add POST callback to allow the status
3464 of POST to be monitored and/or failures induced. Modify fips_test_suite
3465 to use callback. Always run all selftests even if one fails.
3469 * XTS support including algorithm test driver in the fips_gcmtest program.
3470 Note: this does increase the maximum key length from 32 to 64 bytes but
3471 there should be no binary compatibility issues as existing applications
3472 will never use XTS mode.
3476 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3477 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3478 performs algorithm blocking for unapproved PRNG types. Also do not
3479 set PRNG type in FIPS_mode_set(): leave this to the application.
3480 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3481 the standard OpenSSL PRNG: set additional data to a date time vector.
3485 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3486 This shouldn't present any incompatibility problems because applications
3487 shouldn't be using these directly and any that are will need to rethink
3488 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3492 * Extensive self tests and health checking required by SP800-90 DRBG.
3493 Remove strength parameter from FIPS_drbg_instantiate and always
3494 instantiate at maximum supported strength.
3498 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3502 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3506 * New function DH_compute_key_padded() to compute a DH key and pad with
3507 leading zeroes if needed: this complies with SP800-56A et al.
3511 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3512 anything, incomplete, subject to change and largely untested at present.
3516 * Modify fipscanisteronly build option to only build the necessary object
3517 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3521 * Add experimental option FIPSSYMS to give all symbols in
3522 fipscanister.o and FIPS or fips prefix. This will avoid
3523 conflicts with future versions of OpenSSL. Add perl script
3524 util/fipsas.pl to preprocess assembly language source files
3525 and rename any affected symbols.
3529 * Add selftest checks and algorithm block of non-fips algorithms in
3530 FIPS mode. Remove DES2 from selftests.
3534 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3535 return internal method without any ENGINE dependencies. Add new
3536 tiny fips sign and verify functions.
3540 * New build option no-ec2m to disable characteristic 2 code.
3544 * New build option "fipscanisteronly". This only builds fipscanister.o
3545 and (currently) associated fips utilities. Uses the file Makefile.fips
3546 instead of Makefile.org as the prototype.
3550 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3551 Update fips_gcmtest to use IV generator.
3555 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3556 setting output buffer to NULL. The `*Final` function must be
3557 called although it will not retrieve any additional data. The tag
3558 can be set or retrieved with a ctrl. The IV length is by default 12
3559 bytes (96 bits) but can be set to an alternative value. If the IV
3560 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3565 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3566 underlying do_cipher function handles all cipher semantics itself
3567 including padding and finalisation. This is useful if (for example)
3568 an ENGINE cipher handles block padding itself. The behaviour of
3569 do_cipher is subtly changed if this flag is set: the return value
3570 is the number of characters written to the output buffer (zero is
3571 no longer an error code) or a negative error code. Also if the
3572 input buffer is NULL and length 0 finalisation should be performed.
3576 * If a candidate issuer certificate is already part of the constructed
3577 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3581 * Improve forward-security support: add functions
3583 void SSL_CTX_set_not_resumable_session_callback(
3584 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3585 void SSL_set_not_resumable_session_callback(
3586 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3588 for use by SSL/TLS servers; the callback function will be called whenever a
3589 new session is created, and gets to decide whether the session may be
3590 cached to make it resumable (return 0) or not (return 1). (As by the
3591 SSL/TLS protocol specifications, the session_id sent by the server will be
3592 empty to indicate that the session is not resumable; also, the server will
3593 not generate RFC 4507 (RFC 5077) session tickets.)
3595 A simple reasonable callback implementation is to return is_forward_secure.
3596 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3597 by the SSL/TLS server library, indicating whether it can provide forward
3600 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3602 * New -verify_name option in command line utilities to set verification
3607 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3608 Add CMAC pkey methods.
3612 * Experimental renegotiation in s_server -www mode. If the client
3613 browses /reneg connection is renegotiated. If /renegcert it is
3614 renegotiated requesting a certificate.
3618 * Add an "external" session cache for debugging purposes to s_server. This
3619 should help trace issues which normally are only apparent in deployed
3620 multi-process servers.
3624 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3625 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3626 BIO_set_cipher() and some obscure PEM functions were changed so they
3627 can now return an error. The RAND changes required a change to the
3628 RAND_METHOD structure.
3632 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3633 a gcc attribute to warn if the result of a function is ignored. This
3634 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3635 whose return value is often ignored.
3639 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3640 These allow SCTs (signed certificate timestamps) to be requested and
3641 validated when establishing a connection.
3643 *Rob Percival <robpercival@google.com>*
3648 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3650 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3651 used even when parsing explicit parameters, when loading a serialized key
3652 or calling `EC_GROUP_new_from_ecpkparameters()`/
3653 `EC_GROUP_new_from_ecparameters()`.
3654 This prevents bypass of security hardening and performance gains,
3655 especially for curves with specialized EC_METHODs.
3656 By default, if a key encoded with explicit parameters is loaded and later
3657 serialized, the output is still encoded with explicit parameters, even if
3658 internally a "named" EC_GROUP is used for computation.
3662 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3663 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3664 NULL. After this change, only the cofactor parameter can be NULL. It also
3665 does some minimal sanity checks on the passed order.
3670 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3671 An attack is simple, if the first CMS_recipientInfo is valid but the
3672 second CMS_recipientInfo is chosen ciphertext. If the second
3673 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3674 encryption key will be replaced by garbage, and the message cannot be
3675 decoded, but if the RSA decryption fails, the correct encryption key is
3676 used and the recipient will not notice the attack.
3677 As a work around for this potential attack the length of the decrypted
3678 key must be equal to the cipher default key length, in case the
3679 certifiate is not given and all recipientInfo are tried out.
3680 The old behaviour can be re-enabled in the CMS code by setting the
3681 CMS_DEBUG_DECRYPT flag.
3686 * Document issue with installation paths in diverse Windows builds
3688 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3689 binaries and run-time config file.
3694 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3696 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3697 This changes the size when using the genpkey app when no size is given. It
3698 fixes an omission in earlier changes that changed all RSA, DSA and DH
3699 generation apps to use 2048 bits by default.
3703 * Add FIPS support for Android Arm 64-bit
3705 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3706 Module in Version 2.0.10. For some reason, the corresponding target
3707 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3708 built with FIPS support on Android Arm 64-bit. This omission has been
3711 *Matthias St. Pierre*
3713 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3715 * 0-byte record padding oracle
3717 If an application encounters a fatal protocol error and then calls
3718 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3719 then OpenSSL can respond differently to the calling application if a 0 byte
3720 record is received with invalid padding compared to if a 0 byte record is
3721 received with an invalid MAC. If the application then behaves differently
3722 based on that in a way that is detectable to the remote peer, then this
3723 amounts to a padding oracle that could be used to decrypt data.
3725 In order for this to be exploitable "non-stitched" ciphersuites must be in
3726 use. Stitched ciphersuites are optimised implementations of certain
3727 commonly used ciphersuites. Also the application must call SSL_shutdown()
3728 twice even if a protocol error has occurred (applications should not do
3729 this but some do anyway).
3731 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3732 Aviram, with additional investigation by Steven Collison and Andrew
3733 Hourselt. It was reported to OpenSSL on 10th December 2018.
3738 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3742 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3744 * Microarchitecture timing vulnerability in ECC scalar multiplication
3746 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3747 shown to be vulnerable to a microarchitecture timing side channel attack.
3748 An attacker with sufficient access to mount local timing attacks during
3749 ECDSA signature generation could recover the private key.
3751 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3752 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3758 * Timing vulnerability in DSA signature generation
3760 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3761 timing side channel attack. An attacker could use variations in the signing
3762 algorithm to recover the private key.
3764 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3769 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3770 Module, accidentally introduced while backporting security fixes from the
3771 development branch and hindering the use of ECC in FIPS mode.
3775 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3777 * Client DoS due to large DH parameter
3779 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3780 malicious server can send a very large prime value to the client. This will
3781 cause the client to spend an unreasonably long period of time generating a
3782 key for this prime resulting in a hang until the client has finished. This
3783 could be exploited in a Denial Of Service attack.
3785 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3790 * Cache timing vulnerability in RSA Key Generation
3792 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3793 a cache timing side channel attack. An attacker with sufficient access to
3794 mount cache timing attacks during the RSA key generation process could
3795 recover the private key.
3797 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3798 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3803 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3804 parameter is no longer accepted, as it leads to a corrupt table. NULL
3805 pem_str is reserved for alias entries only.
3809 * Revert blinding in ECDSA sign and instead make problematic addition
3810 length-invariant. Switch even to fixed-length Montgomery multiplication.
3814 * Change generating and checking of primes so that the error rate of not
3815 being prime depends on the intended use based on the size of the input.
3816 For larger primes this will result in more rounds of Miller-Rabin.
3817 The maximal error rate for primes with more than 1080 bits is lowered
3820 *Kurt Roeckx, Annie Yousar*
3822 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3826 * Add blinding to ECDSA and DSA signatures to protect against side channel
3827 attacks discovered by Keegan Ryan (NCC Group).
3831 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3832 now allow empty (zero character) pass phrases.
3836 * Certificate time validation (X509_cmp_time) enforces stricter
3837 compliance with RFC 5280. Fractional seconds and timezone offsets
3838 are no longer allowed.
3842 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3844 * Constructed ASN.1 types with a recursive definition could exceed the stack
3846 Constructed ASN.1 types with a recursive definition (such as can be found
3847 in PKCS7) could eventually exceed the stack given malicious input with
3848 excessive recursion. This could result in a Denial Of Service attack. There
3849 are no such structures used within SSL/TLS that come from untrusted sources
3850 so this is considered safe.
3852 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3858 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3860 * Read/write after SSL object in error state
3862 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3863 mechanism. The intent was that if a fatal error occurred during a handshake
3864 then OpenSSL would move into the error state and would immediately fail if
3865 you attempted to continue the handshake. This works as designed for the
3866 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3867 SSL_connect()), however due to a bug it does not work correctly if
3868 SSL_read() or SSL_write() is called directly. In that scenario, if the
3869 handshake fails then a fatal error will be returned in the initial function
3870 call. If SSL_read()/SSL_write() is subsequently called by the application
3871 for the same SSL object then it will succeed and the data is passed without
3872 being decrypted/encrypted directly from the SSL/TLS record layer.
3874 In order to exploit this issue an application bug would have to be present
3875 that resulted in a call to SSL_read()/SSL_write() being issued after having
3876 already received a fatal error.
3878 This issue was reported to OpenSSL by David Benjamin (Google).
3883 * rsaz_1024_mul_avx2 overflow bug on x86_64
3885 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3886 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3887 Analysis suggests that attacks against RSA and DSA as a result of this
3888 defect would be very difficult to perform and are not believed likely.
3889 Attacks against DH1024 are considered just feasible, because most of the
3890 work necessary to deduce information about a private key may be performed
3891 offline. The amount of resources required for such an attack would be
3892 significant. However, for an attack on TLS to be meaningful, the server
3893 would have to share the DH1024 private key among multiple clients, which is
3894 no longer an option since CVE-2016-0701.
3896 This only affects processors that support the AVX2 but not ADX extensions
3897 like Intel Haswell (4th generation).
3899 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3900 was originally found via the OSS-Fuzz project.
3905 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
3907 * bn_sqrx8x_internal carry bug on x86_64
3909 There is a carry propagating bug in the x86_64 Montgomery squaring
3910 procedure. No EC algorithms are affected. Analysis suggests that attacks
3911 against RSA and DSA as a result of this defect would be very difficult to
3912 perform and are not believed likely. Attacks against DH are considered just
3913 feasible (although very difficult) because most of the work necessary to
3914 deduce information about a private key may be performed offline. The amount
3915 of resources required for such an attack would be very significant and
3916 likely only accessible to a limited number of attackers. An attacker would
3917 additionally need online access to an unpatched system using the target
3918 private key in a scenario with persistent DH parameters and a private
3919 key that is shared between multiple clients.
3921 This only affects processors that support the BMI1, BMI2 and ADX extensions
3922 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3924 This issue was reported to OpenSSL by the OSS-Fuzz project.
3929 * Malformed X.509 IPAddressFamily could cause OOB read
3931 If an X.509 certificate has a malformed IPAddressFamily extension,
3932 OpenSSL could do a one-byte buffer overread. The most likely result
3933 would be an erroneous display of the certificate in text format.
3935 This issue was reported to OpenSSL by the OSS-Fuzz project.
3940 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
3942 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3943 platform rather than 'mingw'.
3947 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
3949 * Truncated packet could crash via OOB read
3951 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3952 cipher is being used, then a truncated packet can cause that host to
3953 perform an out-of-bounds read, usually resulting in a crash.
3955 This issue was reported to OpenSSL by Robert Święcki of Google.
3960 * BN_mod_exp may produce incorrect results on x86_64
3962 There is a carry propagating bug in the x86_64 Montgomery squaring
3963 procedure. No EC algorithms are affected. Analysis suggests that attacks
3964 against RSA and DSA as a result of this defect would be very difficult to
3965 perform and are not believed likely. Attacks against DH are considered just
3966 feasible (although very difficult) because most of the work necessary to
3967 deduce information about a private key may be performed offline. The amount
3968 of resources required for such an attack would be very significant and
3969 likely only accessible to a limited number of attackers. An attacker would
3970 additionally need online access to an unpatched system using the target
3971 private key in a scenario with persistent DH parameters and a private
3972 key that is shared between multiple clients. For example this can occur by
3973 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3974 similar to CVE-2015-3193 but must be treated as a separate problem.
3976 This issue was reported to OpenSSL by the OSS-Fuzz project.
3981 * Montgomery multiplication may produce incorrect results
3983 There is a carry propagating bug in the Broadwell-specific Montgomery
3984 multiplication procedure that handles input lengths divisible by, but
3985 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3986 and DH private keys are impossible. This is because the subroutine in
3987 question is not used in operations with the private key itself and an input
3988 of the attacker's direct choice. Otherwise the bug can manifest itself as
3989 transient authentication and key negotiation failures or reproducible
3990 erroneous outcome of public-key operations with specially crafted input.
3991 Among EC algorithms only Brainpool P-512 curves are affected and one
3992 presumably can attack ECDH key negotiation. Impact was not analyzed in
3993 detail, because pre-requisites for attack are considered unlikely. Namely
3994 multiple clients have to choose the curve in question and the server has to
3995 share the private key among them, neither of which is default behaviour.
3996 Even then only clients that chose the curve will be affected.
3998 This issue was publicly reported as transient failures and was not
3999 initially recognized as a security issue. Thanks to Richard Morgan for
4000 providing reproducible case.
4005 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4006 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4007 prevent issues where no progress is being made and the peer continually
4008 sends unrecognised record types, using up resources processing them.
4012 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4014 * Missing CRL sanity check
4016 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4017 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4018 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4020 This issue only affects the OpenSSL 1.0.2i
4025 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4027 * OCSP Status Request extension unbounded memory growth
4029 A malicious client can send an excessively large OCSP Status Request
4030 extension. If that client continually requests renegotiation, sending a
4031 large OCSP Status Request extension each time, then there will be unbounded
4032 memory growth on the server. This will eventually lead to a Denial Of
4033 Service attack through memory exhaustion. Servers with a default
4034 configuration are vulnerable even if they do not support OCSP. Builds using
4035 the "no-ocsp" build time option are not affected.
4037 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4042 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4045 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4051 * OOB write in MDC2_Update()
4053 An overflow can occur in MDC2_Update() either if called directly or
4054 through the EVP_DigestUpdate() function using MDC2. If an attacker
4055 is able to supply very large amounts of input data after a previous
4056 call to EVP_EncryptUpdate() with a partial block then a length check
4057 can overflow resulting in a heap corruption.
4059 The amount of data needed is comparable to SIZE_MAX which is impractical
4062 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4067 * Malformed SHA512 ticket DoS
4069 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4070 DoS attack where a malformed ticket will result in an OOB read which will
4073 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4074 a custom server callback and ticket lookup mechanism.
4076 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4081 * OOB write in BN_bn2dec()
4083 The function BN_bn2dec() does not check the return value of BN_div_word().
4084 This can cause an OOB write if an application uses this function with an
4085 overly large BIGNUM. This could be a problem if an overly large certificate
4086 or CRL is printed out from an untrusted source. TLS is not affected because
4087 record limits will reject an oversized certificate before it is parsed.
4089 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4094 * OOB read in TS_OBJ_print_bio()
4096 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4097 the total length the OID text representation would use and not the amount
4098 of data written. This will result in OOB reads when large OIDs are
4101 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4106 * Pointer arithmetic undefined behaviour
4108 Avoid some undefined pointer arithmetic
4110 A common idiom in the codebase is to check limits in the following manner:
4113 Where "p" points to some malloc'd data of SIZE bytes and
4116 "len" here could be from some externally supplied data (e.g. from a TLS
4119 The rules of C pointer arithmetic are such that "p + len" is only well
4120 defined where len <= SIZE. Therefore the above idiom is actually
4121 undefined behaviour.
4123 For example this could cause problems if some malloc implementation
4124 provides an address for "p" such that "p + len" actually overflows for
4125 values of len that are too big and therefore p + len < limit.
4127 This issue was reported to OpenSSL by Guido Vranken
4132 * Constant time flag not preserved in DSA signing
4134 Operations in the DSA signing algorithm should run in constant time in
4135 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4136 implementation means that a non-constant time codepath is followed for
4137 certain operations. This has been demonstrated through a cache-timing
4138 attack to be sufficient for an attacker to recover the private DSA key.
4140 This issue was reported by César Pereida (Aalto University), Billy Brumley
4141 (Tampere University of Technology), and Yuval Yarom (The University of
4142 Adelaide and NICTA).
4147 * DTLS buffered message DoS
4149 In a DTLS connection where handshake messages are delivered out-of-order
4150 those messages that OpenSSL is not yet ready to process will be buffered
4151 for later use. Under certain circumstances, a flaw in the logic means that
4152 those messages do not get removed from the buffer even though the handshake
4153 has been completed. An attacker could force up to approx. 15 messages to
4154 remain in the buffer when they are no longer required. These messages will
4155 be cleared when the DTLS connection is closed. The default maximum size for
4156 a message is 100k. Therefore the attacker could force an additional 1500k
4157 to be consumed per connection. By opening many simulataneous connections an
4158 attacker could cause a DoS attack through memory exhaustion.
4160 This issue was reported to OpenSSL by Quan Luo.
4165 * DTLS replay protection DoS
4167 A flaw in the DTLS replay attack protection mechanism means that records
4168 that arrive for future epochs update the replay protection "window" before
4169 the MAC for the record has been validated. This could be exploited by an
4170 attacker by sending a record for the next epoch (which does not have to
4171 decrypt or have a valid MAC), with a very large sequence number. This means
4172 that all subsequent legitimate packets are dropped causing a denial of
4173 service for a specific DTLS connection.
4175 This issue was reported to OpenSSL by the OCAP audit team.
4180 * Certificate message OOB reads
4182 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4183 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4184 theoretical DoS risk but this has not been observed in practice on common
4187 The messages affected are client certificate, client certificate request
4188 and server certificate. As a result the attack can only be performed
4189 against a client or a server which enables client authentication.
4191 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4196 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4198 * Prevent padding oracle in AES-NI CBC MAC check
4200 A MITM attacker can use a padding oracle attack to decrypt traffic
4201 when the connection uses an AES CBC cipher and the server support
4204 This issue was introduced as part of the fix for Lucky 13 padding
4205 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4206 constant time by making sure that always the same bytes are read and
4207 compared against either the MAC or padding bytes. But it no longer
4208 checked that there was enough data to have both the MAC and padding
4211 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4216 * Fix EVP_EncodeUpdate overflow
4218 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4219 Base64 encoding of binary data. If an attacker is able to supply very large
4220 amounts of input data then a length check can overflow resulting in a heap
4223 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4224 the `PEM_write_bio*` family of functions. These are mainly used within the
4225 OpenSSL command line applications, so any application which processes data
4226 from an untrusted source and outputs it as a PEM file should be considered
4227 vulnerable to this issue. User applications that call these APIs directly
4228 with large amounts of untrusted data may also be vulnerable.
4230 This issue was reported by Guido Vranken.
4235 * Fix EVP_EncryptUpdate overflow
4237 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4238 is able to supply very large amounts of input data after a previous call to
4239 EVP_EncryptUpdate() with a partial block then a length check can overflow
4240 resulting in a heap corruption. Following an analysis of all OpenSSL
4241 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4242 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4243 the first called function after an EVP_EncryptInit(), and therefore that
4244 specific call must be safe. The second form is where the length passed to
4245 EVP_EncryptUpdate() can be seen from the code to be some small value and
4246 therefore there is no possibility of an overflow. Since all instances are
4247 one of these two forms, it is believed that there can be no overflows in
4248 internal code due to this problem. It should be noted that
4249 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4250 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4251 of these calls have also been analysed too and it is believed there are no
4252 instances in internal usage where an overflow could occur.
4254 This issue was reported by Guido Vranken.
4259 * Prevent ASN.1 BIO excessive memory allocation
4261 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4262 a short invalid encoding can cause allocation of large amounts of memory
4263 potentially consuming excessive resources or exhausting memory.
4265 Any application parsing untrusted data through d2i BIO functions is
4266 affected. The memory based functions such as d2i_X509() are *not* affected.
4267 Since the memory based functions are used by the TLS library, TLS
4268 applications are not affected.
4270 This issue was reported by Brian Carpenter.
4277 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4278 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4279 in arbitrary stack data being returned in the buffer.
4281 This issue was reported by Guido Vranken.
4286 * Modify behavior of ALPN to invoke callback after SNI/servername
4287 callback, such that updates to the SSL_CTX affect ALPN.
4291 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4296 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4297 methods are enabled and ssl2 is disabled the methods return NULL.
4301 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4303 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4304 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4305 provide any "EXPORT" or "LOW" strength ciphers.
4309 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4310 is by default disabled at build-time. Builds that are not configured with
4311 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4312 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4313 will need to explicitly call either of:
4315 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4317 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4319 as appropriate. Even if either of those is used, or the application
4320 explicitly uses the version-specific SSLv2_method() or its client and
4321 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4322 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4323 ciphers, and SSLv2 56-bit DES are no longer available.
4328 * Fix a double-free in DSA code
4330 A double free bug was discovered when OpenSSL parses malformed DSA private
4331 keys and could lead to a DoS attack or memory corruption for applications
4332 that receive DSA private keys from untrusted sources. This scenario is
4335 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4341 * Disable SRP fake user seed to address a server memory leak.
4343 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4345 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4346 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4347 was changed to ignore the "fake user" SRP seed, even if the seed
4350 Users should use SRP_VBASE_get1_by_user instead. Note that in
4351 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4352 also that even though configuring the SRP seed attempts to hide
4353 invalid usernames by continuing the handshake with fake
4354 credentials, this behaviour is not constant time and no strong
4355 guarantees are made that the handshake is indistinguishable from
4356 that of a valid user.
4361 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4363 In the BN_hex2bn function the number of hex digits is calculated using an
4364 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4365 large values of |i| this can result in |bn_expand| not allocating any
4366 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4367 field as NULL leading to a subsequent NULL ptr deref. For very large values
4368 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4369 In this case memory is allocated to the internal BIGNUM data field, but it
4370 is insufficiently sized leading to heap corruption. A similar issue exists
4371 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4372 is ever called by user applications with very large untrusted hex/dec data.
4373 This is anticipated to be a rare occurrence.
4375 All OpenSSL internal usage of these functions use data that is not expected
4376 to be untrusted, e.g. config file data or application command line
4377 arguments. If user developed applications generate config file data based
4378 on untrusted data then it is possible that this could also lead to security
4379 consequences. This is also anticipated to be rare.
4381 This issue was reported to OpenSSL by Guido Vranken.
4386 * Fix memory issues in `BIO_*printf` functions
4388 The internal |fmtstr| function used in processing a "%s" format string in
4389 the `BIO_*printf` functions could overflow while calculating the length of a
4390 string and cause an OOB read when printing very long strings.
4392 Additionally the internal |doapr_outch| function can attempt to write to an
4393 OOB memory location (at an offset from the NULL pointer) in the event of a
4394 memory allocation failure. In 1.0.2 and below this could be caused where
4395 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4396 could be in processing a very long "%s" format string. Memory leaks can
4399 The first issue may mask the second issue dependent on compiler behaviour.
4400 These problems could enable attacks where large amounts of untrusted data
4401 is passed to the `BIO_*printf` functions. If applications use these functions
4402 in this way then they could be vulnerable. OpenSSL itself uses these
4403 functions when printing out human-readable dumps of ASN.1 data. Therefore
4404 applications that print this data could be vulnerable if the data is from
4405 untrusted sources. OpenSSL command line applications could also be
4406 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4407 as command line arguments.
4409 Libssl is not considered directly vulnerable. Additionally certificates etc
4410 received via remote connections via libssl are also unlikely to be able to
4411 trigger these issues because of message size limits enforced within libssl.
4413 This issue was reported to OpenSSL Guido Vranken.
4418 * Side channel attack on modular exponentiation
4420 A side-channel attack was found which makes use of cache-bank conflicts on
4421 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4422 of RSA keys. The ability to exploit this issue is limited as it relies on
4423 an attacker who has control of code in a thread running on the same
4424 hyper-threaded core as the victim thread which is performing decryptions.
4426 This issue was reported to OpenSSL by Yuval Yarom, The University of
4427 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4428 Nadia Heninger, University of Pennsylvania with more information at
4429 <http://cachebleed.info>.
4434 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4435 if no keysize is specified with default_bits. This fixes an
4436 omission in an earlier change that changed all RSA/DSA key generation
4437 apps to use 2048 bits by default.
4441 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4443 * DH small subgroups
4445 Historically OpenSSL only ever generated DH parameters based on "safe"
4446 primes. More recently (in version 1.0.2) support was provided for
4447 generating X9.42 style parameter files such as those required for RFC 5114
4448 support. The primes used in such files may not be "safe". Where an
4449 application is using DH configured with parameters based on primes that are
4450 not "safe" then an attacker could use this fact to find a peer's private
4451 DH exponent. This attack requires that the attacker complete multiple
4452 handshakes in which the peer uses the same private DH exponent. For example
4453 this could be used to discover a TLS server's private DH exponent if it's
4454 reusing the private DH exponent or it's using a static DH ciphersuite.
4456 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4457 TLS. It is not on by default. If the option is not set then the server
4458 reuses the same private DH exponent for the life of the server process and
4459 would be vulnerable to this attack. It is believed that many popular
4460 applications do set this option and would therefore not be at risk.
4462 The fix for this issue adds an additional check where a "q" parameter is
4463 available (as is the case in X9.42 based parameters). This detects the
4464 only known attack, and is the only possible defense for static DH
4465 ciphersuites. This could have some performance impact.
4467 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4468 default and cannot be disabled. This could have some performance impact.
4470 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4475 * SSLv2 doesn't block disabled ciphers
4477 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4478 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4479 been disabled, provided that the SSLv2 protocol was not also disabled via
4482 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4483 and Sebastian Schinzel.
4488 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4490 * BN_mod_exp may produce incorrect results on x86_64
4492 There is a carry propagating bug in the x86_64 Montgomery squaring
4493 procedure. No EC algorithms are affected. Analysis suggests that attacks
4494 against RSA and DSA as a result of this defect would be very difficult to
4495 perform and are not believed likely. Attacks against DH are considered just
4496 feasible (although very difficult) because most of the work necessary to
4497 deduce information about a private key may be performed offline. The amount
4498 of resources required for such an attack would be very significant and
4499 likely only accessible to a limited number of attackers. An attacker would
4500 additionally need online access to an unpatched system using the target
4501 private key in a scenario with persistent DH parameters and a private
4502 key that is shared between multiple clients. For example this can occur by
4503 default in OpenSSL DHE based SSL/TLS ciphersuites.
4505 This issue was reported to OpenSSL by Hanno Böck.
4510 * Certificate verify crash with missing PSS parameter
4512 The signature verification routines will crash with a NULL pointer
4513 dereference if presented with an ASN.1 signature using the RSA PSS
4514 algorithm and absent mask generation function parameter. Since these
4515 routines are used to verify certificate signature algorithms this can be
4516 used to crash any certificate verification operation and exploited in a
4517 DoS attack. Any application which performs certificate verification is
4518 vulnerable including OpenSSL clients and servers which enable client
4521 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4526 * X509_ATTRIBUTE memory leak
4528 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4529 memory. This structure is used by the PKCS#7 and CMS routines so any
4530 application which reads PKCS#7 or CMS data from untrusted sources is
4531 affected. SSL/TLS is not affected.
4533 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4539 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4540 This changes the decoding behaviour for some invalid messages,
4541 though the change is mostly in the more lenient direction, and
4542 legacy behaviour is preserved as much as possible.
4546 * In DSA_generate_parameters_ex, if the provided seed is too short,
4549 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4551 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4553 * Alternate chains certificate forgery
4555 During certificate verification, OpenSSL will attempt to find an
4556 alternative certificate chain if the first attempt to build such a chain
4557 fails. An error in the implementation of this logic can mean that an
4558 attacker could cause certain checks on untrusted certificates to be
4559 bypassed, such as the CA flag, enabling them to use a valid leaf
4560 certificate to act as a CA and "issue" an invalid certificate.
4562 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4567 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4569 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4570 incompatibility in the handling of HMAC. The previous ABI has now been
4575 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4577 * Malformed ECParameters causes infinite loop
4579 When processing an ECParameters structure OpenSSL enters an infinite loop
4580 if the curve specified is over a specially malformed binary polynomial
4583 This can be used to perform denial of service against any
4584 system which processes public keys, certificate requests or
4585 certificates. This includes TLS clients and TLS servers with
4586 client authentication enabled.
4588 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4593 * Exploitable out-of-bounds read in X509_cmp_time
4595 X509_cmp_time does not properly check the length of the ASN1_TIME
4596 string and can read a few bytes out of bounds. In addition,
4597 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4600 An attacker can use this to craft malformed certificates and CRLs of
4601 various sizes and potentially cause a segmentation fault, resulting in
4602 a DoS on applications that verify certificates or CRLs. TLS clients
4603 that verify CRLs are affected. TLS clients and servers with client
4604 authentication enabled may be affected if they use custom verification
4607 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4608 independently by Hanno Böck.
4613 * PKCS7 crash with missing EnvelopedContent
4615 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4616 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4617 with missing content and trigger a NULL pointer dereference on parsing.
4619 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4620 structures from untrusted sources are affected. OpenSSL clients and
4621 servers are not affected.
4623 This issue was reported to OpenSSL by Michal Zalewski (Google).
4628 * CMS verify infinite loop with unknown hash function
4630 When verifying a signedData message the CMS code can enter an infinite loop
4631 if presented with an unknown hash function OID. This can be used to perform
4632 denial of service against any system which verifies signedData messages using
4634 This issue was reported to OpenSSL by Johannes Bauer.
4639 * Race condition handling NewSessionTicket
4641 If a NewSessionTicket is received by a multi-threaded client when attempting to
4642 reuse a previous ticket then a race condition can occur potentially leading to
4643 a double free of the ticket data.
4648 * Only support 256-bit or stronger elliptic curves with the
4649 'ecdh_auto' setting (server) or by default (client). Of supported
4650 curves, prefer P-256 (both).
4654 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4656 * ClientHello sigalgs DoS fix
4658 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4659 invalid signature algorithms extension a NULL pointer dereference will
4660 occur. This can be exploited in a DoS attack against the server.
4662 This issue was was reported to OpenSSL by David Ramos of Stanford
4666 *Stephen Henson and Matt Caswell*
4668 * Multiblock corrupted pointer fix
4670 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4671 feature only applies on 64 bit x86 architecture platforms that support AES
4672 NI instructions. A defect in the implementation of "multiblock" can cause
4673 OpenSSL's internal write buffer to become incorrectly set to NULL when
4674 using non-blocking IO. Typically, when the user application is using a
4675 socket BIO for writing, this will only result in a failed connection.
4676 However if some other BIO is used then it is likely that a segmentation
4677 fault will be triggered, thus enabling a potential DoS attack.
4679 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4684 * Segmentation fault in DTLSv1_listen fix
4686 The DTLSv1_listen function is intended to be stateless and processes the
4687 initial ClientHello from many peers. It is common for user code to loop
4688 over the call to DTLSv1_listen until a valid ClientHello is received with
4689 an associated cookie. A defect in the implementation of DTLSv1_listen means
4690 that state is preserved in the SSL object from one invocation to the next
4691 that can lead to a segmentation fault. Errors processing the initial
4692 ClientHello can trigger this scenario. An example of such an error could be
4693 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4696 This issue was reported to OpenSSL by Per Allansson.
4701 * Segmentation fault in ASN1_TYPE_cmp fix
4703 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4704 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4705 certificate signature algorithm consistency this can be used to crash any
4706 certificate verification operation and exploited in a DoS attack. Any
4707 application which performs certificate verification is vulnerable including
4708 OpenSSL clients and servers which enable client authentication.
4713 * Segmentation fault for invalid PSS parameters fix
4715 The signature verification routines will crash with a NULL pointer
4716 dereference if presented with an ASN.1 signature using the RSA PSS
4717 algorithm and invalid parameters. Since these routines are used to verify
4718 certificate signature algorithms this can be used to crash any
4719 certificate verification operation and exploited in a DoS attack. Any
4720 application which performs certificate verification is vulnerable including
4721 OpenSSL clients and servers which enable client authentication.
4723 This issue was was reported to OpenSSL by Brian Carpenter.
4728 * ASN.1 structure reuse memory corruption fix
4730 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4731 memory corruption via an invalid write. Such reuse is and has been
4732 strongly discouraged and is believed to be rare.
4734 Applications that parse structures containing CHOICE or ANY DEFINED BY
4735 components may be affected. Certificate parsing (d2i_X509 and related
4736 functions) are however not affected. OpenSSL clients and servers are
4742 * PKCS7 NULL pointer dereferences fix
4744 The PKCS#7 parsing code does not handle missing outer ContentInfo
4745 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4746 missing content and trigger a NULL pointer dereference on parsing.
4748 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4749 otherwise parse PKCS#7 structures from untrusted sources are
4750 affected. OpenSSL clients and servers are not affected.
4752 This issue was reported to OpenSSL by Michal Zalewski (Google).
4757 * DoS via reachable assert in SSLv2 servers fix
4759 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4760 servers that both support SSLv2 and enable export cipher suites by sending
4761 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4763 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4764 (OpenSSL development team).
4769 * Empty CKE with client auth and DHE fix
4771 If client auth is used then a server can seg fault in the event of a DHE
4772 ciphersuite being selected and a zero length ClientKeyExchange message
4773 being sent by the client. This could be exploited in a DoS attack.
4778 * Handshake with unseeded PRNG fix
4780 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4781 with an unseeded PRNG. The conditions are:
4782 - The client is on a platform where the PRNG has not been seeded
4783 automatically, and the user has not seeded manually
4784 - A protocol specific client method version has been used (i.e. not
4785 SSL_client_methodv23)
4786 - A ciphersuite is used that does not require additional random data from
4787 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4789 If the handshake succeeds then the client random that has been used will
4790 have been generated from a PRNG with insufficient entropy and therefore the
4791 output may be predictable.
4793 For example using the following command with an unseeded openssl will
4794 succeed on an unpatched platform:
4796 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4801 * Use After Free following d2i_ECPrivatekey error fix
4803 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4804 could cause a use after free condition. This, in turn, could cause a double
4805 free in several private key parsing functions (such as d2i_PrivateKey
4806 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4807 for applications that receive EC private keys from untrusted
4808 sources. This scenario is considered rare.
4810 This issue was discovered by the BoringSSL project and fixed in their
4816 * X509_to_X509_REQ NULL pointer deref fix
4818 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4819 the certificate key is invalid. This function is rarely used in practice.
4821 This issue was discovered by Brian Carpenter.
4826 * Removed the export ciphers from the DEFAULT ciphers
4830 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4832 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4833 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4834 So far those who have to target multiple platforms would compromise
4835 and argue that binary targeting say ARMv5 would still execute on
4836 ARMv8. "Universal" build resolves this compromise by providing
4837 near-optimal performance even on newer platforms.
4841 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4842 (other platforms pending).
4844 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4846 * Add support for the SignedCertificateTimestampList certificate and
4847 OCSP response extensions from RFC6962.
4851 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4852 for corner cases. (Certain input points at infinity could lead to
4853 bogus results, with non-infinity inputs mapped to infinity too.)
4857 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4858 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4859 common cases are optimized and there still is room for further
4860 improvements. Vector Permutation AES for Altivec is also added.
4864 * Add support for little-endian ppc64 Linux target.
4866 *Marcelo Cerri (IBM)*
4868 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4869 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4870 are optimized and there still is room for further improvements.
4871 Both 32- and 64-bit modes are supported.
4873 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4875 * Improved ARMv7 NEON support.
4879 * Support for SPARC Architecture 2011 crypto extensions, first
4880 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4881 SHA256/512, MD5, GHASH and modular exponentiation.
4883 *Andy Polyakov, David Miller*
4885 * Accelerated modular exponentiation for Intel processors, a.k.a.
4888 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4890 * Support for new and upcoming Intel processors, including AVX2,
4891 BMI and SHA ISA extensions. This includes additional "stitched"
4892 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4895 This work was sponsored by Intel Corp.
4899 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4900 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4901 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4905 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4906 this fixes a limitation in previous versions of OpenSSL.
4910 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4911 MGF1 digest and OAEP label.
4915 * Add EVP support for key wrapping algorithms, to avoid problems with
4916 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4917 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4918 algorithms and include tests cases.
4922 * Add functions to allocate and set the fields of an ECDSA_METHOD
4925 *Douglas E. Engert, Steve Henson*
4927 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4928 difference in days and seconds between two tm or ASN1_TIME structures.
4932 * Add -rev test option to s_server to just reverse order of characters
4933 received by client and send back to server. Also prints an abbreviated
4934 summary of the connection parameters.
4938 * New option -brief for s_client and s_server to print out a brief summary
4939 of connection parameters.
4943 * Add callbacks for arbitrary TLS extensions.
4945 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4947 * New option -crl_download in several openssl utilities to download CRLs
4948 from CRLDP extension in certificates.
4952 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4956 * New function X509_CRL_diff to generate a delta CRL from the difference
4957 of two full CRLs. Add support to "crl" utility.
4961 * New functions to set lookup_crls function and to retrieve
4962 X509_STORE from X509_STORE_CTX.
4966 * Print out deprecated issuer and subject unique ID fields in
4971 * Extend OCSP I/O functions so they can be used for simple general purpose
4972 HTTP as well as OCSP. New wrapper function which can be used to download
4973 CRLs using the OCSP API.
4977 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4981 * `SSL_CONF*` functions. These provide a common framework for application
4982 configuration using configuration files or command lines.
4986 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4987 message callback and prints the results. Needs compile time option
4988 "enable-ssl-trace". New options to s_client and s_server to enable
4993 * New ctrl and macro to retrieve supported points extensions.
4994 Print out extension in s_server and s_client.
4998 * New functions to retrieve certificate signature and signature
5003 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5008 * New Suite B modes for TLS code. These use and enforce the requirements
5009 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5010 only use Suite B curves. The Suite B modes can be set by using the
5011 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5015 * New chain verification flags for Suite B levels of security. Check
5016 algorithms are acceptable when flags are set in X509_verify_cert.
5020 * Make tls1_check_chain return a set of flags indicating checks passed
5021 by a certificate chain. Add additional tests to handle client
5022 certificates: checks for matching certificate type and issuer name
5027 * If an attempt is made to use a signature algorithm not in the peer
5028 preference list abort the handshake. If client has no suitable
5029 signature algorithms in response to a certificate request do not
5030 use the certificate.
5034 * If server EC tmp key is not in client preference list abort handshake.
5038 * Add support for certificate stores in CERT structure. This makes it
5039 possible to have different stores per SSL structure or one store in
5040 the parent SSL_CTX. Include distinct stores for certificate chain
5041 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5042 to build and store a certificate chain in CERT structure: returning
5043 an error if the chain cannot be built: this will allow applications
5044 to test if a chain is correctly configured.
5046 Note: if the CERT based stores are not set then the parent SSL_CTX
5047 store is used to retain compatibility with existing behaviour.
5051 * New function ssl_set_client_disabled to set a ciphersuite disabled
5052 mask based on the current session, check mask when sending client
5053 hello and checking the requested ciphersuite.
5057 * New ctrls to retrieve and set certificate types in a certificate
5058 request message. Print out received values in s_client. If certificate
5059 types is not set with custom values set sensible values based on
5060 supported signature algorithms.
5064 * Support for distinct client and server supported signature algorithms.
5068 * Add certificate callback. If set this is called whenever a certificate
5069 is required by client or server. An application can decide which
5070 certificate chain to present based on arbitrary criteria: for example
5071 supported signature algorithms. Add very simple example to s_server.
5072 This fixes many of the problems and restrictions of the existing client
5073 certificate callback: for example you can now clear an existing
5074 certificate and specify the whole chain.
5078 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5079 the certificate can be used for (if anything). Set valid_flags field
5080 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5081 to have similar checks in it.
5083 Add new "cert_flags" field to CERT structure and include a "strict mode".
5084 This enforces some TLS certificate requirements (such as only permitting
5085 certificate signature algorithms contained in the supported algorithms
5086 extension) which some implementations ignore: this option should be used
5087 with caution as it could cause interoperability issues.
5091 * Update and tidy signature algorithm extension processing. Work out
5092 shared signature algorithms based on preferences and peer algorithms
5093 and print them out in s_client and s_server. Abort handshake if no
5094 shared signature algorithms.
5098 * Add new functions to allow customised supported signature algorithms
5099 for SSL and SSL_CTX structures. Add options to s_client and s_server
5104 * New function SSL_certs_clear() to delete all references to certificates
5105 from an SSL structure. Before this once a certificate had been added
5106 it couldn't be removed.
5110 * Integrate hostname, email address and IP address checking with certificate
5111 verification. New verify options supporting checking in openssl utility.
5115 * Fixes and wildcard matching support to hostname and email checking
5116 functions. Add manual page.
5118 *Florian Weimer (Red Hat Product Security Team)*
5120 * New functions to check a hostname email or IP address against a
5121 certificate. Add options x509 utility to print results of checks against
5126 * Fix OCSP checking.
5128 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5130 * Initial experimental support for explicitly trusted non-root CAs.
5131 OpenSSL still tries to build a complete chain to a root but if an
5132 intermediate CA has a trust setting included that is used. The first
5133 setting is used: whether to trust (e.g., -addtrust option to the x509
5138 * Add -trusted_first option which attempts to find certificates in the
5139 trusted store even if an untrusted chain is also supplied.
5143 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5144 platform support for Linux and Android.
5148 * Support for linux-x32, ILP32 environment in x86_64 framework.
5152 * Experimental multi-implementation support for FIPS capable OpenSSL.
5153 When in FIPS mode the approved implementations are used as normal,
5154 when not in FIPS mode the internal unapproved versions are used instead.
5155 This means that the FIPS capable OpenSSL isn't forced to use the
5156 (often lower performance) FIPS implementations outside FIPS mode.
5160 * Transparently support X9.42 DH parameters when calling
5161 PEM_read_bio_DHparameters. This means existing applications can handle
5162 the new parameter format automatically.
5166 * Initial experimental support for X9.42 DH parameter format: mainly
5167 to support use of 'q' parameter for RFC5114 parameters.
5171 * Add DH parameters from RFC5114 including test data to dhtest.
5175 * Support for automatic EC temporary key parameter selection. If enabled
5176 the most preferred EC parameters are automatically used instead of
5177 hardcoded fixed parameters. Now a server just has to call:
5178 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5179 support ECDH and use the most appropriate parameters.
5183 * Enhance and tidy EC curve and point format TLS extension code. Use
5184 static structures instead of allocation if default values are used.
5185 New ctrls to set curves we wish to support and to retrieve shared curves.
5186 Print out shared curves in s_server. New options to s_server and s_client
5187 to set list of supported curves.
5191 * New ctrls to retrieve supported signature algorithms and
5192 supported curve values as an array of NIDs. Extend openssl utility
5193 to print out received values.
5197 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5198 between NIDs and the more common NIST names such as "P-256". Enhance
5199 ecparam utility and ECC method to recognise the NIST names for curves.
5203 * Enhance SSL/TLS certificate chain handling to support different
5204 chains for each certificate instead of one chain in the parent SSL_CTX.
5208 * Support for fixed DH ciphersuite client authentication: where both
5209 server and client use DH certificates with common parameters.
5213 * Support for fixed DH ciphersuites: those requiring DH server
5218 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5220 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5221 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5222 X509_CINF_get_signature were reverted post internal team review.
5227 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5229 * OCSP Status Request extension unbounded memory growth
5231 A malicious client can send an excessively large OCSP Status Request
5232 extension. If that client continually requests renegotiation, sending a
5233 large OCSP Status Request extension each time, then there will be unbounded
5234 memory growth on the server. This will eventually lead to a Denial Of
5235 Service attack through memory exhaustion. Servers with a default
5236 configuration are vulnerable even if they do not support OCSP. Builds using
5237 the "no-ocsp" build time option are not affected.
5239 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5244 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5247 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5253 * OOB write in MDC2_Update()
5255 An overflow can occur in MDC2_Update() either if called directly or
5256 through the EVP_DigestUpdate() function using MDC2. If an attacker
5257 is able to supply very large amounts of input data after a previous
5258 call to EVP_EncryptUpdate() with a partial block then a length check
5259 can overflow resulting in a heap corruption.
5261 The amount of data needed is comparable to SIZE_MAX which is impractical
5264 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5269 * Malformed SHA512 ticket DoS
5271 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5272 DoS attack where a malformed ticket will result in an OOB read which will
5275 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5276 a custom server callback and ticket lookup mechanism.
5278 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5283 * OOB write in BN_bn2dec()
5285 The function BN_bn2dec() does not check the return value of BN_div_word().
5286 This can cause an OOB write if an application uses this function with an
5287 overly large BIGNUM. This could be a problem if an overly large certificate
5288 or CRL is printed out from an untrusted source. TLS is not affected because
5289 record limits will reject an oversized certificate before it is parsed.
5291 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5296 * OOB read in TS_OBJ_print_bio()
5298 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5299 the total length the OID text representation would use and not the amount
5300 of data written. This will result in OOB reads when large OIDs are
5303 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5308 * Pointer arithmetic undefined behaviour
5310 Avoid some undefined pointer arithmetic
5312 A common idiom in the codebase is to check limits in the following manner:
5315 Where "p" points to some malloc'd data of SIZE bytes and
5318 "len" here could be from some externally supplied data (e.g. from a TLS
5321 The rules of C pointer arithmetic are such that "p + len" is only well
5322 defined where len <= SIZE. Therefore the above idiom is actually
5323 undefined behaviour.
5325 For example this could cause problems if some malloc implementation
5326 provides an address for "p" such that "p + len" actually overflows for
5327 values of len that are too big and therefore p + len < limit.
5329 This issue was reported to OpenSSL by Guido Vranken
5334 * Constant time flag not preserved in DSA signing
5336 Operations in the DSA signing algorithm should run in constant time in
5337 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5338 implementation means that a non-constant time codepath is followed for
5339 certain operations. This has been demonstrated through a cache-timing
5340 attack to be sufficient for an attacker to recover the private DSA key.
5342 This issue was reported by César Pereida (Aalto University), Billy Brumley
5343 (Tampere University of Technology), and Yuval Yarom (The University of
5344 Adelaide and NICTA).
5349 * DTLS buffered message DoS
5351 In a DTLS connection where handshake messages are delivered out-of-order
5352 those messages that OpenSSL is not yet ready to process will be buffered
5353 for later use. Under certain circumstances, a flaw in the logic means that
5354 those messages do not get removed from the buffer even though the handshake
5355 has been completed. An attacker could force up to approx. 15 messages to
5356 remain in the buffer when they are no longer required. These messages will
5357 be cleared when the DTLS connection is closed. The default maximum size for
5358 a message is 100k. Therefore the attacker could force an additional 1500k
5359 to be consumed per connection. By opening many simulataneous connections an
5360 attacker could cause a DoS attack through memory exhaustion.
5362 This issue was reported to OpenSSL by Quan Luo.
5367 * DTLS replay protection DoS
5369 A flaw in the DTLS replay attack protection mechanism means that records
5370 that arrive for future epochs update the replay protection "window" before
5371 the MAC for the record has been validated. This could be exploited by an
5372 attacker by sending a record for the next epoch (which does not have to
5373 decrypt or have a valid MAC), with a very large sequence number. This means
5374 that all subsequent legitimate packets are dropped causing a denial of
5375 service for a specific DTLS connection.
5377 This issue was reported to OpenSSL by the OCAP audit team.
5382 * Certificate message OOB reads
5384 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5385 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5386 theoretical DoS risk but this has not been observed in practice on common
5389 The messages affected are client certificate, client certificate request
5390 and server certificate. As a result the attack can only be performed
5391 against a client or a server which enables client authentication.
5393 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5398 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5400 * Prevent padding oracle in AES-NI CBC MAC check
5402 A MITM attacker can use a padding oracle attack to decrypt traffic
5403 when the connection uses an AES CBC cipher and the server support
5406 This issue was introduced as part of the fix for Lucky 13 padding
5407 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5408 constant time by making sure that always the same bytes are read and
5409 compared against either the MAC or padding bytes. But it no longer
5410 checked that there was enough data to have both the MAC and padding
5413 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5418 * Fix EVP_EncodeUpdate overflow
5420 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5421 Base64 encoding of binary data. If an attacker is able to supply very large
5422 amounts of input data then a length check can overflow resulting in a heap
5425 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5426 the `PEM_write_bio*` family of functions. These are mainly used within the
5427 OpenSSL command line applications, so any application which processes data
5428 from an untrusted source and outputs it as a PEM file should be considered
5429 vulnerable to this issue. User applications that call these APIs directly
5430 with large amounts of untrusted data may also be vulnerable.
5432 This issue was reported by Guido Vranken.
5437 * Fix EVP_EncryptUpdate overflow
5439 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5440 is able to supply very large amounts of input data after a previous call to
5441 EVP_EncryptUpdate() with a partial block then a length check can overflow
5442 resulting in a heap corruption. Following an analysis of all OpenSSL
5443 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5444 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5445 the first called function after an EVP_EncryptInit(), and therefore that
5446 specific call must be safe. The second form is where the length passed to
5447 EVP_EncryptUpdate() can be seen from the code to be some small value and
5448 therefore there is no possibility of an overflow. Since all instances are
5449 one of these two forms, it is believed that there can be no overflows in
5450 internal code due to this problem. It should be noted that
5451 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5452 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5453 of these calls have also been analysed too and it is believed there are no
5454 instances in internal usage where an overflow could occur.
5456 This issue was reported by Guido Vranken.
5461 * Prevent ASN.1 BIO excessive memory allocation
5463 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5464 a short invalid encoding can casuse allocation of large amounts of memory
5465 potentially consuming excessive resources or exhausting memory.
5467 Any application parsing untrusted data through d2i BIO functions is
5468 affected. The memory based functions such as d2i_X509() are *not* affected.
5469 Since the memory based functions are used by the TLS library, TLS
5470 applications are not affected.
5472 This issue was reported by Brian Carpenter.
5479 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5480 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5481 in arbitrary stack data being returned in the buffer.
5483 This issue was reported by Guido Vranken.
5488 * Modify behavior of ALPN to invoke callback after SNI/servername
5489 callback, such that updates to the SSL_CTX affect ALPN.
5493 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5498 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5499 methods are enabled and ssl2 is disabled the methods return NULL.
5503 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5505 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5506 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5507 provide any "EXPORT" or "LOW" strength ciphers.
5511 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5512 is by default disabled at build-time. Builds that are not configured with
5513 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5514 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5515 will need to explicitly call either of:
5517 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5519 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5521 as appropriate. Even if either of those is used, or the application
5522 explicitly uses the version-specific SSLv2_method() or its client and
5523 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5524 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5525 ciphers, and SSLv2 56-bit DES are no longer available.
5530 * Fix a double-free in DSA code
5532 A double free bug was discovered when OpenSSL parses malformed DSA private
5533 keys and could lead to a DoS attack or memory corruption for applications
5534 that receive DSA private keys from untrusted sources. This scenario is
5537 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5543 * Disable SRP fake user seed to address a server memory leak.
5545 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5547 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5548 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5549 was changed to ignore the "fake user" SRP seed, even if the seed
5552 Users should use SRP_VBASE_get1_by_user instead. Note that in
5553 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5554 also that even though configuring the SRP seed attempts to hide
5555 invalid usernames by continuing the handshake with fake
5556 credentials, this behaviour is not constant time and no strong
5557 guarantees are made that the handshake is indistinguishable from
5558 that of a valid user.
5563 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5565 In the BN_hex2bn function the number of hex digits is calculated using an
5566 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5567 large values of |i| this can result in |bn_expand| not allocating any
5568 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5569 field as NULL leading to a subsequent NULL ptr deref. For very large values
5570 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5571 In this case memory is allocated to the internal BIGNUM data field, but it
5572 is insufficiently sized leading to heap corruption. A similar issue exists
5573 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5574 is ever called by user applications with very large untrusted hex/dec data.
5575 This is anticipated to be a rare occurrence.
5577 All OpenSSL internal usage of these functions use data that is not expected
5578 to be untrusted, e.g. config file data or application command line
5579 arguments. If user developed applications generate config file data based
5580 on untrusted data then it is possible that this could also lead to security
5581 consequences. This is also anticipated to be rare.
5583 This issue was reported to OpenSSL by Guido Vranken.
5588 * Fix memory issues in `BIO_*printf` functions
5590 The internal |fmtstr| function used in processing a "%s" format string in
5591 the `BIO_*printf` functions could overflow while calculating the length of a
5592 string and cause an OOB read when printing very long strings.
5594 Additionally the internal |doapr_outch| function can attempt to write to an
5595 OOB memory location (at an offset from the NULL pointer) in the event of a
5596 memory allocation failure. In 1.0.2 and below this could be caused where
5597 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5598 could be in processing a very long "%s" format string. Memory leaks can
5601 The first issue may mask the second issue dependent on compiler behaviour.
5602 These problems could enable attacks where large amounts of untrusted data
5603 is passed to the `BIO_*printf` functions. If applications use these functions
5604 in this way then they could be vulnerable. OpenSSL itself uses these
5605 functions when printing out human-readable dumps of ASN.1 data. Therefore
5606 applications that print this data could be vulnerable if the data is from
5607 untrusted sources. OpenSSL command line applications could also be
5608 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5609 as command line arguments.
5611 Libssl is not considered directly vulnerable. Additionally certificates etc
5612 received via remote connections via libssl are also unlikely to be able to
5613 trigger these issues because of message size limits enforced within libssl.
5615 This issue was reported to OpenSSL Guido Vranken.
5620 * Side channel attack on modular exponentiation
5622 A side-channel attack was found which makes use of cache-bank conflicts on
5623 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5624 of RSA keys. The ability to exploit this issue is limited as it relies on
5625 an attacker who has control of code in a thread running on the same
5626 hyper-threaded core as the victim thread which is performing decryptions.
5628 This issue was reported to OpenSSL by Yuval Yarom, The University of
5629 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5630 Nadia Heninger, University of Pennsylvania with more information at
5631 <http://cachebleed.info>.
5636 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5637 if no keysize is specified with default_bits. This fixes an
5638 omission in an earlier change that changed all RSA/DSA key generation
5639 apps to use 2048 bits by default.
5643 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5645 * Protection for DH small subgroup attacks
5647 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5648 switched on by default and cannot be disabled. This could have some
5653 * SSLv2 doesn't block disabled ciphers
5655 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5656 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5657 been disabled, provided that the SSLv2 protocol was not also disabled via
5660 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5661 and Sebastian Schinzel.
5666 * Reject DH handshakes with parameters shorter than 1024 bits.
5670 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5672 * Certificate verify crash with missing PSS parameter
5674 The signature verification routines will crash with a NULL pointer
5675 dereference if presented with an ASN.1 signature using the RSA PSS
5676 algorithm and absent mask generation function parameter. Since these
5677 routines are used to verify certificate signature algorithms this can be
5678 used to crash any certificate verification operation and exploited in a
5679 DoS attack. Any application which performs certificate verification is
5680 vulnerable including OpenSSL clients and servers which enable client
5683 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5688 * X509_ATTRIBUTE memory leak
5690 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5691 memory. This structure is used by the PKCS#7 and CMS routines so any
5692 application which reads PKCS#7 or CMS data from untrusted sources is
5693 affected. SSL/TLS is not affected.
5695 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5701 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5702 This changes the decoding behaviour for some invalid messages,
5703 though the change is mostly in the more lenient direction, and
5704 legacy behaviour is preserved as much as possible.
5708 * In DSA_generate_parameters_ex, if the provided seed is too short,
5709 use a random seed, as already documented.
5711 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5713 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5715 * Alternate chains certificate forgery
5717 During certificate verfification, OpenSSL will attempt to find an
5718 alternative certificate chain if the first attempt to build such a chain
5719 fails. An error in the implementation of this logic can mean that an
5720 attacker could cause certain checks on untrusted certificates to be
5721 bypassed, such as the CA flag, enabling them to use a valid leaf
5722 certificate to act as a CA and "issue" an invalid certificate.
5724 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5730 * Race condition handling PSK identify hint
5732 If PSK identity hints are received by a multi-threaded client then
5733 the values are wrongly updated in the parent SSL_CTX structure. This can
5734 result in a race condition potentially leading to a double free of the
5740 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5742 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5743 incompatibility in the handling of HMAC. The previous ABI has now been
5746 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5748 * Malformed ECParameters causes infinite loop
5750 When processing an ECParameters structure OpenSSL enters an infinite loop
5751 if the curve specified is over a specially malformed binary polynomial
5754 This can be used to perform denial of service against any
5755 system which processes public keys, certificate requests or
5756 certificates. This includes TLS clients and TLS servers with
5757 client authentication enabled.
5759 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5764 * Exploitable out-of-bounds read in X509_cmp_time
5766 X509_cmp_time does not properly check the length of the ASN1_TIME
5767 string and can read a few bytes out of bounds. In addition,
5768 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5771 An attacker can use this to craft malformed certificates and CRLs of
5772 various sizes and potentially cause a segmentation fault, resulting in
5773 a DoS on applications that verify certificates or CRLs. TLS clients
5774 that verify CRLs are affected. TLS clients and servers with client
5775 authentication enabled may be affected if they use custom verification
5778 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5779 independently by Hanno Böck.
5784 * PKCS7 crash with missing EnvelopedContent
5786 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5787 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5788 with missing content and trigger a NULL pointer dereference on parsing.
5790 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5791 structures from untrusted sources are affected. OpenSSL clients and
5792 servers are not affected.
5794 This issue was reported to OpenSSL by Michal Zalewski (Google).
5799 * CMS verify infinite loop with unknown hash function
5801 When verifying a signedData message the CMS code can enter an infinite loop
5802 if presented with an unknown hash function OID. This can be used to perform
5803 denial of service against any system which verifies signedData messages using
5805 This issue was reported to OpenSSL by Johannes Bauer.
5810 * Race condition handling NewSessionTicket
5812 If a NewSessionTicket is received by a multi-threaded client when attempting to
5813 reuse a previous ticket then a race condition can occur potentially leading to
5814 a double free of the ticket data.
5819 * Reject DH handshakes with parameters shorter than 768 bits.
5821 *Kurt Roeckx and Emilia Kasper*
5823 * dhparam: generate 2048-bit parameters by default.
5825 *Kurt Roeckx and Emilia Kasper*
5827 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5829 * Segmentation fault in ASN1_TYPE_cmp fix
5831 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5832 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5833 certificate signature algorithm consistency this can be used to crash any
5834 certificate verification operation and exploited in a DoS attack. Any
5835 application which performs certificate verification is vulnerable including
5836 OpenSSL clients and servers which enable client authentication.
5841 * ASN.1 structure reuse memory corruption fix
5843 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5844 memory corruption via an invalid write. Such reuse is and has been
5845 strongly discouraged and is believed to be rare.
5847 Applications that parse structures containing CHOICE or ANY DEFINED BY
5848 components may be affected. Certificate parsing (d2i_X509 and related
5849 functions) are however not affected. OpenSSL clients and servers are
5855 * PKCS7 NULL pointer dereferences fix
5857 The PKCS#7 parsing code does not handle missing outer ContentInfo
5858 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5859 missing content and trigger a NULL pointer dereference on parsing.
5861 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5862 otherwise parse PKCS#7 structures from untrusted sources are
5863 affected. OpenSSL clients and servers are not affected.
5865 This issue was reported to OpenSSL by Michal Zalewski (Google).
5870 * DoS via reachable assert in SSLv2 servers fix
5872 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5873 servers that both support SSLv2 and enable export cipher suites by sending
5874 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5876 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5877 (OpenSSL development team).
5882 * Use After Free following d2i_ECPrivatekey error fix
5884 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5885 could cause a use after free condition. This, in turn, could cause a double
5886 free in several private key parsing functions (such as d2i_PrivateKey
5887 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5888 for applications that receive EC private keys from untrusted
5889 sources. This scenario is considered rare.
5891 This issue was discovered by the BoringSSL project and fixed in their
5897 * X509_to_X509_REQ NULL pointer deref fix
5899 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5900 the certificate key is invalid. This function is rarely used in practice.
5902 This issue was discovered by Brian Carpenter.
5907 * Removed the export ciphers from the DEFAULT ciphers
5911 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
5913 * Build fixes for the Windows and OpenVMS platforms
5915 *Matt Caswell and Richard Levitte*
5917 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
5919 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5920 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5921 dereference. This could lead to a Denial Of Service attack. Thanks to
5922 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5927 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5928 dtls1_buffer_record function under certain conditions. In particular this
5929 could occur if an attacker sent repeated DTLS records with the same
5930 sequence number but for the next epoch. The memory leak could be exploited
5931 by an attacker in a Denial of Service attack through memory exhaustion.
5932 Thanks to Chris Mueller for reporting this issue.
5937 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5938 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5939 method would be set to NULL which could later result in a NULL pointer
5940 dereference. Thanks to Frank Schmirler for reporting this issue.
5945 * Abort handshake if server key exchange message is omitted for ephemeral
5948 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5949 reporting this issue.
5954 * Remove non-export ephemeral RSA code on client and server. This code
5955 violated the TLS standard by allowing the use of temporary RSA keys in
5956 non-export ciphersuites and could be used by a server to effectively
5957 downgrade the RSA key length used to a value smaller than the server
5958 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5959 INRIA or reporting this issue.
5964 * Fixed issue where DH client certificates are accepted without verification.
5965 An OpenSSL server will accept a DH certificate for client authentication
5966 without the certificate verify message. This effectively allows a client to
5967 authenticate without the use of a private key. This only affects servers
5968 which trust a client certificate authority which issues certificates
5969 containing DH keys: these are extremely rare and hardly ever encountered.
5970 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5976 * Ensure that the session ID context of an SSL is updated when its
5977 SSL_CTX is updated via SSL_set_SSL_CTX.
5979 The session ID context is typically set from the parent SSL_CTX,
5980 and can vary with the CTX.
5984 * Fix various certificate fingerprint issues.
5986 By using non-DER or invalid encodings outside the signed portion of a
5987 certificate the fingerprint can be changed without breaking the signature.
5988 Although no details of the signed portion of the certificate can be changed
5989 this can cause problems with some applications: e.g. those using the
5990 certificate fingerprint for blacklists.
5992 1. Reject signatures with non zero unused bits.
5994 If the BIT STRING containing the signature has non zero unused bits reject
5995 the signature. All current signature algorithms require zero unused bits.
5997 2. Check certificate algorithm consistency.
5999 Check the AlgorithmIdentifier inside TBS matches the one in the
6000 certificate signature. NB: this will result in signature failure
6001 errors for some broken certificates.
6003 Thanks to Konrad Kraszewski from Google for reporting this issue.
6005 3. Check DSA/ECDSA signatures use DER.
6007 Re-encode DSA/ECDSA signatures and compare with the original received
6008 signature. Return an error if there is a mismatch.
6010 This will reject various cases including garbage after signature
6011 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6012 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6013 (negative or with leading zeroes).
6015 Further analysis was conducted and fixes were developed by Stephen Henson
6016 of the OpenSSL core team.
6022 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6023 results on some platforms, including x86_64. This bug occurs at random
6024 with a very low probability, and is not known to be exploitable in any
6025 way, though its exact impact is difficult to determine. Thanks to Pieter
6026 Wuille (Blockstream) who reported this issue and also suggested an initial
6027 fix. Further analysis was conducted by the OpenSSL development team and
6028 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6029 the OpenSSL core team.
6034 * Do not resume sessions on the server if the negotiated protocol
6035 version does not match the session's version. Resuming with a different
6036 version, while not strictly forbidden by the RFC, is of questionable
6037 sanity and breaks all known clients.
6039 *David Benjamin, Emilia Käsper*
6041 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6042 early CCS messages during renegotiation. (Note that because
6043 renegotiation is encrypted, this early CCS was not exploitable.)
6047 * Tighten client-side session ticket handling during renegotiation:
6048 ensure that the client only accepts a session ticket if the server sends
6049 the extension anew in the ServerHello. Previously, a TLS client would
6050 reuse the old extension state and thus accept a session ticket if one was
6051 announced in the initial ServerHello.
6053 Similarly, ensure that the client requires a session ticket if one
6054 was advertised in the ServerHello. Previously, a TLS client would
6055 ignore a missing NewSessionTicket message.
6059 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6063 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6064 sends a carefully crafted handshake message, to cause OpenSSL to fail
6065 to free up to 64k of memory causing a memory leak. This could be
6066 exploited in a Denial Of Service attack. This issue affects OpenSSL
6067 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6068 whether SRTP is used or configured. Implementations of OpenSSL that
6069 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6071 The fix was developed by the OpenSSL team.
6076 * Session Ticket Memory Leak.
6078 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6079 integrity of that ticket is first verified. In the event of a session
6080 ticket integrity check failing, OpenSSL will fail to free memory
6081 causing a memory leak. By sending a large number of invalid session
6082 tickets an attacker could exploit this issue in a Denial Of Service
6088 * Build option no-ssl3 is incomplete.
6090 When OpenSSL is configured with "no-ssl3" as a build option, servers
6091 could accept and complete a SSL 3.0 handshake, and clients could be
6092 configured to send them.
6095 *Akamai and the OpenSSL team*
6097 * Add support for TLS_FALLBACK_SCSV.
6098 Client applications doing fallback retries should call
6099 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6102 *Adam Langley, Bodo Moeller*
6104 * Add additional DigestInfo checks.
6106 Re-encode DigestInto in DER and check against the original when
6107 verifying RSA signature: this will reject any improperly encoded
6108 DigestInfo structures.
6110 Note: this is a precautionary measure and no attacks are currently known.
6114 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6116 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6117 SRP code can be overrun an internal buffer. Add sanity check that
6118 g, A, B < N to SRP code.
6120 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6121 Group for discovering this issue.
6126 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6127 TLS 1.0 instead of higher protocol versions when the ClientHello message
6128 is badly fragmented. This allows a man-in-the-middle attacker to force a
6129 downgrade to TLS 1.0 even if both the server and the client support a
6130 higher protocol version, by modifying the client's TLS records.
6132 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6133 researching this issue.
6138 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6139 to a denial of service attack. A malicious server can crash the client
6140 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6141 ciphersuite and sending carefully crafted handshake messages.
6143 Thanks to Felix Gröbert (Google) for discovering and researching this
6149 * By sending carefully crafted DTLS packets an attacker could cause openssl
6150 to leak memory. This can be exploited through a Denial of Service attack.
6151 Thanks to Adam Langley for discovering and researching this issue.
6156 * An attacker can force openssl to consume large amounts of memory whilst
6157 processing DTLS handshake messages. This can be exploited through a
6158 Denial of Service attack.
6159 Thanks to Adam Langley for discovering and researching this issue.
6164 * An attacker can force an error condition which causes openssl to crash
6165 whilst processing DTLS packets due to memory being freed twice. This
6166 can be exploited through a Denial of Service attack.
6167 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6173 * If a multithreaded client connects to a malicious server using a resumed
6174 session and the server sends an ec point format extension it could write
6175 up to 255 bytes to freed memory.
6177 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6183 * A malicious server can crash an OpenSSL client with a null pointer
6184 dereference (read) by specifying an SRP ciphersuite even though it was not
6185 properly negotiated with the client. This can be exploited through a
6186 Denial of Service attack.
6188 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6189 discovering and researching this issue.
6194 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6195 X509_name_oneline, X509_name_print_ex et al. to leak some information
6196 from the stack. Applications may be affected if they echo pretty printing
6197 output to the attacker.
6199 Thanks to Ivan Fratric (Google) for discovering this issue.
6202 *Emilia Käsper, and Steve Henson*
6204 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6205 for corner cases. (Certain input points at infinity could lead to
6206 bogus results, with non-infinity inputs mapped to infinity too.)
6210 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6212 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6213 handshake can force the use of weak keying material in OpenSSL
6214 SSL/TLS clients and servers.
6216 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6217 researching this issue. [CVE-2014-0224][]
6219 *KIKUCHI Masashi, Steve Henson*
6221 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6222 OpenSSL DTLS client the code can be made to recurse eventually crashing
6225 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6228 *Imre Rad, Steve Henson*
6230 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6231 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6232 client or server. This is potentially exploitable to run arbitrary
6233 code on a vulnerable client or server.
6235 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6237 *Jüri Aedla, Steve Henson*
6239 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6240 are subject to a denial of service attack.
6242 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6243 this issue. [CVE-2014-3470][]
6245 *Felix Gröbert, Ivan Fratric, Steve Henson*
6247 * Harmonize version and its documentation. -f flag is used to display
6250 *mancha <mancha1@zoho.com>*
6252 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6253 in i2d_ECPrivateKey.
6255 *mancha <mancha1@zoho.com>*
6257 * Fix some double frees. These are not thought to be exploitable.
6259 *mancha <mancha1@zoho.com>*
6261 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6263 * A missing bounds check in the handling of the TLS heartbeat extension
6264 can be used to reveal up to 64k of memory to a connected client or
6267 Thanks for Neel Mehta of Google Security for discovering this bug and to
6268 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6269 preparing the fix [CVE-2014-0160][]
6271 *Adam Langley, Bodo Moeller*
6273 * Fix for the attack described in the paper "Recovering OpenSSL
6274 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6275 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6276 <http://eprint.iacr.org/2014/140>
6278 Thanks to Yuval Yarom and Naomi Benger for discovering this
6279 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6281 *Yuval Yarom and Naomi Benger*
6283 * TLS pad extension: draft-agl-tls-padding-03
6285 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6286 TLS client Hello record length value would otherwise be > 255 and
6287 less that 512 pad with a dummy extension containing zeroes so it
6288 is at least 512 bytes long.
6290 *Adam Langley, Steve Henson*
6292 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6294 * Fix for TLS record tampering bug. A carefully crafted invalid
6295 handshake could crash OpenSSL with a NULL pointer exception.
6296 Thanks to Anton Johansson for reporting this issues.
6299 * Keep original DTLS digest and encryption contexts in retransmission
6300 structures so we can use the previous session parameters if they need
6301 to be resent. [CVE-2013-6450][]
6305 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6306 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6307 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6308 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6309 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6310 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6312 *Rob Stradling, Adam Langley*
6314 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6316 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6317 supporting platforms or when small records were transferred.
6319 *Andy Polyakov, Steve Henson*
6321 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6323 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6325 This addresses the flaw in CBC record processing discovered by
6326 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6327 at: <http://www.isg.rhul.ac.uk/tls/>
6329 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6330 Security Group at Royal Holloway, University of London
6331 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6332 Emilia Käsper for the initial patch.
6335 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6337 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6338 ciphersuites which can be exploited in a denial of service attack.
6339 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6340 and detecting this bug and to Wolfgang Ettlinger
6341 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6346 * Return an error when checking OCSP signatures when key is NULL.
6347 This fixes a DoS attack. [CVE-2013-0166][]
6351 * Make openssl verify return errors.
6353 *Chris Palmer <palmer@google.com> and Ben Laurie*
6355 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6356 the right response is stapled. Also change SSL_get_certificate()
6357 so it returns the certificate actually sent.
6358 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6360 *Rob Stradling <rob.stradling@comodo.com>*
6362 * Fix possible deadlock when decoding public keys.
6366 * Don't use TLS 1.0 record version number in initial client hello
6371 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6373 * Sanity check record length before skipping explicit IV in TLS
6374 1.2, 1.1 and DTLS to fix DoS attack.
6376 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6377 fuzzing as a service testing platform.
6382 * Initialise tkeylen properly when encrypting CMS messages.
6383 Thanks to Solar Designer of Openwall for reporting this issue.
6387 * In FIPS mode don't try to use composite ciphers as they are not
6392 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6394 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6395 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6396 mean any application compiled against OpenSSL 1.0.0 headers setting
6397 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6398 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6399 0x10000000L Any application which was previously compiled against
6400 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6401 will need to be recompiled as a result. Letting be results in
6402 inability to disable specifically TLS 1.1 and in client context,
6403 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6407 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6408 disable just protocol X, but all protocols above X *if* there are
6409 protocols *below* X still enabled. In more practical terms it means
6410 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6411 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6412 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6417 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6419 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6420 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6421 in CRYPTO_realloc_clean.
6423 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6424 issue and to Adam Langley <agl@chromium.org> for fixing it.
6427 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6429 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6433 * Workarounds for some broken servers that "hang" if a client hello
6434 record length exceeds 255 bytes.
6436 1. Do not use record version number > TLS 1.0 in initial client
6437 hello: some (but not all) hanging servers will now work.
6438 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6439 the number of ciphers sent in the client hello. This should be
6440 set to an even number, such as 50, for example by passing:
6441 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6442 Most broken servers should now work.
6443 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6444 TLS 1.2 client support entirely.
6448 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6452 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6454 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6455 STRING form instead of a DigestInfo.
6459 * The format used for MDC2 RSA signatures is inconsistent between EVP
6460 and the RSA_sign/RSA_verify functions. This was made more apparent when
6461 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6462 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6463 the correct format in RSA_verify so both forms transparently work.
6467 * Some servers which support TLS 1.0 can choke if we initially indicate
6468 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6469 encrypted premaster secret. As a workaround use the maximum permitted
6470 client version in client hello, this should keep such servers happy
6471 and still work with previous versions of OpenSSL.
6475 * Add support for TLS/DTLS heartbeats.
6477 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6479 * Add support for SCTP.
6481 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6483 * Improved PRNG seeding for VOS.
6485 *Paul Green <Paul.Green@stratus.com>*
6487 * Extensive assembler packs updates, most notably:
6489 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6490 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6491 - x86_64: bit-sliced AES implementation;
6492 - ARM: NEON support, contemporary platforms optimizations;
6493 - s390x: z196 support;
6494 - `*`: GHASH and GF(2^m) multiplication implementations;
6498 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6499 (removal of unnecessary code)
6501 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6503 * Add TLS key material exporter from RFC 5705.
6507 * Add DTLS-SRTP negotiation from RFC 5764.
6511 * Add Next Protocol Negotiation,
6512 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6513 disabled with a no-npn flag to config or Configure. Code donated
6516 *Adam Langley <agl@google.com> and Ben Laurie*
6518 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6519 NIST-P256, NIST-P521, with constant-time single point multiplication on
6520 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6521 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6522 Code made available under Apache License version 2.0.
6524 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6525 line to include this in your build of OpenSSL, and run "make depend" (or
6526 "make update"). This enables the following EC_METHODs:
6528 EC_GFp_nistp224_method()
6529 EC_GFp_nistp256_method()
6530 EC_GFp_nistp521_method()
6532 EC_GROUP_new_by_curve_name() will automatically use these (while
6533 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6536 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6538 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6539 all platforms. Move ssize_t definition from e_os.h to the public
6540 header file e_os2.h as it now appears in public header file cms.h
6544 * New -sigopt option to the ca, req and x509 utilities. Additional
6545 signature parameters can be passed using this option and in
6550 * Add RSA PSS signing function. This will generate and set the
6551 appropriate AlgorithmIdentifiers for PSS based on those in the
6552 corresponding EVP_MD_CTX structure. No application support yet.
6556 * Support for companion algorithm specific ASN1 signing routines.
6557 New function ASN1_item_sign_ctx() signs a pre-initialised
6558 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6559 the appropriate parameters.
6563 * Add new algorithm specific ASN1 verification initialisation function
6564 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6565 handling will be the same no matter what EVP_PKEY_METHOD is used.
6566 Add a PSS handler to support verification of PSS signatures: checked
6567 against a number of sample certificates.
6571 * Add signature printing for PSS. Add PSS OIDs.
6573 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6575 * Add algorithm specific signature printing. An individual ASN1 method
6576 can now print out signatures instead of the standard hex dump.
6578 More complex signatures (e.g. PSS) can print out more meaningful
6579 information. Include DSA version that prints out the signature
6584 * Password based recipient info support for CMS library: implementing
6589 * Split password based encryption into PBES2 and PBKDF2 functions. This
6590 neatly separates the code into cipher and PBE sections and is required
6591 for some algorithms that split PBES2 into separate pieces (such as
6592 password based CMS).
6596 * Session-handling fixes:
6597 - Fix handling of connections that are resuming with a session ID,
6598 but also support Session Tickets.
6599 - Fix a bug that suppressed issuing of a new ticket if the client
6600 presented a ticket with an expired session.
6601 - Try to set the ticket lifetime hint to something reasonable.
6602 - Make tickets shorter by excluding irrelevant information.
6603 - On the client side, don't ignore renewed tickets.
6605 *Adam Langley, Bodo Moeller (Google)*
6607 * Fix PSK session representation.
6611 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6613 This work was sponsored by Intel.
6617 * Add GCM support to TLS library. Some custom code is needed to split
6618 the IV between the fixed (from PRF) and explicit (from TLS record)
6619 portions. This adds all GCM ciphersuites supported by RFC5288 and
6620 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6621 add a special AESGCM string for GCM only.
6625 * Expand range of ctrls for AES GCM. Permit setting invocation
6626 field on decrypt and retrieval of invocation field only on encrypt.
6630 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6631 As required by RFC5289 these ciphersuites cannot be used if for
6632 versions of TLS earlier than 1.2.
6636 * For FIPS capable OpenSSL interpret a NULL default public key method
6637 as unset and return the appropriate default but do *not* set the default.
6638 This means we can return the appropriate method in applications that
6639 switch between FIPS and non-FIPS modes.
6643 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6644 ENGINE is used then we cannot handle that in the FIPS module so we
6645 keep original code iff non-FIPS operations are allowed.
6649 * Add -attime option to openssl utilities.
6651 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6653 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6657 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6658 FIPS EC methods unconditionally for now.
6662 * New build option no-ec2m to disable characteristic 2 code.
6666 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6667 all cases can be covered as some introduce binary incompatibilities.
6671 * Redirect RSA operations to FIPS module including keygen,
6672 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6676 * Add similar low level API blocking to ciphers.
6680 * Low level digest APIs are not approved in FIPS mode: any attempt
6681 to use these will cause a fatal error. Applications that *really* want
6682 to use them can use the `private_*` version instead.
6686 * Redirect cipher operations to FIPS module for FIPS builds.
6690 * Redirect digest operations to FIPS module for FIPS builds.
6694 * Update build system to add "fips" flag which will link in fipscanister.o
6695 for static and shared library builds embedding a signature if needed.
6699 * Output TLS supported curves in preference order instead of numerical
6700 order. This is currently hardcoded for the highest order curves first.
6701 This should be configurable so applications can judge speed vs strength.
6705 * Add TLS v1.2 server support for client authentication.
6709 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6714 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6715 FIPS modules versions.
6719 * Add TLS v1.2 client side support for client authentication. Keep cache
6720 of handshake records longer as we don't know the hash algorithm to use
6721 until after the certificate request message is received.
6725 * Initial TLS v1.2 client support. Add a default signature algorithms
6726 extension including all the algorithms we support. Parse new signature
6727 format in client key exchange. Relax some ECC signing restrictions for
6728 TLS v1.2 as indicated in RFC5246.
6732 * Add server support for TLS v1.2 signature algorithms extension. Switch
6733 to new signature format when needed using client digest preference.
6734 All server ciphersuites should now work correctly in TLS v1.2. No client
6735 support yet and no support for client certificates.
6739 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6740 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6741 ciphersuites. At present only RSA key exchange ciphersuites work with
6742 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6743 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6744 and version checking.
6748 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6749 with this defined it will not be affected by any changes to ssl internal
6750 structures. Add several utility functions to allow openssl application
6751 to work with OPENSSL_NO_SSL_INTERN defined.
6755 * A long standing patch to add support for SRP from EdelWeb (Peter
6756 Sylvester and Christophe Renou) was integrated.
6757 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6758 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6761 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6765 * Permit abbreviated handshakes when renegotiating using the function
6766 SSL_renegotiate_abbreviated().
6768 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6770 * Add call to ENGINE_register_all_complete() to
6771 ENGINE_load_builtin_engines(), so some implementations get used
6772 automatically instead of needing explicit application support.
6776 * Add support for TLS key exporter as described in RFC5705.
6778 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6780 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6781 a few changes are required:
6783 Add SSL_OP_NO_TLSv1_1 flag.
6784 Add TLSv1_1 methods.
6785 Update version checking logic to handle version 1.1.
6786 Add explicit IV handling (ported from DTLS code).
6787 Add command line options to s_client/s_server.
6794 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6796 * X509_ATTRIBUTE memory leak
6798 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6799 memory. This structure is used by the PKCS#7 and CMS routines so any
6800 application which reads PKCS#7 or CMS data from untrusted sources is
6801 affected. SSL/TLS is not affected.
6803 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6809 * Race condition handling PSK identify hint
6811 If PSK identity hints are received by a multi-threaded client then
6812 the values are wrongly updated in the parent SSL_CTX structure. This can
6813 result in a race condition potentially leading to a double free of the
6819 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6821 * Malformed ECParameters causes infinite loop
6823 When processing an ECParameters structure OpenSSL enters an infinite loop
6824 if the curve specified is over a specially malformed binary polynomial
6827 This can be used to perform denial of service against any
6828 system which processes public keys, certificate requests or
6829 certificates. This includes TLS clients and TLS servers with
6830 client authentication enabled.
6832 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6837 * Exploitable out-of-bounds read in X509_cmp_time
6839 X509_cmp_time does not properly check the length of the ASN1_TIME
6840 string and can read a few bytes out of bounds. In addition,
6841 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6844 An attacker can use this to craft malformed certificates and CRLs of
6845 various sizes and potentially cause a segmentation fault, resulting in
6846 a DoS on applications that verify certificates or CRLs. TLS clients
6847 that verify CRLs are affected. TLS clients and servers with client
6848 authentication enabled may be affected if they use custom verification
6851 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6852 independently by Hanno Böck.
6857 * PKCS7 crash with missing EnvelopedContent
6859 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6860 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6861 with missing content and trigger a NULL pointer dereference on parsing.
6863 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6864 structures from untrusted sources are affected. OpenSSL clients and
6865 servers are not affected.
6867 This issue was reported to OpenSSL by Michal Zalewski (Google).
6872 * CMS verify infinite loop with unknown hash function
6874 When verifying a signedData message the CMS code can enter an infinite loop
6875 if presented with an unknown hash function OID. This can be used to perform
6876 denial of service against any system which verifies signedData messages using
6878 This issue was reported to OpenSSL by Johannes Bauer.
6883 * Race condition handling NewSessionTicket
6885 If a NewSessionTicket is received by a multi-threaded client when attempting to
6886 reuse a previous ticket then a race condition can occur potentially leading to
6887 a double free of the ticket data.
6892 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
6894 * Segmentation fault in ASN1_TYPE_cmp fix
6896 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6897 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6898 certificate signature algorithm consistency this can be used to crash any
6899 certificate verification operation and exploited in a DoS attack. Any
6900 application which performs certificate verification is vulnerable including
6901 OpenSSL clients and servers which enable client authentication.
6906 * ASN.1 structure reuse memory corruption fix
6908 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6909 memory corruption via an invalid write. Such reuse is and has been
6910 strongly discouraged and is believed to be rare.
6912 Applications that parse structures containing CHOICE or ANY DEFINED BY
6913 components may be affected. Certificate parsing (d2i_X509 and related
6914 functions) are however not affected. OpenSSL clients and servers are
6920 * PKCS7 NULL pointer dereferences fix
6922 The PKCS#7 parsing code does not handle missing outer ContentInfo
6923 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6924 missing content and trigger a NULL pointer dereference on parsing.
6926 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6927 otherwise parse PKCS#7 structures from untrusted sources are
6928 affected. OpenSSL clients and servers are not affected.
6930 This issue was reported to OpenSSL by Michal Zalewski (Google).
6935 * DoS via reachable assert in SSLv2 servers fix
6937 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6938 servers that both support SSLv2 and enable export cipher suites by sending
6939 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6941 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6942 (OpenSSL development team).
6947 * Use After Free following d2i_ECPrivatekey error fix
6949 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6950 could cause a use after free condition. This, in turn, could cause a double
6951 free in several private key parsing functions (such as d2i_PrivateKey
6952 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6953 for applications that receive EC private keys from untrusted
6954 sources. This scenario is considered rare.
6956 This issue was discovered by the BoringSSL project and fixed in their
6962 * X509_to_X509_REQ NULL pointer deref fix
6964 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6965 the certificate key is invalid. This function is rarely used in practice.
6967 This issue was discovered by Brian Carpenter.
6972 * Removed the export ciphers from the DEFAULT ciphers
6976 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
6978 * Build fixes for the Windows and OpenVMS platforms
6980 *Matt Caswell and Richard Levitte*
6982 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
6984 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6985 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6986 dereference. This could lead to a Denial Of Service attack. Thanks to
6987 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6992 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6993 dtls1_buffer_record function under certain conditions. In particular this
6994 could occur if an attacker sent repeated DTLS records with the same
6995 sequence number but for the next epoch. The memory leak could be exploited
6996 by an attacker in a Denial of Service attack through memory exhaustion.
6997 Thanks to Chris Mueller for reporting this issue.
7002 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7003 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7004 method would be set to NULL which could later result in a NULL pointer
7005 dereference. Thanks to Frank Schmirler for reporting this issue.
7010 * Abort handshake if server key exchange message is omitted for ephemeral
7013 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7014 reporting this issue.
7019 * Remove non-export ephemeral RSA code on client and server. This code
7020 violated the TLS standard by allowing the use of temporary RSA keys in
7021 non-export ciphersuites and could be used by a server to effectively
7022 downgrade the RSA key length used to a value smaller than the server
7023 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7024 INRIA or reporting this issue.
7029 * Fixed issue where DH client certificates are accepted without verification.
7030 An OpenSSL server will accept a DH certificate for client authentication
7031 without the certificate verify message. This effectively allows a client to
7032 authenticate without the use of a private key. This only affects servers
7033 which trust a client certificate authority which issues certificates
7034 containing DH keys: these are extremely rare and hardly ever encountered.
7035 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7041 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7042 results on some platforms, including x86_64. This bug occurs at random
7043 with a very low probability, and is not known to be exploitable in any
7044 way, though its exact impact is difficult to determine. Thanks to Pieter
7045 Wuille (Blockstream) who reported this issue and also suggested an initial
7046 fix. Further analysis was conducted by the OpenSSL development team and
7047 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7048 the OpenSSL core team.
7053 * Fix various certificate fingerprint issues.
7055 By using non-DER or invalid encodings outside the signed portion of a
7056 certificate the fingerprint can be changed without breaking the signature.
7057 Although no details of the signed portion of the certificate can be changed
7058 this can cause problems with some applications: e.g. those using the
7059 certificate fingerprint for blacklists.
7061 1. Reject signatures with non zero unused bits.
7063 If the BIT STRING containing the signature has non zero unused bits reject
7064 the signature. All current signature algorithms require zero unused bits.
7066 2. Check certificate algorithm consistency.
7068 Check the AlgorithmIdentifier inside TBS matches the one in the
7069 certificate signature. NB: this will result in signature failure
7070 errors for some broken certificates.
7072 Thanks to Konrad Kraszewski from Google for reporting this issue.
7074 3. Check DSA/ECDSA signatures use DER.
7076 Reencode DSA/ECDSA signatures and compare with the original received
7077 signature. Return an error if there is a mismatch.
7079 This will reject various cases including garbage after signature
7080 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7081 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7082 (negative or with leading zeroes).
7084 Further analysis was conducted and fixes were developed by Stephen Henson
7085 of the OpenSSL core team.
7091 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7093 * Session Ticket Memory Leak.
7095 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7096 integrity of that ticket is first verified. In the event of a session
7097 ticket integrity check failing, OpenSSL will fail to free memory
7098 causing a memory leak. By sending a large number of invalid session
7099 tickets an attacker could exploit this issue in a Denial Of Service
7105 * Build option no-ssl3 is incomplete.
7107 When OpenSSL is configured with "no-ssl3" as a build option, servers
7108 could accept and complete a SSL 3.0 handshake, and clients could be
7109 configured to send them.
7112 *Akamai and the OpenSSL team*
7114 * Add support for TLS_FALLBACK_SCSV.
7115 Client applications doing fallback retries should call
7116 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7119 *Adam Langley, Bodo Moeller*
7121 * Add additional DigestInfo checks.
7123 Reencode DigestInto in DER and check against the original when
7124 verifying RSA signature: this will reject any improperly encoded
7125 DigestInfo structures.
7127 Note: this is a precautionary measure and no attacks are currently known.
7131 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7133 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7134 to a denial of service attack. A malicious server can crash the client
7135 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7136 ciphersuite and sending carefully crafted handshake messages.
7138 Thanks to Felix Gröbert (Google) for discovering and researching this
7144 * By sending carefully crafted DTLS packets an attacker could cause openssl
7145 to leak memory. This can be exploited through a Denial of Service attack.
7146 Thanks to Adam Langley for discovering and researching this issue.
7151 * An attacker can force openssl to consume large amounts of memory whilst
7152 processing DTLS handshake messages. This can be exploited through a
7153 Denial of Service attack.
7154 Thanks to Adam Langley for discovering and researching this issue.
7159 * An attacker can force an error condition which causes openssl to crash
7160 whilst processing DTLS packets due to memory being freed twice. This
7161 can be exploited through a Denial of Service attack.
7162 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7168 * If a multithreaded client connects to a malicious server using a resumed
7169 session and the server sends an ec point format extension it could write
7170 up to 255 bytes to freed memory.
7172 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7178 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7179 X509_name_oneline, X509_name_print_ex et al. to leak some information
7180 from the stack. Applications may be affected if they echo pretty printing
7181 output to the attacker.
7183 Thanks to Ivan Fratric (Google) for discovering this issue.
7186 *Emilia Käsper, and Steve Henson*
7188 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7189 for corner cases. (Certain input points at infinity could lead to
7190 bogus results, with non-infinity inputs mapped to infinity too.)
7194 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7196 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7197 handshake can force the use of weak keying material in OpenSSL
7198 SSL/TLS clients and servers.
7200 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7201 researching this issue. [CVE-2014-0224][]
7203 *KIKUCHI Masashi, Steve Henson*
7205 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7206 OpenSSL DTLS client the code can be made to recurse eventually crashing
7209 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7212 *Imre Rad, Steve Henson*
7214 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7215 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7216 client or server. This is potentially exploitable to run arbitrary
7217 code on a vulnerable client or server.
7219 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7221 *Jüri Aedla, Steve Henson*
7223 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7224 are subject to a denial of service attack.
7226 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7227 this issue. [CVE-2014-3470][]
7229 *Felix Gröbert, Ivan Fratric, Steve Henson*
7231 * Harmonize version and its documentation. -f flag is used to display
7234 *mancha <mancha1@zoho.com>*
7236 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7237 in i2d_ECPrivateKey.
7239 *mancha <mancha1@zoho.com>*
7241 * Fix some double frees. These are not thought to be exploitable.
7243 *mancha <mancha1@zoho.com>*
7245 * Fix for the attack described in the paper "Recovering OpenSSL
7246 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7247 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7248 <http://eprint.iacr.org/2014/140>
7250 Thanks to Yuval Yarom and Naomi Benger for discovering this
7251 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7253 *Yuval Yarom and Naomi Benger*
7255 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7257 * Keep original DTLS digest and encryption contexts in retransmission
7258 structures so we can use the previous session parameters if they need
7259 to be resent. [CVE-2013-6450][]
7263 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7264 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7265 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7266 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7267 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7268 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7270 *Rob Stradling, Adam Langley*
7272 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7274 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7276 This addresses the flaw in CBC record processing discovered by
7277 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7278 at: <http://www.isg.rhul.ac.uk/tls/>
7280 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7281 Security Group at Royal Holloway, University of London
7282 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7283 Emilia Käsper for the initial patch.
7286 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7288 * Return an error when checking OCSP signatures when key is NULL.
7289 This fixes a DoS attack. [CVE-2013-0166][]
7293 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7294 the right response is stapled. Also change SSL_get_certificate()
7295 so it returns the certificate actually sent.
7296 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7297 (This is a backport)
7299 *Rob Stradling <rob.stradling@comodo.com>*
7301 * Fix possible deadlock when decoding public keys.
7305 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7307 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7310 * Sanity check record length before skipping explicit IV in DTLS
7313 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7314 fuzzing as a service testing platform.
7319 * Initialise tkeylen properly when encrypting CMS messages.
7320 Thanks to Solar Designer of Openwall for reporting this issue.
7324 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7326 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7327 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7328 in CRYPTO_realloc_clean.
7330 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7331 issue and to Adam Langley <agl@chromium.org> for fixing it.
7334 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7336 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7338 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7339 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7340 content decryption and always return the same error. Note: this attack
7341 needs on average 2^20 messages so it only affects automated senders. The
7342 old behaviour can be re-enabled in the CMS code by setting the
7343 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7344 an MMA defence is not necessary.
7345 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7346 this issue. [CVE-2012-0884][]
7350 * Fix CVE-2011-4619: make sure we really are receiving a
7351 client hello before rejecting multiple SGC restarts. Thanks to
7352 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7356 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7358 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7359 Thanks to Antonio Martin, Enterprise Secure Access Research and
7360 Development, Cisco Systems, Inc. for discovering this bug and
7361 preparing a fix. [CVE-2012-0050][]
7365 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7367 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7368 of the Vaudenay padding oracle attack on CBC mode encryption
7369 which enables an efficient plaintext recovery attack against
7370 the OpenSSL implementation of DTLS. Their attack exploits timing
7371 differences arising during decryption processing. A research
7372 paper describing this attack can be found at:
7373 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7374 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7375 Security Group at Royal Holloway, University of London
7376 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7377 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7378 for preparing the fix. [CVE-2011-4108][]
7380 *Robin Seggelmann, Michael Tuexen*
7382 * Clear bytes used for block padding of SSL 3.0 records.
7385 *Adam Langley (Google)*
7387 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7388 Kadianakis <desnacked@gmail.com> for discovering this issue and
7389 Adam Langley for preparing the fix. [CVE-2011-4619][]
7391 *Adam Langley (Google)*
7393 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7395 *Andrey Kulikov <amdeich@gmail.com>*
7397 * Prevent malformed RFC3779 data triggering an assertion failure.
7398 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7399 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7401 *Rob Austein <sra@hactrn.net>*
7403 * Improved PRNG seeding for VOS.
7405 *Paul Green <Paul.Green@stratus.com>*
7407 * Fix ssl_ciph.c set-up race.
7409 *Adam Langley (Google)*
7411 * Fix spurious failures in ecdsatest.c.
7413 *Emilia Käsper (Google)*
7415 * Fix the BIO_f_buffer() implementation (which was mixing different
7416 interpretations of the `..._len` fields).
7418 *Adam Langley (Google)*
7420 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7421 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7422 threads won't reuse the same blinding coefficients.
7424 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7425 lock to call BN_BLINDING_invert_ex, and avoids one use of
7426 BN_BLINDING_update for each BN_BLINDING structure (previously,
7427 the last update always remained unused).
7429 *Emilia Käsper (Google)*
7431 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7433 *Bob Buckholz (Google)*
7435 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7437 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7438 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7440 *Kaspar Brand <ossl@velox.ch>*
7442 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7443 for multi-threaded use of ECDH. [CVE-2011-3210][]
7445 *Adam Langley (Google)*
7447 * Fix x509_name_ex_d2i memory leak on bad inputs.
7451 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7452 signature public key algorithm by using OID xref utilities instead.
7453 Before this you could only use some ECC ciphersuites with SHA1 only.
7457 * Add protection against ECDSA timing attacks as mentioned in the paper
7458 by Billy Bob Brumley and Nicola Tuveri, see:
7459 <http://eprint.iacr.org/2011/232.pdf>
7461 *Billy Bob Brumley and Nicola Tuveri*
7463 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7465 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7467 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7469 * Fix bug in string printing code: if *any* escaping is enabled we must
7470 escape the escape character (backslash) or the resulting string is
7475 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7477 * Disable code workaround for ancient and obsolete Netscape browsers
7478 and servers: an attacker can use it in a ciphersuite downgrade attack.
7479 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7483 * Fixed J-PAKE implementation error, originally discovered by
7484 Sebastien Martini, further info and confirmation from Stefan
7485 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7489 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7491 * Fix extension code to avoid race conditions which can result in a buffer
7492 overrun vulnerability: resumed sessions must not be modified as they can
7493 be shared by multiple threads. CVE-2010-3864
7497 * Fix WIN32 build system to correctly link an ENGINE directory into
7502 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7504 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7507 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7509 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7511 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7512 context. The operation can be customised via the ctrl mechanism in
7513 case ENGINEs want to include additional functionality.
7517 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7521 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7522 output hashes compatible with older versions of OpenSSL.
7524 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7526 * Fix compression algorithm handling: if resuming a session use the
7527 compression algorithm of the resumed session instead of determining
7528 it from client hello again. Don't allow server to change algorithm.
7532 * Add load_crls() function to apps tidying load_certs() too. Add option
7533 to verify utility to allow additional CRLs to be included.
7537 * Update OCSP request code to permit adding custom headers to the request:
7538 some responders need this.
7542 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7545 *Julia Lawall <julia@diku.dk>*
7547 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7548 needlessly dereferenced structures, used obsolete functions and
7549 didn't handle all updated verify codes correctly.
7553 * Disable MD2 in the default configuration.
7557 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7558 indicate the initial BIO being pushed or popped. This makes it possible
7559 to determine whether the BIO is the one explicitly called or as a result
7560 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7561 it handles reference counts correctly and doesn't zero out the I/O bio
7562 when it is not being explicitly popped. WARNING: applications which
7563 included workarounds for the old buggy behaviour will need to be modified
7564 or they could free up already freed BIOs.
7568 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7569 renaming to all platforms (within the 0.9.8 branch, this was
7570 done conditionally on Netware platforms to avoid a name clash).
7572 *Guenter <lists@gknw.net>*
7574 * Add ECDHE and PSK support to DTLS.
7576 *Michael Tuexen <tuexen@fh-muenster.de>*
7578 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7583 * Add "missing" function EVP_MD_flags() (without this the only way to
7584 retrieve a digest flags is by accessing the structure directly. Update
7585 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7586 or cipher is registered as in the "from" argument. Print out all
7587 registered digests in the dgst usage message instead of manually
7588 attempting to work them out.
7592 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7593 this allows the use of compression and extensions. Change default cipher
7594 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7595 by default unless an application cipher string requests it.
7599 * Alter match criteria in PKCS12_parse(). It used to try to use local
7600 key ids to find matching certificates and keys but some PKCS#12 files
7601 don't follow the (somewhat unwritten) rules and this strategy fails.
7602 Now just gather all certificates together and the first private key
7603 then look for the first certificate that matches the key.
7607 * Support use of registered digest and cipher names for dgst and cipher
7608 commands instead of having to add each one as a special case. So now
7615 openssl dgst -sha256 foo
7617 and this works for ENGINE based algorithms too.
7621 * Update Gost ENGINE to support parameter files.
7623 *Victor B. Wagner <vitus@cryptocom.ru>*
7625 * Support GeneralizedTime in ca utility.
7627 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7629 * Enhance the hash format used for certificate directory links. The new
7630 form uses the canonical encoding (meaning equivalent names will work
7631 even if they aren't identical) and uses SHA1 instead of MD5. This form
7632 is incompatible with the older format and as a result c_rehash should
7633 be used to rebuild symbolic links.
7637 * Make PKCS#8 the default write format for private keys, replacing the
7638 traditional format. This form is standardised, more secure and doesn't
7639 include an implicit MD5 dependency.
7643 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7644 committed to OpenSSL should pass this lot as a minimum.
7648 * Add session ticket override functionality for use by EAP-FAST.
7650 *Jouni Malinen <j@w1.fi>*
7652 * Modify HMAC functions to return a value. Since these can be implemented
7653 in an ENGINE errors can occur.
7657 * Type-checked OBJ_bsearch_ex.
7661 * Type-checked OBJ_bsearch. Also some constification necessitated
7662 by type-checking. Still to come: TXT_DB, bsearch(?),
7663 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7668 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7669 seconds to a tm structure directly, instead of going through OS
7670 specific date routines. This avoids any issues with OS routines such
7671 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7672 and X509_time_adj_ex() to cover the extended range. The existing
7673 X509_time_adj() is still usable and will no longer have any date issues.
7677 * Delta CRL support. New use deltas option which will attempt to locate
7678 and search any appropriate delta CRLs available.
7680 This work was sponsored by Google.
7684 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7685 code and add additional score elements. Validate alternate CRL paths
7686 as part of the CRL checking and indicate a new error "CRL path validation
7687 error" in this case. Applications wanting additional details can use
7688 the verify callback and check the new "parent" field. If this is not
7689 NULL CRL path validation is taking place. Existing applications won't
7690 see this because it requires extended CRL support which is off by
7693 This work was sponsored by Google.
7697 * Support for freshest CRL extension.
7699 This work was sponsored by Google.
7703 * Initial indirect CRL support. Currently only supported in the CRLs
7704 passed directly and not via lookup. Process certificate issuer
7705 CRL entry extension and lookup CRL entries by bother issuer name
7706 and serial number. Check and process CRL issuer entry in IDP extension.
7708 This work was sponsored by Google.
7712 * Add support for distinct certificate and CRL paths. The CRL issuer
7713 certificate is validated separately in this case. Only enabled if
7714 an extended CRL support flag is set: this flag will enable additional
7715 CRL functionality in future.
7717 This work was sponsored by Google.
7721 * Add support for policy mappings extension.
7723 This work was sponsored by Google.
7727 * Fixes to pathlength constraint, self issued certificate handling,
7728 policy processing to align with RFC3280 and PKITS tests.
7730 This work was sponsored by Google.
7734 * Support for name constraints certificate extension. DN, email, DNS
7735 and URI types are currently supported.
7737 This work was sponsored by Google.
7741 * To cater for systems that provide a pointer-based thread ID rather
7742 than numeric, deprecate the current numeric thread ID mechanism and
7743 replace it with a structure and associated callback type. This
7744 mechanism allows a numeric "hash" to be extracted from a thread ID in
7745 either case, and on platforms where pointers are larger than 'long',
7746 mixing is done to help ensure the numeric 'hash' is usable even if it
7747 can't be guaranteed unique. The default mechanism is to use "&errno"
7748 as a pointer-based thread ID to distinguish between threads.
7750 Applications that want to provide their own thread IDs should now use
7751 CRYPTO_THREADID_set_callback() to register a callback that will call
7752 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7754 Note that ERR_remove_state() is now deprecated, because it is tied
7755 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7756 to free the current thread's error state should be replaced by
7757 ERR_remove_thread_state(NULL).
7759 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7760 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7761 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7762 application was previously providing a numeric thread callback that
7763 was inappropriate for distinguishing threads, then uniqueness might
7764 have been obtained with &errno that happened immediately in the
7765 intermediate development versions of OpenSSL; this is no longer the
7766 case, the numeric thread callback will now override the automatic use
7769 *Geoff Thorpe, with help from Bodo Moeller*
7771 * Initial support for different CRL issuing certificates. This covers a
7772 simple case where the self issued certificates in the chain exist and
7773 the real CRL issuer is higher in the existing chain.
7775 This work was sponsored by Google.
7779 * Removed effectively defunct crypto/store from the build.
7783 * Revamp of STACK to provide stronger type-checking. Still to come:
7784 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7785 ASN1_STRING, CONF_VALUE.
7789 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7790 RAM on SSL connections. This option can save about 34k per idle SSL.
7794 * Revamp of LHASH to provide stronger type-checking. Still to come:
7795 STACK, TXT_DB, bsearch, qsort.
7799 * Initial support for Cryptographic Message Syntax (aka CMS) based
7800 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7801 support for data, signedData, compressedData, digestedData and
7802 encryptedData, envelopedData types included. Scripts to check against
7803 RFC4134 examples draft and interop and consistency checks of many
7804 content types and variants.
7808 * Add options to enc utility to support use of zlib compression BIO.
7812 * Extend mk1mf to support importing of options and assembly language
7813 files from Configure script, currently only included in VC-WIN32.
7814 The assembly language rules can now optionally generate the source
7815 files from the associated perl scripts.
7819 * Implement remaining functionality needed to support GOST ciphersuites.
7820 Interop testing has been performed using CryptoPro implementations.
7822 *Victor B. Wagner <vitus@cryptocom.ru>*
7824 * s390x assembler pack.
7828 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7833 * Implement Opaque PRF Input TLS extension as specified in
7834 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7835 official specification yet and no extension type assignment by
7836 IANA exists, this extension (for now) will have to be explicitly
7837 enabled when building OpenSSL by providing the extension number
7838 to use. For example, specify an option
7840 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7842 to the "config" or "Configure" script to enable the extension,
7843 assuming extension number 0x9527 (which is a completely arbitrary
7844 and unofficial assignment based on the MD5 hash of the Internet
7845 Draft). Note that by doing so, you potentially lose
7846 interoperability with other TLS implementations since these might
7847 be using the same extension number for other purposes.
7849 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7850 opaque PRF input value to use in the handshake. This will create
7851 an internal copy of the length-'len' string at 'src', and will
7852 return non-zero for success.
7854 To get more control and flexibility, provide a callback function
7857 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7858 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7862 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7865 Callback function 'cb' will be called in handshakes, and is
7866 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7867 Argument 'arg' is for application purposes (the value as given to
7868 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7869 be provided to the callback function). The callback function
7870 has to return non-zero to report success: usually 1 to use opaque
7871 PRF input just if possible, or 2 to enforce use of the opaque PRF
7872 input. In the latter case, the library will abort the handshake
7873 if opaque PRF input is not successfully negotiated.
7875 Arguments 'peerinput' and 'len' given to the callback function
7876 will always be NULL and 0 in the case of a client. A server will
7877 see the client's opaque PRF input through these variables if
7878 available (NULL and 0 otherwise). Note that if the server
7879 provides an opaque PRF input, the length must be the same as the
7880 length of the client's opaque PRF input.
7882 Note that the callback function will only be called when creating
7883 a new session (session resumption can resume whatever was
7884 previously negotiated), and will not be called in SSL 2.0
7885 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7886 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7887 for applications that need to enforce opaque PRF input.
7891 * Update ssl code to support digests other than SHA1+MD5 for handshake
7894 *Victor B. Wagner <vitus@cryptocom.ru>*
7896 * Add RFC4507 support to OpenSSL. This includes the corrections in
7897 RFC4507bis. The encrypted ticket format is an encrypted encoded
7898 SSL_SESSION structure, that way new session features are automatically
7901 If a client application caches session in an SSL_SESSION structure
7902 support is transparent because tickets are now stored in the encoded
7905 The SSL_CTX structure automatically generates keys for ticket
7906 protection in servers so again support should be possible
7907 with no application modification.
7909 If a client or server wishes to disable RFC4507 support then the option
7910 SSL_OP_NO_TICKET can be set.
7912 Add a TLS extension debugging callback to allow the contents of any client
7913 or server extensions to be examined.
7915 This work was sponsored by Google.
7919 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7920 OpenSSL should now compile cleanly on gcc 4.2
7922 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7924 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7925 support including streaming MAC support: this is required for GOST
7926 ciphersuite support.
7928 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7930 * Add option -stream to use PKCS#7 streaming in smime utility. New
7931 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7932 to output in BER and PEM format.
7936 * Experimental support for use of HMAC via EVP_PKEY interface. This
7937 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
7938 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7939 ENGINE support for HMAC keys which are unextractable. New -mac and
7940 -macopt options to dgst utility.
7944 * New option -sigopt to dgst utility. Update dgst to use
7945 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
7946 alternative signing parameters such as X9.31 or PSS in the dgst
7951 * Change ssl_cipher_apply_rule(), the internal function that does
7952 the work each time a ciphersuite string requests enabling
7953 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7954 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7955 the order of disabled ciphersuites such that those ciphersuites
7956 that most recently went from enabled to disabled not only stay
7957 in order with respect to each other, but also have higher priority
7958 than other disabled ciphersuites the next time ciphersuites are
7961 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7962 the same ciphersuites as with "HIGH" alone, but in a specific
7963 order where the PSK ciphersuites come first (since they are the
7964 most recently disabled ciphersuites when "HIGH" is parsed).
7966 Also, change ssl_create_cipher_list() (using this new
7967 functionality) such that between otherwise identical
7968 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7973 * Change ssl_create_cipher_list() so that it automatically
7974 arranges the ciphersuites in reasonable order before starting
7975 to process the rule string. Thus, the definition for "DEFAULT"
7976 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7977 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7978 This makes it much easier to arrive at a reasonable default order
7979 in applications for which anonymous ciphers are OK (meaning
7980 that you can't actually use DEFAULT).
7982 *Bodo Moeller; suggested by Victor Duchovni*
7984 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7985 processing) into multiple integers instead of setting
7986 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7987 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7988 (These masks as well as the individual bit definitions are hidden
7989 away into the non-exported interface ssl/ssl_locl.h, so this
7990 change to the definition of the SSL_CIPHER structure shouldn't
7991 affect applications.) This give us more bits for each of these
7992 categories, so there is no longer a need to coagulate AES128 and
7993 AES256 into a single algorithm bit, and to coagulate Camellia128
7994 and Camellia256 into a single algorithm bit, which has led to all
7997 Thus, among other things, the kludge introduced in 0.9.7m and
7998 0.9.8e for masking out AES256 independently of AES128 or masking
7999 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8001 With the change, we also introduce new ciphersuite aliases that
8002 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8007 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8008 Use the leftmost N bytes of the signature input if the input is
8009 larger than the prime q (with N being the size in bytes of q).
8013 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8014 it yet and it is largely untested.
8018 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8022 * Initial incomplete changes to avoid need for function casts in OpenSSL
8023 some compilers (gcc 4.2 and later) reject their use. Safestack is
8024 reimplemented. Update ASN1 to avoid use of legacy functions.
8028 * Win32/64 targets are linked with Winsock2.
8032 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8033 to external functions. This can be used to increase CRL handling
8034 efficiency especially when CRLs are very large by (for example) storing
8035 the CRL revoked certificates in a database.
8039 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8040 new CRLs added to a directory can be used. New command line option
8041 -verify_return_error to s_client and s_server. This causes real errors
8042 to be returned by the verify callback instead of carrying on no matter
8043 what. This reflects the way a "real world" verify callback would behave.
8047 * GOST engine, supporting several GOST algorithms and public key formats.
8048 Kindly donated by Cryptocom.
8052 * Partial support for Issuing Distribution Point CRL extension. CRLs
8053 partitioned by DP are handled but no indirect CRL or reason partitioning
8054 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8055 selected via a scoring technique which handles IDP and AKID in CRLs.
8059 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8060 will ultimately be used for all verify operations: this will remove the
8061 X509_STORE dependency on certificate verification and allow alternative
8062 lookup methods. X509_STORE based implementations of these two callbacks.
8066 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8067 Modify get_crl() to find a valid (unexpired) CRL if possible.
8071 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8072 this would be called X509_CRL_cmp() but that name is already used by
8073 a function that just compares CRL issuer names. Cache several CRL
8074 extensions in X509_CRL structure and cache CRLDP in X509.
8078 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8079 this maps equivalent X509_NAME structures into a consistent structure.
8080 Name comparison can then be performed rapidly using memcmp().
8084 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8089 * Allow digests to supply their own micalg string for S/MIME type using
8090 the ctrl EVP_MD_CTRL_MICALG.
8094 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8095 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8096 ctrl. It can then customise the structure before and/or after signing
8101 * New function OBJ_add_sigid() to allow application defined signature OIDs
8102 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8103 to free up any added signature OIDs.
8107 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8108 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8109 digest and cipher tables. New options added to openssl utility:
8110 list-message-digest-algorithms and list-cipher-algorithms.
8114 * Change the array representation of binary polynomials: the list
8115 of degrees of non-zero coefficients is now terminated with -1.
8116 Previously it was terminated with 0, which was also part of the
8117 value; thus, the array representation was not applicable to
8118 polynomials where t^0 has coefficient zero. This change makes
8119 the array representation useful in a more general context.
8123 * Various modifications and fixes to SSL/TLS cipher string
8124 handling. For ECC, the code now distinguishes between fixed ECDH
8125 with RSA certificates on the one hand and with ECDSA certificates
8126 on the other hand, since these are separate ciphersuites. The
8127 unused code for Fortezza ciphersuites has been removed.
8129 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8130 (not "ECDHE"). For consistency with the code for DH
8131 certificates, use of ECDH certificates is now considered ECDH
8132 authentication, not RSA or ECDSA authentication (the latter is
8133 merely the CA's signing algorithm and not actively used in the
8136 The temporary ciphersuite alias "ECCdraft" is no longer
8137 available, and ECC ciphersuites are no longer excluded from "ALL"
8138 and "DEFAULT". The following aliases now exist for RFC 4492
8139 ciphersuites, most of these by analogy with the DH case:
8141 kECDHr - ECDH cert, signed with RSA
8142 kECDHe - ECDH cert, signed with ECDSA
8143 kECDH - ECDH cert (signed with either RSA or ECDSA)
8144 kEECDH - ephemeral ECDH
8145 ECDH - ECDH cert or ephemeral ECDH
8151 AECDH - anonymous ECDH
8152 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8156 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8157 Use correct micalg parameters depending on digest(s) in signed message.
8161 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8162 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8166 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8167 an engine to register a method. Add ENGINE lookups for methods and
8168 functional reference processing.
8172 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8173 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8178 * New -resign option to smime utility. This adds one or more signers
8179 to an existing PKCS#7 signedData structure. Also -md option to use an
8180 alternative message digest algorithm for signing.
8184 * Tidy up PKCS#7 routines and add new functions to make it easier to
8185 create PKCS7 structures containing multiple signers. Update smime
8186 application to support multiple signers.
8190 * New -macalg option to pkcs12 utility to allow setting of an alternative
8195 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8196 Reorganize PBE internals to lookup from a static table using NIDs,
8197 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8198 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8199 PRF which will be automatically used with PBES2.
8203 * Replace the algorithm specific calls to generate keys in "req" with the
8208 * Update PKCS#7 enveloped data routines to use new API. This is now
8209 supported by any public key method supporting the encrypt operation. A
8210 ctrl is added to allow the public key algorithm to examine or modify
8211 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8216 * Add a ctrl to asn1 method to allow a public key algorithm to express
8217 a default digest type to use. In most cases this will be SHA1 but some
8218 algorithms (such as GOST) need to specify an alternative digest. The
8219 return value indicates how strong the preference is 1 means optional and
8220 2 is mandatory (that is it is the only supported type). Modify
8221 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8222 use the default md. Update openssl utilities to use the default digest
8223 type for signing if it is not explicitly indicated.
8227 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8228 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8229 signing method from the key type. This effectively removes the link
8230 between digests and public key types.
8234 * Add an OID cross reference table and utility functions. Its purpose is to
8235 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8236 rsaEncryption. This will allow some of the algorithm specific hackery
8237 needed to use the correct OID to be removed.
8241 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8242 structures for PKCS7_sign(). They are now set up by the relevant public
8247 * Add provisional EC pkey method with support for ECDSA and ECDH.
8251 * Add support for key derivation (agreement) in the API, DH method and
8256 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8257 public and private key formats. As a side effect these add additional
8258 command line functionality not previously available: DSA signatures can be
8259 generated and verified using pkeyutl and DH key support and generation in
8266 *Oliver Tappe <zooey@hirschkaefer.de>*
8268 * New make target "install_html_docs" installs HTML renditions of the
8271 *Oliver Tappe <zooey@hirschkaefer.de>*
8273 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8274 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8275 support key and parameter generation and add initial key generation
8276 functionality for RSA.
8280 * Add functions for main EVP_PKEY_method operations. The undocumented
8281 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8282 `EVP_PKEY_{encrypt,decrypt}_old`.
8286 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8287 key API, doesn't do much yet.
8291 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8292 public key algorithms. New option to openssl utility:
8293 "list-public-key-algorithms" to print out info.
8297 * Implement the Supported Elliptic Curves Extension for
8298 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8302 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8303 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8307 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8308 utilities such as rsa, dsa, dsaparam etc except they process any key
8313 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8314 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8315 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8320 * Initial support for pluggable public key ASN1.
8321 De-spaghettify the public key ASN1 handling. Move public and private
8322 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8323 algorithm specific handling to a single module within the relevant
8324 algorithm directory. Add functions to allow (near) opaque processing
8325 of public and private key structures.
8329 * Implement the Supported Point Formats Extension for
8330 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8334 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8335 for the psk identity [hint] and the psk callback functions to the
8336 SSL_SESSION, SSL and SSL_CTX structure.
8339 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8343 SSL_CTX_use_psk_identity_hint
8344 SSL_get_psk_identity_hint
8345 SSL_get_psk_identity
8346 SSL_use_psk_identity_hint
8348 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8350 * Add RFC 3161 compliant time stamp request creation, response generation
8351 and response verification functionality.
8353 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8355 * Add initial support for TLS extensions, specifically for the server_name
8356 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8357 have new members for a host name. The SSL data structure has an
8358 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8359 stored in that context to allow for session resumption, even after the
8360 SSL has been switched to a new SSL_CTX in reaction to a client's
8361 server_name extension.
8363 New functions (subject to change):
8365 SSL_get_servername()
8366 SSL_get_servername_type()
8369 New CTRL codes and macros (subject to change):
8371 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8372 - SSL_CTX_set_tlsext_servername_callback()
8373 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8374 - SSL_CTX_set_tlsext_servername_arg()
8375 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8377 openssl s_client has a new '-servername ...' option.
8379 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8380 '-key2 ...', '-servername_fatal' (subject to change). This allows
8381 testing the HostName extension for a specific single host name ('-cert'
8382 and '-key' remain fallbacks for handshakes without HostName
8383 negotiation). If the unrecognized_name alert has to be sent, this by
8384 default is a warning; it becomes fatal with the '-servername_fatal'
8387 *Peter Sylvester, Remy Allais, Christophe Renou*
8389 * Whirlpool hash implementation is added.
8393 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8394 bn(64,32). Because of instruction set limitations it doesn't have
8395 any negative impact on performance. This was done mostly in order
8396 to make it possible to share assembler modules, such as bn_mul_mont
8397 implementations, between 32- and 64-bit builds without hassle.
8401 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8402 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8407 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8408 dedicated Montgomery multiplication procedure, is introduced.
8409 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8410 "64-bit" performance on certain 32-bit targets.
8414 * New option SSL_OP_NO_COMP to disable use of compression selectively
8415 in SSL structures. New SSL ctrl to set maximum send fragment size.
8416 Save memory by setting the I/O buffer sizes dynamically instead of
8417 using the maximum available value.
8421 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8422 in addition to the text details.
8426 * Very, very preliminary EXPERIMENTAL support for printing of general
8427 ASN1 structures. This currently produces rather ugly output and doesn't
8428 handle several customised structures at all.
8432 * Integrated support for PVK file format and some related formats such
8433 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8434 these in the 'rsa' and 'dsa' utilities.
8438 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8442 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8443 place for the (very old) "NETSCAPE" format certificates which are now
8444 handled using new ASN1 code equivalents.
8448 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8449 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8450 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8454 * Modify CRL distribution points extension code to print out previously
8455 unsupported fields. Enhance extension setting code to allow setting of
8460 * Add print and set support for Issuing Distribution Point CRL extension.
8464 * Change 'Configure' script to enable Camellia by default.
8471 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8473 * When rejecting SSL/TLS records due to an incorrect version number, never
8474 update s->server with a new major version number. As of
8475 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8476 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8477 the previous behavior could result in a read attempt at NULL when
8478 receiving specific incorrect SSL/TLS records once record payload
8479 protection is active. [CVE-2010-0740][]
8481 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8483 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8484 could be crashed if the relevant tables were not present (e.g. chrooted).
8486 *Tomas Hoger <thoger@redhat.com>*
8488 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8490 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8492 *Martin Olsson, Neel Mehta*
8494 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8495 accommodate for stack sorting, always a write lock!).
8499 * On some versions of WIN32 Heap32Next is very slow. This can cause
8500 excessive delays in the RAND_poll(): over a minute. As a workaround
8501 include a time check in the inner Heap32Next loop too.
8505 * The code that handled flushing of data in SSL/TLS originally used the
8506 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8507 the problem outlined in PR#1949. The fix suggested there however can
8508 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8509 of Apache). So instead simplify the code to flush unconditionally.
8510 This should be fine since flushing with no data to flush is a no op.
8514 * Handle TLS versions 2.0 and later properly and correctly use the
8515 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8516 off ancient servers have a habit of sticking around for a while...
8520 * Modify compression code so it frees up structures without using the
8521 ex_data callbacks. This works around a problem where some applications
8522 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8523 restarting) then use compression (e.g. SSL with compression) later.
8524 This results in significant per-connection memory leaks and
8525 has caused some security issues including CVE-2008-1678 and
8530 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8531 change when encrypting or decrypting.
8535 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8536 connect and renegotiate with servers which do not support RI.
8537 Until RI is more widely deployed this option is enabled by default.
8541 * Add "missing" ssl ctrls to clear options and mode.
8545 * If client attempts to renegotiate and doesn't support RI respond with
8546 a no_renegotiation alert as required by RFC5746. Some renegotiating
8547 TLS clients will continue a connection gracefully when they receive
8548 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8549 waiting for a server hello which it will never receive. Now we treat a
8550 received no_renegotiation alert as a fatal error. This is because
8551 applications requesting a renegotiation might well expect it to succeed
8552 and would have no code in place to handle the server denying it so the
8553 only safe thing to do is to terminate the connection.
8557 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8558 peer supports secure renegotiation and 0 otherwise. Print out peer
8559 renegotiation support in s_client/s_server.
8563 * Replace the highly broken and deprecated SPKAC certification method with
8564 the updated NID creation version. This should correctly handle UTF8.
8568 * Implement RFC5746. Re-enable renegotiation but require the extension
8569 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8570 turns out to be a bad idea. It has been replaced by
8571 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8572 SSL_CTX_set_options(). This is really not recommended unless you
8573 know what you are doing.
8575 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8577 * Fixes to stateless session resumption handling. Use initial_ctx when
8578 issuing and attempting to decrypt tickets in case it has changed during
8579 servername handling. Use a non-zero length session ID when attempting
8580 stateless session resumption: this makes it possible to determine if
8581 a resumption has occurred immediately after receiving server hello
8582 (several places in OpenSSL subtly assume this) instead of later in
8587 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8588 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8589 fixes for a few places where the return code is not checked
8592 *Julia Lawall <julia@diku.dk>*
8594 * Add --strict-warnings option to Configure script to include devteam
8595 warnings in other configurations.
8599 * Add support for --libdir option and LIBDIR variable in makefiles. This
8600 makes it possible to install openssl libraries in locations which
8601 have names other than "lib", for example "/usr/lib64" which some
8604 *Steve Henson, based on patch from Jeremy Utley*
8606 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8607 X690 8.9.12 and can produce some misleading textual output of OIDs.
8609 *Steve Henson, reported by Dan Kaminsky*
8611 * Delete MD2 from algorithm tables. This follows the recommendation in
8612 several standards that it is not used in new applications due to
8613 several cryptographic weaknesses. For binary compatibility reasons
8614 the MD2 API is still compiled in by default.
8618 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8623 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8624 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8627 *Guenter <lists@gknw.net>*
8629 * Fix the server certificate chain building code to use X509_verify_cert(),
8630 it used to have an ad-hoc builder which was unable to cope with anything
8631 other than a simple chain.
8633 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8635 * Don't check self signed certificate signatures in X509_verify_cert()
8636 by default (a flag can override this): it just wastes time without
8637 adding any security. As a useful side effect self signed root CAs
8638 with non-FIPS digests are now usable in FIPS mode.
8642 * In dtls1_process_out_of_seq_message() the check if the current message
8643 is already buffered was missing. For every new message was memory
8644 allocated, allowing an attacker to perform an denial of service attack
8645 with sending out of seq handshake messages until there is no memory
8646 left. Additionally every future message was buffered, even if the
8647 sequence number made no sense and would be part of another handshake.
8648 So only messages with sequence numbers less than 10 in advance will be
8649 buffered. [CVE-2009-1378][]
8651 *Robin Seggelmann, discovered by Daniel Mentz*
8653 * Records are buffered if they arrive with a future epoch to be
8654 processed after finishing the corresponding handshake. There is
8655 currently no limitation to this buffer allowing an attacker to perform
8656 a DOS attack with sending records with future epochs until there is no
8657 memory left. This patch adds the pqueue_size() function to determine
8658 the size of a buffer and limits the record buffer to 100 entries.
8661 *Robin Seggelmann, discovered by Daniel Mentz*
8663 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8664 parent structure is freed. [CVE-2009-1379][]
8668 * Handle non-blocking I/O properly in SSL_shutdown() call.
8670 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8672 * Add `2.5.4.*` OIDs
8674 *Ilya O. <vrghost@gmail.com>*
8676 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8678 * Disable renegotiation completely - this fixes a severe security
8679 problem [CVE-2009-3555][] at the cost of breaking all
8680 renegotiation. Renegotiation can be re-enabled by setting
8681 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8682 run-time. This is really not recommended unless you know what
8687 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8689 * Don't set val to NULL when freeing up structures, it is freed up by
8690 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8691 zeroing past the valid field. [CVE-2009-0789][]
8693 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8695 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8696 checked correctly. This would allow some invalid signed attributes to
8697 appear to verify correctly. [CVE-2009-0591][]
8699 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8701 * Reject UniversalString and BMPString types with invalid lengths. This
8702 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8703 a legal length. [CVE-2009-0590][]
8707 * Set S/MIME signing as the default purpose rather than setting it
8708 unconditionally. This allows applications to override it at the store
8713 * Permit restricted recursion of ASN1 strings. This is needed in practice
8714 to handle some structures.
8718 * Improve efficiency of mem_gets: don't search whole buffer each time
8721 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8723 * New -hex option for openssl rand.
8727 * Print out UTF8String and NumericString when parsing ASN1.
8731 * Support NumericString type for name components.
8735 * Allow CC in the environment to override the automatically chosen
8736 compiler. Note that nothing is done to ensure flags work with the
8741 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8743 * Properly check EVP_VerifyFinal() and similar return values
8746 *Ben Laurie, Bodo Moeller, Google Security Team*
8748 * Enable TLS extensions by default.
8752 * Allow the CHIL engine to be loaded, whether the application is
8753 multithreaded or not. (This does not release the developer from the
8754 obligation to set up the dynamic locking callbacks.)
8756 *Sander Temme <sander@temme.net>*
8758 * Use correct exit code if there is an error in dgst command.
8760 *Steve Henson; problem pointed out by Roland Dirlewanger*
8762 * Tweak Configure so that you need to say "experimental-jpake" to enable
8763 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8767 * Add experimental JPAKE support, including demo authentication in
8768 s_client and s_server.
8772 * Set the comparison function in v3_addr_canonize().
8774 *Rob Austein <sra@hactrn.net>*
8776 * Add support for XMPP STARTTLS in s_client.
8778 *Philip Paeps <philip@freebsd.org>*
8780 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8781 to ensure that even with this option, only ciphersuites in the
8782 server's preference list will be accepted. (Note that the option
8783 applies only when resuming a session, so the earlier behavior was
8784 just about the algorithm choice for symmetric cryptography.)
8788 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8790 * Fix NULL pointer dereference if a DTLS server received
8791 ChangeCipherSpec as first record [CVE-2009-1386][].
8795 * Fix a state transition in s3_srvr.c and d1_srvr.c
8796 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8800 * The fix in 0.9.8c that supposedly got rid of unsafe
8801 double-checked locking was incomplete for RSA blinding,
8802 addressing just one layer of what turns out to have been
8803 doubly unsafe triple-checked locking.
8805 So now fix this for real by retiring the MONT_HELPER macro
8806 in crypto/rsa/rsa_eay.c.
8808 *Bodo Moeller; problem pointed out by Marius Schilder*
8810 * Various precautionary measures:
8812 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8814 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8815 (NB: This would require knowledge of the secret session ticket key
8816 to exploit, in which case you'd be SOL either way.)
8818 - Change bn_nist.c so that it will properly handle input BIGNUMs
8819 outside the expected range.
8821 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8824 *Neel Mehta, Bodo Moeller*
8826 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8827 the load fails. Useful for distros.
8829 *Ben Laurie and the FreeBSD team*
8831 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8835 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8839 * Expand ENGINE to support engine supplied SSL client certificate functions.
8841 This work was sponsored by Logica.
8845 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8846 keystores. Support for SSL/TLS client authentication too.
8847 Not compiled unless enable-capieng specified to Configure.
8849 This work was sponsored by Logica.
8853 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8854 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8855 attribute creation routines such as certificate requests and PKCS#12
8860 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8862 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8863 handshake which could lead to a client crash as found using the
8864 Codenomicon TLS test suite [CVE-2008-1672][]
8866 *Steve Henson, Mark Cox*
8868 * Fix double free in TLS server name extensions which could lead to
8869 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8873 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8875 Clear the error queue to ensure that error entries left from
8876 older function calls do not interfere with the correct operation.
8878 *Lutz Jaenicke, Erik de Castro Lopo*
8880 * Remove root CA certificates of commercial CAs:
8882 The OpenSSL project does not recommend any specific CA and does not
8883 have any policy with respect to including or excluding any CA.
8884 Therefore it does not make any sense to ship an arbitrary selection
8885 of root CA certificates with the OpenSSL software.
8889 * RSA OAEP patches to fix two separate invalid memory reads.
8890 The first one involves inputs when 'lzero' is greater than
8891 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8892 before the beginning of from). The second one involves inputs where
8893 the 'db' section contains nothing but zeroes (there is a one-byte
8894 invalid read after the end of 'db').
8896 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8898 * Partial backport from 0.9.9-dev:
8900 Introduce bn_mul_mont (dedicated Montgomery multiplication
8901 procedure) as a candidate for BIGNUM assembler implementation.
8902 While 0.9.9-dev uses assembler for various architectures, only
8903 x86_64 is available by default here in the 0.9.8 branch, and
8904 32-bit x86 is available through a compile-time setting.
8906 To try the 32-bit x86 assembler implementation, use Configure
8907 option "enable-montasm" (which exists only for this backport).
8909 As "enable-montasm" for 32-bit x86 disclaims code stability
8910 anyway, in this constellation we activate additional code
8911 backported from 0.9.9-dev for further performance improvements,
8912 namely BN_from_montgomery_word. (To enable this otherwise,
8913 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8915 *Andy Polyakov (backport partially by Bodo Moeller)*
8917 * Add TLS session ticket callback. This allows an application to set
8918 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8919 values. This is useful for key rollover for example where several key
8920 sets may exist with different names.
8924 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8925 This was broken until now in 0.9.8 releases, such that the only way
8926 a registered ENGINE could be used (assuming it initialises
8927 successfully on the host) was to explicitly set it as the default
8928 for the relevant algorithms. This is in contradiction with 0.9.7
8929 behaviour and the documentation. With this fix, when an ENGINE is
8930 registered into a given algorithm's table of implementations, the
8931 'uptodate' flag is reset so that auto-discovery will be used next
8932 time a new context for that algorithm attempts to select an
8935 *Ian Lister (tweaked by Geoff Thorpe)*
8937 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8938 implementation in the following ways:
8940 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8943 Lack of BER streaming support means one pass streaming processing is
8944 only supported if data is detached: setting the streaming flag is
8945 ignored for embedded content.
8947 CMS support is disabled by default and must be explicitly enabled
8948 with the enable-cms configuration option.
8952 * Update the GMP engine glue to do direct copies between BIGNUM and
8953 mpz_t when openssl and GMP use the same limb size. Otherwise the
8954 existing "conversion via a text string export" trick is still used.
8956 *Paul Sheer <paulsheer@gmail.com>*
8958 * Zlib compression BIO. This is a filter BIO which compressed and
8959 uncompresses any data passed through it.
8963 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8964 RFC3394 compatible AES key wrapping.
8968 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8969 sets string data without copying. X509_ALGOR_set0() and
8970 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8971 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8972 from an X509_ATTRIBUTE structure optionally checking it occurs only
8973 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8978 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8979 to get the expected BN_FLG_CONSTTIME behavior.
8981 *Bodo Moeller (Google)*
8985 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8986 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8987 - added some more tests to do_tests.pl
8988 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8989 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8990 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8991 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8992 - various changes to netware.pl to enable gcc-cross builds on Win32
8994 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8995 - various changes to fix missing prototype warnings
8996 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8997 - added AES, WHIRLPOOL and CPUID assembler code to build files
8998 - added missing AES assembler make rules to mk1mf.pl
8999 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9001 *Guenter Knauf <eflash@gmx.net>*
9003 * Implement certificate status request TLS extension defined in RFC3546.
9004 A client can set the appropriate parameters and receive the encoded
9005 OCSP response via a callback. A server can query the supplied parameters
9006 and set the encoded OCSP response in the callback. Add simplified examples
9007 to s_client and s_server.
9011 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9014 + Binary incompatibility of ssl_ctx_st structure
9015 + DTLS interoperation with non-compliant servers
9016 + Don't call get_session_cb() without proposed session
9017 + Fix ia64 assembler code
9019 *Andy Polyakov, Steve Henson*
9021 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9023 * DTLS Handshake overhaul. There were longstanding issues with
9024 OpenSSL DTLS implementation, which were making it impossible for
9025 RFC 4347 compliant client to communicate with OpenSSL server.
9026 Unfortunately just fixing these incompatibilities would "cut off"
9027 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9028 server keeps tolerating non RFC compliant syntax. The opposite is
9029 not true, 0.9.8f client can not communicate with earlier server.
9030 This update even addresses CVE-2007-4995.
9034 * Changes to avoid need for function casts in OpenSSL: some compilers
9035 (gcc 4.2 and later) reject their use.
9036 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9039 * Add RFC4507 support to OpenSSL. This includes the corrections in
9040 RFC4507bis. The encrypted ticket format is an encrypted encoded
9041 SSL_SESSION structure, that way new session features are automatically
9044 If a client application caches session in an SSL_SESSION structure
9045 support is transparent because tickets are now stored in the encoded
9048 The SSL_CTX structure automatically generates keys for ticket
9049 protection in servers so again support should be possible
9050 with no application modification.
9052 If a client or server wishes to disable RFC4507 support then the option
9053 SSL_OP_NO_TICKET can be set.
9055 Add a TLS extension debugging callback to allow the contents of any client
9056 or server extensions to be examined.
9058 This work was sponsored by Google.
9062 * Add initial support for TLS extensions, specifically for the server_name
9063 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9064 have new members for a host name. The SSL data structure has an
9065 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9066 stored in that context to allow for session resumption, even after the
9067 SSL has been switched to a new SSL_CTX in reaction to a client's
9068 server_name extension.
9070 New functions (subject to change):
9072 SSL_get_servername()
9073 SSL_get_servername_type()
9076 New CTRL codes and macros (subject to change):
9078 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9079 - SSL_CTX_set_tlsext_servername_callback()
9080 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9081 - SSL_CTX_set_tlsext_servername_arg()
9082 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9084 openssl s_client has a new '-servername ...' option.
9086 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9087 '-key2 ...', '-servername_fatal' (subject to change). This allows
9088 testing the HostName extension for a specific single host name ('-cert'
9089 and '-key' remain fallbacks for handshakes without HostName
9090 negotiation). If the unrecognized_name alert has to be sent, this by
9091 default is a warning; it becomes fatal with the '-servername_fatal'
9094 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9096 * Add AES and SSE2 assembly language support to VC++ build.
9100 * Mitigate attack on final subtraction in Montgomery reduction.
9104 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9105 (which previously caused an internal error).
9109 * Squeeze another 10% out of IGE mode when in != out.
9113 * AES IGE mode speedup.
9115 *Dean Gaudet (Google)*
9117 * Add the Korean symmetric 128-bit cipher SEED (see
9118 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9119 add SEED ciphersuites from RFC 4162:
9121 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9122 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9123 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9124 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9126 To minimize changes between patchlevels in the OpenSSL 0.9.8
9127 series, SEED remains excluded from compilation unless OpenSSL
9128 is configured with 'enable-seed'.
9130 *KISA, Bodo Moeller*
9132 * Mitigate branch prediction attacks, which can be practical if a
9133 single processor is shared, allowing a spy process to extract
9134 information. For detailed background information, see
9135 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9136 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9137 and Necessary Software Countermeasures"). The core of the change
9138 are new versions BN_div_no_branch() and
9139 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9140 respectively, which are slower, but avoid the security-relevant
9141 conditional branches. These are automatically called by BN_div()
9142 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9143 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9144 remove a conditional branch.
9146 BN_FLG_CONSTTIME is the new name for the previous
9147 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9148 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9149 in the exponent causes BN_mod_exp_mont() to use the alternative
9150 implementation in BN_mod_exp_mont_consttime().) The old name
9151 remains as a deprecated alias.
9153 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9154 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9155 constant-time implementations for more than just exponentiation.
9156 Here too the old name is kept as a deprecated alias.
9158 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9159 the BN_BLINDING structure gets an independent copy of the
9160 modulus. This means that the previous `BIGNUM *m` argument to
9161 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9162 essentially becomes `const BIGNUM *m`, although we can't actually
9163 change this in the header file before 0.9.9. It allows
9164 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9165 enable BN_FLG_CONSTTIME.
9167 *Matthew D Wood (Intel Corp)*
9169 * In the SSL/TLS server implementation, be strict about session ID
9170 context matching (which matters if an application uses a single
9171 external cache for different purposes). Previously,
9172 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9173 set. This did ensure strict client verification, but meant that,
9174 with applications using a single external cache for quite
9175 different requirements, clients could circumvent ciphersuite
9176 restrictions for a given session ID context by starting a session
9177 in a different context.
9181 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9182 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9183 authentication-only ciphersuites.
9187 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9188 not complete and could lead to a possible single byte overflow
9189 [CVE-2007-5135][] [Ben Laurie]
9191 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9193 * Since AES128 and AES256 (and similarly Camellia128 and
9194 Camellia256) share a single mask bit in the logic of
9195 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9196 kludge to work properly if AES128 is available and AES256 isn't
9197 (or if Camellia128 is available and Camellia256 isn't).
9201 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9202 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9203 When a point or a seed is encoded in a BIT STRING, we need to
9204 prevent the removal of trailing zero bits to get the proper DER
9205 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9206 of a NamedBitList, for which trailing 0 bits need to be removed.)
9210 * Have SSL/TLS server implementation tolerate "mismatched" record
9211 protocol version while receiving ClientHello even if the
9212 ClientHello is fragmented. (The server can't insist on the
9213 particular protocol version it has chosen before the ServerHello
9214 message has informed the client about his choice.)
9218 * Add RFC 3779 support.
9220 *Rob Austein for ARIN, Ben Laurie*
9222 * Load error codes if they are not already present instead of using a
9223 static variable. This allows them to be cleanly unloaded and reloaded.
9224 Improve header file function name parsing.
9228 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9229 or CAPABILITY handshake as required by RFCs.
9233 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9235 * Introduce limits to prevent malicious keys being able to
9236 cause a denial of service. [CVE-2006-2940][]
9238 *Steve Henson, Bodo Moeller*
9240 * Fix ASN.1 parsing of certain invalid structures that can result
9241 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9243 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9244 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9246 * Fix SSL client code which could crash if connecting to a
9247 malicious SSLv2 server. [CVE-2006-4343][]
9249 *Tavis Ormandy and Will Drewry, Google Security Team*
9251 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9252 match only those. Before that, "AES256-SHA" would be interpreted
9253 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9254 the same strength classification in 0.9.7h) as we currently only
9255 have a single AES bit in the ciphersuite description bitmap.
9256 That change, however, also applied to ciphersuite strings such as
9257 "RC4-MD5" that intentionally matched multiple ciphersuites --
9258 namely, SSL 2.0 ciphersuites in addition to the more common ones
9259 from SSL 3.0/TLS 1.0.
9261 So we change the selection algorithm again: Naming an explicit
9262 ciphersuite selects this one ciphersuite, and any other similar
9263 ciphersuite (same bitmap) from *other* protocol versions.
9264 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9265 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9267 Since SSL 2.0 does not have any ciphersuites for which the
9268 128/256 bit distinction would be relevant, this works for now.
9269 The proper fix will be to use different bits for AES128 and
9270 AES256, which would have avoided the problems from the beginning;
9271 however, bits are scarce, so we can only do this in a new release
9272 (not just a patchlevel) when we can change the SSL_CIPHER
9273 definition to split the single 'unsigned long mask' bitmap into
9274 multiple values to extend the available space.
9278 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9280 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9281 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9283 * Add AES IGE and biIGE modes.
9287 * Change the Unix randomness entropy gathering to use poll() when
9288 possible instead of select(), since the latter has some
9289 undesirable limitations.
9291 *Darryl Miles via Richard Levitte and Bodo Moeller*
9293 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9294 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9295 cannot be implicitly activated as part of, e.g., the "AES" alias.
9296 However, please upgrade to OpenSSL 0.9.9[-dev] for
9297 non-experimental use of the ECC ciphersuites to get TLS extension
9298 support, which is required for curve and point format negotiation
9299 to avoid potential handshake problems.
9303 * Disable rogue ciphersuites:
9305 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9306 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9307 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9309 The latter two were purportedly from
9310 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9313 Also deactivate the remaining ciphersuites from
9314 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9315 unofficial, and the ID has long expired.
9319 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9320 dual-core machines) and other potential thread-safety issues.
9324 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9325 versions), which is now available for royalty-free use
9326 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9327 Also, add Camellia TLS ciphersuites from RFC 4132.
9329 To minimize changes between patchlevels in the OpenSSL 0.9.8
9330 series, Camellia remains excluded from compilation unless OpenSSL
9331 is configured with 'enable-camellia'.
9335 * Disable the padding bug check when compression is in use. The padding
9336 bug check assumes the first packet is of even length, this is not
9337 necessarily true if compression is enabled and can result in false
9338 positives causing handshake failure. The actual bug test is ancient
9339 code so it is hoped that implementations will either have fixed it by
9340 now or any which still have the bug do not support compression.
9344 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9346 * When applying a cipher rule check to see if string match is an explicit
9347 cipher suite and only match that one cipher suite if it is.
9351 * Link in manifests for VC++ if needed.
9353 *Austin Ziegler <halostatue@gmail.com>*
9355 * Update support for ECC-based TLS ciphersuites according to
9356 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9357 TLS extensions, which are supported starting with the 0.9.9
9358 branch, not in the OpenSSL 0.9.8 branch).
9362 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9363 opaque EVP_CIPHER_CTX handling.
9367 * Fixes and enhancements to zlib compression code. We now only use
9368 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9369 to conform with the standards mentioned here:
9370 <http://www.zlib.net/DLL_FAQ.txt>
9371 Static zlib linking now works on Windows and the new --with-zlib-include
9372 --with-zlib-lib options to Configure can be used to supply the location
9373 of the headers and library. Gracefully handle case where zlib library
9378 * Several fixes and enhancements to the OID generation code. The old code
9379 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9380 handle numbers larger than ULONG_MAX, truncated printing and had a
9381 non standard OBJ_obj2txt() behaviour.
9385 * Add support for building of engines under engine/ as shared libraries
9386 under VC++ build system.
9390 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9391 Hopefully, we will not see any false combination of paths any more.
9395 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9397 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9398 (part of SSL_OP_ALL). This option used to disable the
9399 countermeasure against man-in-the-middle protocol-version
9400 rollback in the SSL 2.0 server implementation, which is a bad
9401 idea. [CVE-2005-2969][]
9403 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9404 for Information Security, National Institute of Advanced Industrial
9405 Science and Technology [AIST], Japan)*
9407 * Add two function to clear and return the verify parameter flags.
9411 * Keep cipherlists sorted in the source instead of sorting them at
9412 runtime, thus removing the need for a lock.
9416 * Avoid some small subgroup attacks in Diffie-Hellman.
9418 *Nick Mathewson and Ben Laurie*
9420 * Add functions for well-known primes.
9424 * Extended Windows CE support.
9426 *Satoshi Nakamura and Andy Polyakov*
9428 * Initialize SSL_METHOD structures at compile time instead of during
9429 runtime, thus removing the need for a lock.
9433 * Make PKCS7_decrypt() work even if no certificate is supplied by
9434 attempting to decrypt each encrypted key in turn. Add support to
9439 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9441 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9444 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9448 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9449 key into the same file any more.
9453 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9457 * Add -utf8 command line and config file option to 'ca'.
9459 *Stefan <stf@udoma.org*
9461 * Removed the macro des_crypt(), as it seems to conflict with some
9462 libraries. Use DES_crypt().
9466 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9467 involves renaming the source and generated shared-libs for
9468 both. The engines will accept the corrected or legacy ids
9469 ('ncipher' and '4758_cca' respectively) when binding. NB,
9470 this only applies when building 'shared'.
9472 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9474 * Add attribute functions to EVP_PKEY structure. Modify
9475 PKCS12_create() to recognize a CSP name attribute and
9476 use it. Make -CSP option work again in pkcs12 utility.
9480 * Add new functionality to the bn blinding code:
9481 - automatic re-creation of the BN_BLINDING parameters after
9482 a fixed number of uses (currently 32)
9483 - add new function for parameter creation
9484 - introduce flags to control the update behaviour of the
9485 BN_BLINDING parameters
9486 - hide BN_BLINDING structure
9487 Add a second BN_BLINDING slot to the RSA structure to improve
9488 performance when a single RSA object is shared among several
9493 * Add support for DTLS.
9495 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9497 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9498 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9502 * Remove buggy and incomplete DH cert support from
9503 ssl/ssl_rsa.c and ssl/s3_both.c
9507 * Use SHA-1 instead of MD5 as the default digest algorithm for
9508 the apps/openssl applications.
9512 * Compile clean with "-Wall -Wmissing-prototypes
9513 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9514 DEBUG_SAFESTACK must also be set.
9518 * Change ./Configure so that certain algorithms can be disabled by default.
9519 The new counterpiece to "no-xxx" is "enable-xxx".
9521 The patented RC5 and MDC2 algorithms will now be disabled unless
9522 "enable-rc5" and "enable-mdc2", respectively, are specified.
9524 (IDEA remains enabled despite being patented. This is because IDEA
9525 is frequently required for interoperability, and there is no license
9526 fee for non-commercial use. As before, "no-idea" can be used to
9527 avoid this algorithm.)
9531 * Add processing of proxy certificates (see RFC 3820). This work was
9532 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9533 EGEE (Enabling Grids for E-science in Europe).
9537 * RC4 performance overhaul on modern architectures/implementations, such
9538 as Intel P4, IA-64 and AMD64.
9542 * New utility extract-section.pl. This can be used specify an alternative
9543 section number in a pod file instead of having to treat each file as
9544 a separate case in Makefile. This can be done by adding two lines to the
9547 =for comment openssl_section:XXX
9549 The blank line is mandatory.
9553 * New arguments -certform, -keyform and -pass for s_client and s_server
9554 to allow alternative format key and certificate files and passphrase
9559 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9560 update associated structures and add various utility functions.
9562 Add new policy related verify parameters, include policy checking in
9563 standard verify code. Enhance 'smime' application with extra parameters
9564 to support policy checking and print out.
9568 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9569 Nehemiah processors. These extensions support AES encryption in hardware
9570 as well as RNG (though RNG support is currently disabled).
9572 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9574 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9578 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9580 *Andy Polyakov and a number of other people*
9582 * Improved PowerPC platform support. Most notably BIGNUM assembler
9583 implementation contributed by IBM.
9585 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9587 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9588 exponent rather than 'unsigned long'. There is a corresponding change to
9589 the new 'rsa_keygen' element of the RSA_METHOD structure.
9591 *Jelte Jansen, Geoff Thorpe*
9593 * Functionality for creating the initial serial number file is now
9594 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9596 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9597 number file to 1, which is bound to cause problems. To avoid
9598 the problems while respecting compatibility between different 0.9.7
9599 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9600 CA.pl for serial number initialization. With the new release 0.9.8,
9601 we can fix the problem directly in the 'ca' utility.)
9605 * Reduced header interdependencies by declaring more opaque objects in
9606 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9607 give fewer recursive includes, which could break lazy source code - so
9608 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9609 developers should define this symbol when building and using openssl to
9610 ensure they track the recommended behaviour, interfaces, [etc], but
9611 backwards-compatible behaviour prevails when this isn't defined.
9615 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9619 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9620 This will generate a random key of the appropriate length based on the
9621 cipher context. The EVP_CIPHER can provide its own random key generation
9622 routine to support keys of a specific form. This is used in the des and
9623 3des routines to generate a key of the correct parity. Update S/MIME
9624 code to use new functions and hence generate correct parity DES keys.
9625 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9626 valid (weak or incorrect parity).
9630 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9631 as looking them up. This is useful when the verified structure may contain
9632 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9633 present unless the new PKCS7_NO_CRL flag is asserted.
9637 * Extend ASN1 oid configuration module. It now additionally accepts the
9640 shortName = some long name, 1.2.3.4
9644 * Reimplemented the BN_CTX implementation. There is now no more static
9645 limitation on the number of variables it can handle nor the depth of the
9646 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9647 information can now expand as required, and rather than having a single
9648 static array of bignums, BN_CTX now uses a linked-list of such arrays
9649 allowing it to expand on demand whilst maintaining the usefulness of
9650 BN_CTX's "bundling".
9654 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9655 to allow all RSA operations to function using a single BN_CTX.
9659 * Preliminary support for certificate policy evaluation and checking. This
9660 is initially intended to pass the tests outlined in "Conformance Testing
9661 of Relying Party Client Certificate Path Processing Logic" v1.07.
9665 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9666 remained unused and not that useful. A variety of other little bignum
9667 tweaks and fixes have also been made continuing on from the audit (see
9672 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9673 associated ASN1, EVP and SSL functions and old ASN1 macros.
9677 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9678 and this should never fail. So the return value from the use of
9679 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9680 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9684 * BN_CTX_get() should return zero-valued bignums, providing the same
9685 initialised value as BN_new().
9687 *Geoff Thorpe, suggested by Ulf Möller*
9689 * Support for inhibitAnyPolicy certificate extension.
9693 * An audit of the BIGNUM code is underway, for which debugging code is
9694 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9695 is considered valid when processing BIGNUMs, and causes execution to
9696 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9697 further steps are taken to deliberately pollute unused data in BIGNUM
9698 structures to try and expose faulty code further on. For now, openssl will
9699 (in its default mode of operation) continue to tolerate the inconsistent
9700 forms that it has tolerated in the past, but authors and packagers should
9701 consider trying openssl and their own applications when compiled with
9702 these debugging symbols defined. It will help highlight potential bugs in
9703 their own code, and will improve the test coverage for OpenSSL itself. At
9704 some point, these tighter rules will become openssl's default to improve
9705 maintainability, though the assert()s and other overheads will remain only
9706 in debugging configurations. See bn.h for more details.
9708 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9710 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9711 that can only be obtained through BN_CTX_new() (which implicitly
9712 initialises it). The presence of this function only made it possible
9713 to overwrite an existing structure (and cause memory leaks).
9717 * Because of the callback-based approach for implementing LHASH as a
9718 template type, lh_insert() adds opaque objects to hash-tables and
9719 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9720 to clean up those corresponding objects before destroying the hash table
9721 (and losing the object pointers). So some over-zealous constifications in
9722 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9723 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9724 prototyped to have "const" restrictions on the object pointers they are
9725 given (and so aren't required to cast them away any more).
9729 * The tmdiff.h API was so ugly and minimal that our own timing utility
9730 (speed) prefers to use its own implementation. The two implementations
9731 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9732 its object type properly exposed (MS_TM) instead of casting to/from
9733 `char *`. This may still change yet if someone realises MS_TM and
9735 aren't necessarily the greatest nomenclatures - but this is what was used
9736 internally to the implementation so I've used that for now.
9740 * Ensure that deprecated functions do not get compiled when
9741 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9742 the self-tests were still using deprecated key-generation functions so
9743 these have been updated also.
9747 * Reorganise PKCS#7 code to separate the digest location functionality
9748 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9749 New function PKCS7_set_digest() to set the digest type for PKCS#7
9750 digestedData type. Add additional code to correctly generate the
9751 digestedData type and add support for this type in PKCS7 initialization
9756 * New function PKCS7_set0_type_other() this initializes a PKCS7
9757 structure of type "other".
9761 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9762 sure the loop does correctly stop and breaking ("division by zero")
9763 modulus operations are not performed. The (pre-generated) prime
9764 table crypto/bn/bn_prime.h was already correct, but it could not be
9765 re-generated on some platforms because of the "division by zero"
9766 situation in the script.
9768 *Ralf S. Engelschall*
9770 * Update support for ECC-based TLS ciphersuites according to
9771 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9772 SHA-1 now is only used for "small" curves (where the
9773 representation of a field element takes up to 24 bytes); for
9774 larger curves, the field element resulting from ECDH is directly
9775 used as premaster secret.
9777 *Douglas Stebila (Sun Microsystems Laboratories)*
9779 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9780 curve secp160r1 to the tests.
9782 *Douglas Stebila (Sun Microsystems Laboratories)*
9784 * Add the possibility to load symbols globally with DSO.
9786 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9788 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9789 control of the error stack.
9793 * Add support for STORE in ENGINE.
9797 * Add the STORE type. The intention is to provide a common interface
9798 to certificate and key stores, be they simple file-based stores, or
9799 HSM-type store, or LDAP stores, or...
9800 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9804 * Add a generic structure called OPENSSL_ITEM. This can be used to
9805 pass a list of arguments to any function as well as provide a way
9806 for a function to pass data back to the caller.
9810 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9811 works like BUF_strdup() but can be used to duplicate a portion of
9812 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9817 * Add the function sk_find_ex() which works like sk_find(), but will
9818 return an index to an element even if an exact match couldn't be
9819 found. The index is guaranteed to point at the element where the
9820 searched-for key would be inserted to preserve sorting order.
9824 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9825 takes an extra flags argument for optional functionality. Currently,
9826 the following flags are defined:
9828 OBJ_BSEARCH_VALUE_ON_NOMATCH
9829 This one gets OBJ_bsearch_ex() to return a pointer to the first
9830 element where the comparing function returns a negative or zero
9833 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9834 This one gets OBJ_bsearch_ex() to return a pointer to the first
9835 element where the comparing function returns zero. This is useful
9836 if there are more than one element where the comparing function
9841 * Make it possible to create self-signed certificates with 'openssl ca'
9842 in such a way that the self-signed certificate becomes part of the
9843 CA database and uses the same mechanisms for serial number generation
9844 as all other certificate signing. The new flag '-selfsign' enables
9845 this functionality. Adapt CA.sh and CA.pl.in.
9849 * Add functionality to check the public key of a certificate request
9850 against a given private. This is useful to check that a certificate
9851 request can be signed by that key (self-signing).
9855 * Make it possible to have multiple active certificates with the same
9856 subject in the CA index file. This is done only if the keyword
9857 'unique_subject' is set to 'no' in the main CA section (default
9858 if 'CA_default') of the configuration file. The value is saved
9859 with the database itself in a separate index attribute file,
9860 named like the index file with '.attr' appended to the name.
9864 * Generate multi-valued AVAs using '+' notation in config files for
9869 * Support for nameConstraints certificate extension.
9873 * Support for policyConstraints certificate extension.
9877 * Support for policyMappings certificate extension.
9881 * Make sure the default DSA_METHOD implementation only uses its
9882 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9883 and change its own handlers to be NULL so as to remove unnecessary
9884 indirection. This lets alternative implementations fallback to the
9885 default implementation more easily.
9889 * Support for directoryName in GeneralName related extensions
9894 * Make it possible to link applications using Makefile.shared.
9895 Make that possible even when linking against static libraries!
9899 * Support for single pass processing for S/MIME signing. This now
9900 means that S/MIME signing can be done from a pipe, in addition
9901 cleartext signing (multipart/signed type) is effectively streaming
9902 and the signed data does not need to be all held in memory.
9904 This is done with a new flag PKCS7_STREAM. When this flag is set
9905 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9906 is done after the data is output (and digests calculated) in
9907 SMIME_write_PKCS7().
9911 * Add full support for -rpath/-R, both in shared libraries and
9912 applications, at least on the platforms where it's known how
9917 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9918 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9919 will now compute a table of multiples of the generator that
9920 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9921 faster (notably in the case of a single point multiplication,
9922 scalar * generator).
9924 *Nils Larsch, Bodo Moeller*
9926 * IPv6 support for certificate extensions. The various extensions
9927 which use the IP:a.b.c.d can now take IPv6 addresses using the
9928 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9933 * Added an ENGINE that implements RSA by performing private key
9934 exponentiations with the GMP library. The conversions to and from
9935 GMP's mpz_t format aren't optimised nor are any montgomery forms
9936 cached, and on x86 it appears OpenSSL's own performance has caught up.
9937 However there are likely to be other architectures where GMP could
9938 provide a boost. This ENGINE is not built in by default, but it can be
9939 specified at Configure time and should be accompanied by the necessary
9940 linker additions, eg;
9941 ./config -DOPENSSL_USE_GMP -lgmp
9945 * "openssl engine" will not display ENGINE/DSO load failure errors when
9946 testing availability of engines with "-t" - the old behaviour is
9947 produced by increasing the feature's verbosity with "-tt".
9951 * ECDSA routines: under certain error conditions uninitialized BN objects
9952 could be freed. Solution: make sure initialization is performed early
9953 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9958 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9959 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9960 software implementations. For DSA and DH, parameter generation can
9961 also be overridden by providing the appropriate method callbacks.
9965 * Change the "progress" mechanism used in key-generation and
9966 primality testing to functions that take a new BN_GENCB pointer in
9967 place of callback/argument pairs. The new API functions have `_ex`
9968 postfixes and the older functions are reimplemented as wrappers for
9969 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9970 declarations of the old functions to help (graceful) attempts to
9971 migrate to the new functions. Also, the new key-generation API
9972 functions operate on a caller-supplied key-structure and return
9973 success/failure rather than returning a key or NULL - this is to
9974 help make "keygen" another member function of RSA_METHOD etc.
9976 Example for using the new callback interface:
9978 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9982 BN_GENCB_set(&my_cb, my_callback, my_arg);
9984 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9985 /* For the meaning of a, b in calls to my_callback(), see the
9986 * documentation of the function that calls the callback.
9987 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9988 * my_callback should return 1 if it wants BN_is_prime_ex()
9989 * to continue, or 0 to stop.
9994 * Change the ZLIB compression method to be stateful, and make it
9995 available to TLS with the number defined in
9996 draft-ietf-tls-compression-04.txt.
10000 * Add the ASN.1 structures and functions for CertificatePair, which
10001 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10003 CertificatePair ::= SEQUENCE {
10004 forward [0] Certificate OPTIONAL,
10005 reverse [1] Certificate OPTIONAL,
10006 -- at least one of the pair shall be present -- }
10008 Also implement the PEM functions to read and write certificate
10009 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10011 This needed to be defined, mostly for the sake of the LDAP
10012 attribute crossCertificatePair, but may prove useful elsewhere as
10017 * Make it possible to inhibit symlinking of shared libraries in
10018 Makefile.shared, for Cygwin's sake.
10022 * Extend the BIGNUM API by creating a function
10023 void BN_set_negative(BIGNUM *a, int neg);
10024 and a macro that behave like
10025 int BN_is_negative(const BIGNUM *a);
10027 to avoid the need to access 'a->neg' directly in applications.
10031 * Implement fast modular reduction for pseudo-Mersenne primes
10032 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10033 EC_GROUP_new_curve_GFp() will now automatically use this
10036 *Nils Larsch <nla@trustcenter.de>*
10038 * Add new lock type (CRYPTO_LOCK_BN).
10042 * Change the ENGINE framework to automatically load engines
10043 dynamically from specific directories unless they could be
10044 found to already be built in or loaded. Move all the
10045 current engines except for the cryptodev one to a new
10046 directory engines/.
10047 The engines in engines/ are built as shared libraries if
10048 the "shared" options was given to ./Configure or ./config.
10049 Otherwise, they are inserted in libcrypto.a.
10050 /usr/local/ssl/engines is the default directory for dynamic
10051 engines, but that can be overridden at configure time through
10052 the usual use of --prefix and/or --openssldir, and at run
10053 time with the environment variable OPENSSL_ENGINES.
10055 *Geoff Thorpe and Richard Levitte*
10057 * Add Makefile.shared, a helper makefile to build shared
10058 libraries. Adapt Makefile.org.
10062 * Add version info to Win32 DLLs.
10064 *Peter 'Luna' Runestig" <peter@runestig.com>*
10066 * Add new 'medium level' PKCS#12 API. Certificates and keys
10067 can be added using this API to created arbitrary PKCS#12
10068 files while avoiding the low level API.
10070 New options to PKCS12_create(), key or cert can be NULL and
10071 will then be omitted from the output file. The encryption
10072 algorithm NIDs can be set to -1 for no encryption, the mac
10073 iteration count can be set to 0 to omit the mac.
10075 Enhance pkcs12 utility by making the -nokeys and -nocerts
10076 options work when creating a PKCS#12 file. New option -nomac
10077 to omit the mac, NONE can be set for an encryption algorithm.
10078 New code is modified to use the enhanced PKCS12_create()
10079 instead of the low level API.
10083 * Extend ASN1 encoder to support indefinite length constructed
10084 encoding. This can output sequences tags and octet strings in
10085 this form. Modify pk7_asn1.c to support indefinite length
10086 encoding. This is experimental and needs additional code to
10087 be useful, such as an ASN1 bio and some enhanced streaming
10090 Extend template encode functionality so that tagging is passed
10091 down to the template encoder.
10095 * Let 'openssl req' fail if an argument to '-newkey' is not
10096 recognized instead of using RSA as a default.
10100 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10101 As these are not official, they are not included in "ALL";
10102 the "ECCdraft" ciphersuite group alias can be used to select them.
10104 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10106 * Add ECDH engine support.
10108 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10110 * Add ECDH in new directory crypto/ecdh/.
10112 *Douglas Stebila (Sun Microsystems Laboratories)*
10114 * Let BN_rand_range() abort with an error after 100 iterations
10115 without success (which indicates a broken PRNG).
10119 * Change BN_mod_sqrt() so that it verifies that the input value
10120 is really the square of the return value. (Previously,
10121 BN_mod_sqrt would show GIGO behaviour.)
10125 * Add named elliptic curves over binary fields from X9.62, SECG,
10126 and WAP/WTLS; add OIDs that were still missing.
10128 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10130 * Extend the EC library for elliptic curves over binary fields
10131 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10134 EC_GF2m_simple_method
10138 EC_GROUP_new_curve_GF2m
10139 EC_GROUP_set_curve_GF2m
10140 EC_GROUP_get_curve_GF2m
10141 EC_POINT_set_affine_coordinates_GF2m
10142 EC_POINT_get_affine_coordinates_GF2m
10143 EC_POINT_set_compressed_coordinates_GF2m
10145 Point compression for binary fields is disabled by default for
10146 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10149 As binary polynomials are represented as BIGNUMs, various members
10150 of the EC_GROUP and EC_POINT data structures can be shared
10151 between the implementations for prime fields and binary fields;
10152 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10153 are essentially identical to their `..._GFp` counterparts.
10154 (For simplicity, the `..._GFp` prefix has been dropped from
10155 various internal method names.)
10157 An internal 'field_div' method (similar to 'field_mul' and
10158 'field_sqr') has been added; this is used only for binary fields.
10160 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10162 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10163 through methods ('mul', 'precompute_mult').
10165 The generic implementations (now internally called 'ec_wNAF_mul'
10166 and 'ec_wNAF_precomputed_mult') remain the default if these
10167 methods are undefined.
10169 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10171 * New function EC_GROUP_get_degree, which is defined through
10172 EC_METHOD. For curves over prime fields, this returns the bit
10173 length of the modulus.
10175 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10177 * New functions EC_GROUP_dup, EC_POINT_dup.
10178 (These simply call ..._new and ..._copy).
10180 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10182 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10183 Polynomials are represented as BIGNUMs (where the sign bit is not
10184 used) in the following functions [macros]:
10187 BN_GF2m_sub [= BN_GF2m_add]
10188 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10189 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10190 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10192 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10193 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10194 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10195 BN_GF2m_cmp [= BN_ucmp]
10197 (Note that only the 'mod' functions are actually for fields GF(2^m).
10198 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10200 For some functions, an the irreducible polynomial defining a
10201 field can be given as an 'unsigned int[]' with strictly
10202 decreasing elements giving the indices of those bits that are set;
10203 i.e., p[] represents the polynomial
10204 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10206 p[0] > p[1] > ... > p[k] = 0.
10207 This applies to the following functions:
10210 BN_GF2m_mod_mul_arr
10211 BN_GF2m_mod_sqr_arr
10212 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10213 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10214 BN_GF2m_mod_exp_arr
10215 BN_GF2m_mod_sqrt_arr
10216 BN_GF2m_mod_solve_quad_arr
10220 Conversion can be performed by the following functions:
10225 bntest.c has additional tests for binary polynomial arithmetic.
10227 Two implementations for BN_GF2m_mod_div() are available.
10228 The default algorithm simply uses BN_GF2m_mod_inv() and
10229 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10230 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10231 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10233 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10235 * Add new error code 'ERR_R_DISABLED' that can be used when some
10236 functionality is disabled at compile-time.
10238 *Douglas Stebila <douglas.stebila@sun.com>*
10240 * Change default behaviour of 'openssl asn1parse' so that more
10241 information is visible when viewing, e.g., a certificate:
10243 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10244 mode the content of non-printable OCTET STRINGs is output in a
10245 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10246 avoid the appearance of a printable string.
10248 *Nils Larsch <nla@trustcenter.de>*
10250 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10252 EC_GROUP_set_asn1_flag()
10253 EC_GROUP_get_asn1_flag()
10254 EC_GROUP_set_point_conversion_form()
10255 EC_GROUP_get_point_conversion_form()
10256 These control ASN1 encoding details:
10257 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10258 has been set to OPENSSL_EC_NAMED_CURVE.
10259 - Points are encoded in uncompressed form by default; options for
10260 asn1_for are as for point2oct, namely
10261 POINT_CONVERSION_COMPRESSED
10262 POINT_CONVERSION_UNCOMPRESSED
10263 POINT_CONVERSION_HYBRID
10265 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10267 EC_GROUP_set_seed()
10268 EC_GROUP_get0_seed()
10269 EC_GROUP_get_seed_len()
10270 This is used only for ASN1 purposes (so far).
10272 *Nils Larsch <nla@trustcenter.de>*
10274 * Add 'field_type' member to EC_METHOD, which holds the NID
10275 of the appropriate field type OID. The new function
10276 EC_METHOD_get_field_type() returns this value.
10278 *Nils Larsch <nla@trustcenter.de>*
10281 EC_POINT_point2bn()
10282 EC_POINT_bn2point()
10283 EC_POINT_point2hex()
10284 EC_POINT_hex2point()
10285 providing useful interfaces to EC_POINT_point2oct() and
10286 EC_POINT_oct2point().
10288 *Nils Larsch <nla@trustcenter.de>*
10290 * Change internals of the EC library so that the functions
10291 EC_GROUP_set_generator()
10292 EC_GROUP_get_generator()
10293 EC_GROUP_get_order()
10294 EC_GROUP_get_cofactor()
10295 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10296 to methods, which would lead to unnecessary code duplication when
10297 adding different types of curves.
10299 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10301 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10302 arithmetic, and such that modified wNAFs are generated
10303 (which avoid length expansion in many cases).
10307 * Add a function EC_GROUP_check_discriminant() (defined via
10308 EC_METHOD) that verifies that the curve discriminant is non-zero.
10310 Add a function EC_GROUP_check() that makes some sanity tests
10311 on a EC_GROUP, its generator and order. This includes
10312 EC_GROUP_check_discriminant().
10314 *Nils Larsch <nla@trustcenter.de>*
10316 * Add ECDSA in new directory crypto/ecdsa/.
10318 Add applications 'openssl ecparam' and 'openssl ecdsa'
10319 (these are based on 'openssl dsaparam' and 'openssl dsa').
10321 ECDSA support is also included in various other files across the
10322 library. Most notably,
10323 - 'openssl req' now has a '-newkey ecdsa:file' option;
10324 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10325 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10326 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10327 them suitable for ECDSA where domain parameters must be
10328 extracted before the specific public key;
10329 - ECDSA engine support has been added.
10331 *Nils Larsch <nla@trustcenter.de>*
10333 * Include some named elliptic curves, and add OIDs from X9.62,
10334 SECG, and WAP/WTLS. Each curve can be obtained from the new
10336 EC_GROUP_new_by_curve_name(),
10337 and the list of available named curves can be obtained with
10338 EC_get_builtin_curves().
10339 Also add a 'curve_name' member to EC_GROUP objects, which can be
10341 EC_GROUP_set_curve_name()
10342 EC_GROUP_get_curve_name()
10344 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10346 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10347 was actually never needed) and in BN_mul(). The removal in BN_mul()
10348 required a small change in bn_mul_part_recursive() and the addition
10349 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10350 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10351 bn_sub_words() and bn_add_words() except they take arrays with
10356 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10358 * Cleanse PEM buffers before freeing them since they may contain
10361 *Benjamin Bennett <ben@psc.edu>*
10363 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10364 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10365 authentication-only ciphersuites.
10369 * Since AES128 and AES256 share a single mask bit in the logic of
10370 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10371 kludge to work properly if AES128 is available and AES256 isn't.
10375 * Expand security boundary to match 1.1.1 module.
10379 * Remove redundant features: hash file source, editing of test vectors
10380 modify fipsld to use external fips_premain.c signature.
10384 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10385 run algorithm test programs.
10389 * Make algorithm test programs more tolerant of whitespace.
10393 * Have SSL/TLS server implementation tolerate "mismatched" record
10394 protocol version while receiving ClientHello even if the
10395 ClientHello is fragmented. (The server can't insist on the
10396 particular protocol version it has chosen before the ServerHello
10397 message has informed the client about his choice.)
10401 * Load error codes if they are not already present instead of using a
10402 static variable. This allows them to be cleanly unloaded and reloaded.
10406 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10408 * Introduce limits to prevent malicious keys being able to
10409 cause a denial of service. [CVE-2006-2940][]
10411 *Steve Henson, Bodo Moeller*
10413 * Fix ASN.1 parsing of certain invalid structures that can result
10414 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10416 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10417 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10419 * Fix SSL client code which could crash if connecting to a
10420 malicious SSLv2 server. [CVE-2006-4343][]
10422 *Tavis Ormandy and Will Drewry, Google Security Team*
10424 * Change ciphersuite string processing so that an explicit
10425 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10426 will no longer include "AES128-SHA"), and any other similar
10427 ciphersuite (same bitmap) from *other* protocol versions (so that
10428 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10429 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10430 changes from 0.9.8b and 0.9.8d.
10434 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10436 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10437 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10439 * Change the Unix randomness entropy gathering to use poll() when
10440 possible instead of select(), since the latter has some
10441 undesirable limitations.
10443 *Darryl Miles via Richard Levitte and Bodo Moeller*
10445 * Disable rogue ciphersuites:
10447 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10448 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10449 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10451 The latter two were purportedly from
10452 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10455 Also deactivate the remaining ciphersuites from
10456 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10457 unofficial, and the ID has long expired.
10461 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10462 dual-core machines) and other potential thread-safety issues.
10466 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10468 * Adapt fipsld and the build system to link against the validated FIPS
10469 module in FIPS mode.
10473 * Fixes for VC++ 2005 build under Windows.
10477 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10478 from a Windows bash shell such as MSYS. It is autodetected from the
10479 "config" script when run from a VC++ environment. Modify standard VC++
10480 build to use fipscanister.o from the GNU make build.
10484 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10486 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10487 The value now differs depending on if you build for FIPS or not.
10488 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10489 safely run with a non-FIPSed libcrypto, as it may crash because of
10490 the difference induced by this change.
10494 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10496 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10497 (part of SSL_OP_ALL). This option used to disable the
10498 countermeasure against man-in-the-middle protocol-version
10499 rollback in the SSL 2.0 server implementation, which is a bad
10500 idea. [CVE-2005-2969][]
10502 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10503 for Information Security, National Institute of Advanced Industrial
10504 Science and Technology [AIST, Japan)]*
10506 * Minimal support for X9.31 signatures and PSS padding modes. This is
10507 mainly for FIPS compliance and not fully integrated at this stage.
10511 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10512 the exponentiation using a fixed-length exponent. (Otherwise,
10513 the information leaked through timing could expose the secret key
10514 after many signatures; cf. Bleichenbacher's attack on DSA with
10519 * Make a new fixed-window mod_exp implementation the default for
10520 RSA, DSA, and DH private-key operations so that the sequence of
10521 squares and multiplies and the memory access pattern are
10522 independent of the particular secret key. This will mitigate
10523 cache-timing and potential related attacks.
10525 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10526 and this is automatically used by BN_mod_exp_mont() if the new flag
10527 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10528 will use this BN flag for private exponents unless the flag
10529 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10530 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10532 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10534 * Change the client implementation for SSLv23_method() and
10535 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10536 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10537 (Previously, the SSL 2.0 backwards compatible Client Hello
10538 message format would be used even with SSL_OP_NO_SSLv2.)
10542 * Add support for smime-type MIME parameter in S/MIME messages which some
10547 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10548 a threadsafe manner. Modify rsa code to use new function and add calls
10549 to dsa and dh code (which had race conditions before).
10553 * Include the fixed error library code in the C error file definitions
10554 instead of fixing them up at runtime. This keeps the error code
10555 structures constant.
10559 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10561 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10564 * Fixes for newer kerberos headers. NB: the casts are needed because
10565 the 'length' field is signed on one version and unsigned on another
10566 with no (?) obvious way to tell the difference, without these VC++
10567 complains. Also the "definition" of FAR (blank) is no longer included
10568 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10569 some needed definitions.
10573 * Undo Cygwin change.
10577 * Added support for proxy certificates according to RFC 3820.
10578 Because they may be a security thread to unaware applications,
10579 they must be explicitly allowed in run-time. See
10580 docs/HOWTO/proxy_certificates.txt for further information.
10584 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10586 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10587 server and client random values. Previously
10588 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10589 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10591 This change has negligible security impact because:
10593 1. Server and client random values still have 24 bytes of pseudo random
10596 2. Server and client random values are sent in the clear in the initial
10599 3. The master secret is derived using the premaster secret (48 bytes in
10600 size for static RSA ciphersuites) as well as client server and random
10603 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10606 *Stephen Henson, reported by UK NISCC*
10608 * Use Windows randomness collection on Cygwin.
10612 * Fix hang in EGD/PRNGD query when communication socket is closed
10613 prematurely by EGD/PRNGD.
10615 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10617 * Prompt for pass phrases when appropriate for PKCS12 input format.
10621 * Back-port of selected performance improvements from development
10622 branch, as well as improved support for PowerPC platforms.
10626 * Add lots of checks for memory allocation failure, error codes to indicate
10627 failure and freeing up memory if a failure occurs.
10629 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10631 * Add new -passin argument to dgst.
10635 * Perform some character comparisons of different types in X509_NAME_cmp:
10636 this is needed for some certificates that re-encode DNs into UTF8Strings
10637 (in violation of RFC3280) and can't or won't issue name rollover
10642 * Make an explicit check during certificate validation to see that
10643 the CA setting in each certificate on the chain is correct. As a
10644 side effect always do the following basic checks on extensions,
10645 not just when there's an associated purpose to the check:
10647 - if there is an unhandled critical extension (unless the user
10648 has chosen to ignore this fault)
10649 - if the path length has been exceeded (if one is set at all)
10650 - that certain extensions fit the associated purpose (if one has
10655 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10657 * Avoid a race condition when CRLs are checked in a multi threaded
10658 environment. This would happen due to the reordering of the revoked
10659 entries during signature checking and serial number lookup. Now the
10660 encoding is cached and the serial number sort performed under a lock.
10661 Add new STACK function sk_is_sorted().
10665 * Add Delta CRL to the extension code.
10669 * Various fixes to s3_pkt.c so alerts are sent properly.
10671 *David Holmes <d.holmes@f5.com>*
10673 * Reduce the chances of duplicate issuer name and serial numbers (in
10674 violation of RFC3280) using the OpenSSL certificate creation utilities.
10675 This is done by creating a random 64 bit value for the initial serial
10676 number when a serial number file is created or when a self signed
10677 certificate is created using 'openssl req -x509'. The initial serial
10678 number file is created using 'openssl x509 -next_serial' in CA.pl
10679 rather than being initialized to 1.
10683 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10685 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10686 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10688 *Joe Orton, Steve Henson*
10690 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10693 *Joe Orton, Steve Henson*
10695 * Make it possible to have multiple active certificates with the same
10696 subject in the CA index file. This is done only if the keyword
10697 'unique_subject' is set to 'no' in the main CA section (default
10698 if 'CA_default') of the configuration file. The value is saved
10699 with the database itself in a separate index attribute file,
10700 named like the index file with '.attr' appended to the name.
10704 * X509 verify fixes. Disable broken certificate workarounds when
10705 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10706 keyUsage extension present. Don't accept CRLs with unhandled critical
10707 extensions: since verify currently doesn't process CRL extensions this
10708 rejects a CRL with *any* critical extensions. Add new verify error codes
10713 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10714 A clarification of RFC2560 will require the use of OCTET STRINGs and
10715 some implementations cannot handle the current raw format. Since OpenSSL
10716 copies and compares OCSP nonces as opaque blobs without any attempt at
10717 parsing them this should not create any compatibility issues.
10721 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10722 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10723 this HMAC (and other) operations are several times slower than OpenSSL
10728 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10730 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10732 * Use the correct content when signing type "other".
10736 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10738 * Fix various bugs revealed by running the NISCC test suite:
10740 Stop out of bounds reads in the ASN1 code when presented with
10741 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10743 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10745 If verify callback ignores invalid public key errors don't try to check
10746 certificate signature with the NULL public key.
10750 * New -ignore_err option in ocsp application to stop the server
10751 exiting on the first error in a request.
10755 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10756 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10761 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10762 extra data after the compression methods not only for TLS 1.0
10763 but also for SSL 3.0 (as required by the specification).
10765 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10767 * Change X509_certificate_type() to mark the key as exported/exportable
10768 when it's 512 *bits* long, not 512 bytes.
10772 * Change AES_cbc_encrypt() so it outputs exact multiple of
10773 blocks during encryption.
10777 * Various fixes to base64 BIO and non blocking I/O. On write
10778 flushes were not handled properly if the BIO retried. On read
10779 data was not being buffered properly and had various logic bugs.
10780 This also affects blocking I/O when the data being decoded is a
10785 * Various S/MIME bugfixes and compatibility changes:
10786 output correct application/pkcs7 MIME type if
10787 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10788 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10789 of files as .eml work). Correctly handle very long lines in MIME
10794 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10796 * Countermeasure against the Klima-Pokorny-Rosa extension of
10797 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10798 a protocol version number mismatch like a decryption error
10799 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10803 * Turn on RSA blinding by default in the default implementation
10804 to avoid a timing attack. Applications that don't want it can call
10805 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10806 They would be ill-advised to do so in most cases.
10808 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10810 * Change RSA blinding code so that it works when the PRNG is not
10811 seeded (in this case, the secret RSA exponent is abused as
10812 an unpredictable seed -- if it is not unpredictable, there
10813 is no point in blinding anyway). Make RSA blinding thread-safe
10814 by remembering the creator's thread ID in rsa->blinding and
10815 having all other threads use local one-time blinding factors
10816 (this requires more computation than sharing rsa->blinding, but
10817 avoids excessive locking; and if an RSA object is not shared
10818 between threads, blinding will still be very fast).
10822 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10823 ENGINE as defaults for all supported algorithms irrespective of
10824 the 'flags' parameter. 'flags' is now honoured, so applications
10825 should make sure they are passing it correctly.
10829 * Target "mingw" now allows native Windows code to be generated in
10830 the Cygwin environment as well as with the MinGW compiler.
10834 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10836 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10837 via timing by performing a MAC computation even if incorrect
10838 block cipher padding has been found. This is a countermeasure
10839 against active attacks where the attacker has to distinguish
10840 between bad padding and a MAC verification error. [CVE-2003-0078][]
10842 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10843 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10844 Martin Vuagnoux (EPFL, Ilion)*
10846 * Make the no-err option work as intended. The intention with no-err
10847 is not to have the whole error stack handling routines removed from
10848 libcrypto, it's only intended to remove all the function name and
10849 reason texts, thereby removing some of the footprint that may not
10850 be interesting if those errors aren't displayed anyway.
10852 NOTE: it's still possible for any application or module to have its
10853 own set of error texts inserted. The routines are there, just not
10854 used by default when no-err is given.
10858 * Add support for FreeBSD on IA64.
10860 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10862 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10863 Kerberos function mit_des_cbc_cksum(). Before this change,
10864 the value returned by DES_cbc_cksum() was like the one from
10865 mit_des_cbc_cksum(), except the bytes were swapped.
10867 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10869 * Allow an application to disable the automatic SSL chain building.
10870 Before this a rather primitive chain build was always performed in
10871 ssl3_output_cert_chain(): an application had no way to send the
10872 correct chain if the automatic operation produced an incorrect result.
10874 Now the chain builder is disabled if either:
10876 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10878 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10880 The reasoning behind this is that an application would not want the
10881 auto chain building to take place if extra chain certificates are
10882 present and it might also want a means of sending no additional
10883 certificates (for example the chain has two certificates and the
10888 * Add the possibility to build without the ENGINE framework.
10890 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10892 * Under Win32 gmtime() can return NULL: check return value in
10893 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10897 * DSA routines: under certain error conditions uninitialized BN objects
10898 could be freed. Solution: make sure initialization is performed early
10899 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10900 Nils Larsch <nla@trustcenter.de> via PR#459)
10904 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10905 checked on reconnect on the client side, therefore session resumption
10906 could still fail with a "ssl session id is different" error. This
10907 behaviour is masked when SSL_OP_ALL is used due to
10908 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10909 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10910 followup to PR #377.
10914 * IA-32 assembler support enhancements: unified ELF targets, support
10915 for SCO/Caldera platforms, fix for Cygwin shared build.
10919 * Add support for FreeBSD on sparc64. As a consequence, support for
10920 FreeBSD on non-x86 processors is separate from x86 processors on
10921 the config script, much like the NetBSD support.
10923 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10925 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
10927 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10930 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10931 code (06) was taken as the first octet of the session ID and the last
10932 octet was ignored consequently. As a result SSLv2 client side session
10933 caching could not have worked due to the session ID mismatch between
10935 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10940 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10941 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10946 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10947 seems that in spite of existing for more than a year, many application
10948 author have done nothing to provide the necessary callbacks, which
10949 means that this particular engine will not work properly anywhere.
10950 This is a very unfortunate situation which forces us, in the name
10951 of usability, to give the hw_ncipher.c a static lock, which is part
10953 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10954 appear in 0.9.8 or later. We EXPECT application authors to have
10955 dealt properly with this when 0.9.8 is released (unless we actually
10956 make such changes in the libcrypto locking code that changes will
10957 have to be made anyway).
10961 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10962 octets have been read, EOF or an error occurs. Without this change
10963 some truncated ASN1 structures will not produce an error.
10967 * Disable Heimdal support, since it hasn't been fully implemented.
10968 Still give the possibility to force the use of Heimdal, but with
10969 warnings and a request that patches get sent to openssl-dev.
10973 * Add the VC-CE target, introduce the WINCE sysname, and add
10974 INSTALL.WCE and appropriate conditionals to make it build.
10976 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10978 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10979 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10980 edit numbers of the version.
10982 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10984 * Introduce safe string copy and catenation functions
10985 (BUF_strlcpy() and BUF_strlcat()).
10987 *Ben Laurie (CHATS) and Richard Levitte*
10989 * Avoid using fixed-size buffers for one-line DNs.
10991 *Ben Laurie (CHATS)*
10993 * Add BUF_MEM_grow_clean() to avoid information leakage when
10994 resizing buffers containing secrets, and use where appropriate.
10996 *Ben Laurie (CHATS)*
10998 * Avoid using fixed size buffers for configuration file location.
11000 *Ben Laurie (CHATS)*
11002 * Avoid filename truncation for various CA files.
11004 *Ben Laurie (CHATS)*
11006 * Use sizeof in preference to magic numbers.
11008 *Ben Laurie (CHATS)*
11010 * Avoid filename truncation in cert requests.
11012 *Ben Laurie (CHATS)*
11014 * Add assertions to check for (supposedly impossible) buffer
11017 *Ben Laurie (CHATS)*
11019 * Don't cache truncated DNS entries in the local cache (this could
11020 potentially lead to a spoofing attack).
11022 *Ben Laurie (CHATS)*
11024 * Fix various buffers to be large enough for hex/decimal
11025 representations in a platform independent manner.
11027 *Ben Laurie (CHATS)*
11029 * Add CRYPTO_realloc_clean() to avoid information leakage when
11030 resizing buffers containing secrets, and use where appropriate.
11032 *Ben Laurie (CHATS)*
11034 * Add BIO_indent() to avoid much slightly worrying code to do
11037 *Ben Laurie (CHATS)*
11039 * Convert sprintf()/BIO_puts() to BIO_printf().
11041 *Ben Laurie (CHATS)*
11043 * buffer_gets() could terminate with the buffer only half
11046 *Ben Laurie (CHATS)*
11048 * Add assertions to prevent user-supplied crypto functions from
11049 overflowing internal buffers by having large block sizes, etc.
11051 *Ben Laurie (CHATS)*
11053 * New OPENSSL_assert() macro (similar to assert(), but enabled
11056 *Ben Laurie (CHATS)*
11058 * Eliminate unused copy of key in RC4.
11060 *Ben Laurie (CHATS)*
11062 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11064 *Ben Laurie (CHATS)*
11066 * Fix off-by-one error in EGD path.
11068 *Ben Laurie (CHATS)*
11070 * If RANDFILE path is too long, ignore instead of truncating.
11072 *Ben Laurie (CHATS)*
11074 * Eliminate unused and incorrectly sized X.509 structure
11077 *Ben Laurie (CHATS)*
11079 * Eliminate unused and dangerous function knumber().
11081 *Ben Laurie (CHATS)*
11083 * Eliminate unused and dangerous structure, KSSL_ERR.
11085 *Ben Laurie (CHATS)*
11087 * Protect against overlong session ID context length in an encoded
11088 session object. Since these are local, this does not appear to be
11091 *Ben Laurie (CHATS)*
11093 * Change from security patch (see 0.9.6e below) that did not affect
11094 the 0.9.6 release series:
11096 Remote buffer overflow in SSL3 protocol - an attacker could
11097 supply an oversized master key in Kerberos-enabled versions.
11100 *Ben Laurie (CHATS)*
11102 * Change the SSL kerb5 codes to match RFC 2712.
11106 * Make -nameopt work fully for req and add -reqopt switch.
11108 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11110 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11112 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11114 * Make sure tests can be performed even if the corresponding algorithms
11115 have been removed entirely. This was also the last step to make
11116 OpenSSL compilable with DJGPP under all reasonable conditions.
11118 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11120 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11121 to allow version independent disabling of normally unselected ciphers,
11122 which may be activated as a side-effect of selecting a single cipher.
11124 (E.g., cipher list string "RSA" enables ciphersuites that are left
11125 out of "ALL" because they do not provide symmetric encryption.
11126 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11128 *Lutz Jaenicke, Bodo Moeller*
11130 * Add appropriate support for separate platform-dependent build
11131 directories. The recommended way to make a platform-dependent
11132 build directory is the following (tested on Linux), maybe with
11135 # Place yourself outside of the OpenSSL source tree. In
11136 # this example, the environment variable OPENSSL_SOURCE
11137 # is assumed to contain the absolute OpenSSL source directory.
11138 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11139 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11140 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11141 mkdir -p `dirname $F`
11142 ln -s $OPENSSL_SOURCE/$F $F
11145 To be absolutely sure not to disturb the source tree, a "make clean"
11146 is a good thing. If it isn't successful, don't worry about it,
11147 it probably means the source directory is very clean.
11151 * Make sure any ENGINE control commands make local copies of string
11152 pointers passed to them whenever necessary. Otherwise it is possible
11153 the caller may have overwritten (or deallocated) the original string
11154 data when a later ENGINE operation tries to use the stored values.
11156 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11158 * Improve diagnostics in file reading and command-line digests.
11160 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11162 * Add AES modes CFB and OFB to the object database. Correct an
11163 error in AES-CFB decryption.
11167 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11168 allows existing EVP_CIPHER_CTX structures to be reused after
11169 calling `EVP_*Final()`. This behaviour is used by encryption
11170 BIOs and some applications. This has the side effect that
11171 applications must explicitly clean up cipher contexts with
11172 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11176 * Check the values of dna and dnb in bn_mul_recursive before calling
11177 bn_mul_comba (a non zero value means the a or b arrays do not contain
11178 n2 elements) and fallback to bn_mul_normal if either is not zero.
11182 * Fix escaping of non-ASCII characters when using the -subj option
11183 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11187 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11188 form for "surname", serialNumber has no short form.
11189 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11190 therefore remove "mail" short name for "internet 7".
11191 The OID for unique identifiers in X509 certificates is
11192 x500UniqueIdentifier, not uniqueIdentifier.
11193 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11197 * Add an "init" command to the ENGINE config module and auto initialize
11198 ENGINEs. Without any "init" command the ENGINE will be initialized
11199 after all ctrl commands have been executed on it. If init=1 the
11200 ENGINE is initialized at that point (ctrls before that point are run
11201 on the uninitialized ENGINE and after on the initialized one). If
11202 init=0 then the ENGINE will not be initialized at all.
11206 * Fix the 'app_verify_callback' interface so that the user-defined
11207 argument is actually passed to the callback: In the
11208 SSL_CTX_set_cert_verify_callback() prototype, the callback
11209 declaration has been changed from
11212 int (*cb)(X509_STORE_CTX *,void *);
11213 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11214 i=s->ctx->app_verify_callback(&ctx)
11215 has been changed into
11216 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11218 To update applications using SSL_CTX_set_cert_verify_callback(),
11219 a dummy argument can be added to their callback functions.
11221 *D. K. Smetters <smetters@parc.xerox.com>*
11223 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11225 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11227 * Add and OPENSSL_LOAD_CONF define which will cause
11228 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11229 This allows older applications to transparently support certain
11230 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11231 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11232 load the config file and OPENSSL_add_all_algorithms_conf() which will
11233 always load it have also been added.
11237 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11238 Adjust NIDs and EVP layer.
11240 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11242 * Config modules support in openssl utility.
11244 Most commands now load modules from the config file,
11245 though in a few (such as version) this isn't done
11246 because it couldn't be used for anything.
11248 In the case of ca and req the config file used is
11249 the same as the utility itself: that is the -config
11250 command line option can be used to specify an
11255 * Move default behaviour from OPENSSL_config(). If appname is NULL
11256 use "openssl_conf" if filename is NULL use default openssl config file.
11260 * Add an argument to OPENSSL_config() to allow the use of an alternative
11261 config section name. Add a new flag to tolerate a missing config file
11262 and move code to CONF_modules_load_file().
11266 * Support for crypto accelerator cards from Accelerated Encryption
11267 Processing, www.aep.ie. (Use engine 'aep')
11268 The support was copied from 0.9.6c [engine] and adapted/corrected
11269 to work with the new engine framework.
11271 *AEP Inc. and Richard Levitte*
11273 * Support for SureWare crypto accelerator cards from Baltimore
11274 Technologies. (Use engine 'sureware')
11275 The support was copied from 0.9.6c [engine] and adapted
11276 to work with the new engine framework.
11280 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11281 make the newer ENGINE framework commands for the CHIL engine work.
11283 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11285 * Make it possible to produce shared libraries on ReliantUNIX.
11287 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11289 * Add the configuration target debug-linux-ppro.
11290 Make 'openssl rsa' use the general key loading routines
11291 implemented in apps.c, and make those routines able to
11292 handle the key format FORMAT_NETSCAPE and the variant
11295 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11297 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11299 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11301 * Add -keyform to rsautl, and document -engine.
11303 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11305 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11306 BIO_R_NO_SUCH_FILE error code rather than the generic
11307 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11311 * Add new functions
11312 ERR_peek_last_error
11313 ERR_peek_last_error_line
11314 ERR_peek_last_error_line_data.
11315 These are similar to
11317 ERR_peek_error_line
11318 ERR_peek_error_line_data,
11319 but report on the latest error recorded rather than the first one
11320 still in the error queue.
11322 *Ben Laurie, Bodo Moeller*
11324 * default_algorithms option in ENGINE config module. This allows things
11326 default_algorithms = ALL
11327 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11331 * Preliminary ENGINE config module.
11335 * New experimental application configuration code.
11339 * Change the AES code to follow the same name structure as all other
11340 symmetric ciphers, and behave the same way. Move everything to
11341 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11343 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11345 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11347 *Ben Laurie and Theo de Raadt*
11349 * Add option to output public keys in req command.
11351 *Massimiliano Pala madwolf@openca.org*
11353 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11354 (up to about 10% better than before for P-192 and P-224).
11358 * New functions/macros
11360 SSL_CTX_set_msg_callback(ctx, cb)
11361 SSL_CTX_set_msg_callback_arg(ctx, arg)
11362 SSL_set_msg_callback(ssl, cb)
11363 SSL_set_msg_callback_arg(ssl, arg)
11365 to request calling a callback function
11367 void cb(int write_p, int version, int content_type,
11368 const void *buf, size_t len, SSL *ssl, void *arg)
11370 whenever a protocol message has been completely received
11371 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11372 protocol version according to which the SSL library interprets
11373 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11374 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11375 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11376 specification (change_cipher_spec(20), alert(21), handshake(22)).
11377 'buf' and 'len' point to the actual message, 'ssl' to the
11378 SSL object, and 'arg' is the application-defined value set by
11379 SSL[_CTX]_set_msg_callback_arg().
11381 'openssl s_client' and 'openssl s_server' have new '-msg' options
11382 to enable a callback that displays all protocol messages.
11386 * Change the shared library support so shared libraries are built as
11387 soon as the corresponding static library is finished, and thereby get
11388 openssl and the test programs linked against the shared library.
11389 This still only happens when the keyword "shard" has been given to
11390 the configuration scripts.
11392 NOTE: shared library support is still an experimental thing, and
11393 backward binary compatibility is still not guaranteed.
11395 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11397 * Add support for Subject Information Access extension.
11399 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11401 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11402 additional bytes when new memory had to be allocated, not just
11403 when reusing an existing buffer.
11407 * New command line and configuration option 'utf8' for the req command.
11408 This allows field values to be specified as UTF8 strings.
11412 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11413 runs for the former and machine-readable output for the latter.
11417 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11418 of the e-mail address in the DN (i.e., it will go into a certificate
11419 extension only). The new configuration file option 'email_in_dn = no'
11420 has the same effect.
11422 *Massimiliano Pala madwolf@openca.org*
11424 * Change all functions with names starting with `des_` to be starting
11425 with `DES_` instead. Add wrappers that are compatible with libdes,
11426 but are named `_ossl_old_des_*`. Finally, add macros that map the
11427 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11428 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11429 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11432 Since we provide two compatibility mappings, the user needs to
11433 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11434 compatibility is desired. The default (i.e., when that macro
11435 isn't defined) is OpenSSL 0.9.6c compatibility.
11437 There are also macros that enable and disable the support of old
11438 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11439 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11440 are defined, the default will apply: to support the old des routines.
11442 In either case, one must include openssl/des.h to get the correct
11443 definitions. Do not try to just include openssl/des_old.h, that
11446 NOTE: This is a major break of an old API into a new one. Software
11447 authors are encouraged to switch to the `DES_` style functions. Some
11448 time in the future, des_old.h and the libdes compatibility functions
11449 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11450 default), and then completely removed.
11454 * Test for certificates which contain unsupported critical extensions.
11455 If such a certificate is found during a verify operation it is
11456 rejected by default: this behaviour can be overridden by either
11457 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11458 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11459 X509_supported_extension() has also been added which returns 1 if a
11460 particular extension is supported.
11464 * Modify the behaviour of EVP cipher functions in similar way to digests
11465 to retain compatibility with existing code.
11469 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11470 compatibility with existing code. In particular the 'ctx' parameter does
11471 not have to be to be initialized before the call to EVP_DigestInit() and
11472 it is tidied up after a call to EVP_DigestFinal(). New function
11473 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11474 EVP_MD_CTX_copy() changed to not require the destination to be
11475 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11476 requires the destination to be valid.
11478 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11479 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11483 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11484 so that complete 'Handshake' protocol structures are kept in memory
11485 instead of overwriting 'msg_type' and 'length' with 'body' data.
11489 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11491 *Massimo Santin via Richard Levitte*
11493 * Major restructuring to the underlying ENGINE code. This includes
11494 reduction of linker bloat, separation of pure "ENGINE" manipulation
11495 (initialisation, etc) from functionality dealing with implementations
11496 of specific crypto interfaces. This change also introduces integrated
11497 support for symmetric ciphers and digest implementations - so ENGINEs
11498 can now accelerate these by providing EVP_CIPHER and EVP_MD
11499 implementations of their own. This is detailed in crypto/engine/README
11500 as it couldn't be adequately described here. However, there are a few
11501 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11502 were changed in the original introduction of ENGINE code have now
11503 reverted back - the hooking from this code to ENGINE is now a good
11504 deal more passive and at run-time, operations deal directly with
11505 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11506 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11507 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11508 they were not being used by the framework as there is no concept of a
11509 BIGNUM_METHOD and they could not be generalised to the new
11510 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11511 ENGINE_cpy() has been removed as it cannot be consistently defined in
11516 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11520 * Change mkdef.pl to sort symbols that get the same entry number,
11521 and make sure the automatically generated functions `ERR_load_*`
11522 become part of libeay.num as well.
11526 * New function SSL_renegotiate_pending(). This returns true once
11527 renegotiation has been requested (either SSL_renegotiate() call
11528 or HelloRequest/ClientHello received from the peer) and becomes
11529 false once a handshake has been completed.
11530 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11531 sends a HelloRequest, but does not ensure that a handshake takes
11532 place. SSL_renegotiate_pending() is useful for checking if the
11533 client has followed the request.)
11537 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11538 By default, clients may request session resumption even during
11539 renegotiation (if session ID contexts permit); with this option,
11540 session resumption is possible only in the first handshake.
11542 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11543 more bits available for options that should not be part of
11544 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11548 * Add some demos for certificate and certificate request creation.
11552 * Make maximum certificate chain size accepted from the peer application
11553 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11554 "Douglas E. Engert" <deengert@anl.gov>.
11558 * Add support for shared libraries for Unixware-7
11559 (Boyd Lynn Gerber <gerberb@zenez.com>).
11563 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11564 be done prior to destruction. Use this to unload error strings from
11565 ENGINEs that load their own error strings. NB: This adds two new API
11566 functions to "get" and "set" this destroy handler in an ENGINE.
11570 * Alter all existing ENGINE implementations (except "openssl" and
11571 "openbsd") to dynamically instantiate their own error strings. This
11572 makes them more flexible to be built both as statically-linked ENGINEs
11573 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11574 Also, add stub code to each that makes building them as self-contained
11575 shared-libraries easier (see README.ENGINE).
11579 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11580 implementations into applications that are completely implemented in
11581 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11582 commands that can be used to configure what shared-library to load and
11583 to control aspects of the way it is handled. Also, made an update to
11584 the README.ENGINE file that brings its information up-to-date and
11585 provides some information and instructions on the "dynamic" ENGINE
11586 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11590 * Make it possible to unload ranges of ERR strings with a new
11591 "ERR_unload_strings" function.
11595 * Add a copy() function to EVP_MD.
11599 * Make EVP_MD routines take a context pointer instead of just the
11600 md_data void pointer.
11604 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11605 that the digest can only process a single chunk of data
11606 (typically because it is provided by a piece of
11607 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11608 is only going to provide a single chunk of data, and hence the
11609 framework needn't accumulate the data for oneshot drivers.
11613 * As with "ERR", make it possible to replace the underlying "ex_data"
11614 functions. This change also alters the storage and management of global
11615 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11616 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11617 index counters. The API functions that use this state have been changed
11618 to take a "class_index" rather than pointers to the class's local STACK
11619 and counter, and there is now an API function to dynamically create new
11620 classes. This centralisation allows us to (a) plug a lot of the
11621 thread-safety problems that existed, and (b) makes it possible to clean
11622 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11623 such data would previously have always leaked in application code and
11624 workarounds were in place to make the memory debugging turn a blind eye
11625 to it. Application code that doesn't use this new function will still
11626 leak as before, but their memory debugging output will announce it now
11627 rather than letting it slide.
11629 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11630 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11631 has a return value to indicate success or failure.
11635 * Make it possible to replace the underlying "ERR" functions such that the
11636 global state (2 LHASH tables and 2 locks) is only used by the "default"
11637 implementation. This change also adds two functions to "get" and "set"
11638 the implementation prior to it being automatically set the first time
11639 any other ERR function takes place. Ie. an application can call "get",
11640 pass the return value to a module it has just loaded, and that module
11641 can call its own "set" function using that value. This means the
11642 module's "ERR" operations will use (and modify) the error state in the
11643 application and not in its own statically linked copy of OpenSSL code.
11647 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11648 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11649 the operation, and provides a more encapsulated way for external code
11650 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11651 to use these functions rather than manually incrementing the counts.
11653 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11657 * Add EVP test program.
11661 * Add symmetric cipher support to ENGINE. Expect the API to change!
11665 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11666 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11667 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11668 These allow a CRL to be built without having to access X509_CRL fields
11669 directly. Modify 'ca' application to use new functions.
11673 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11674 bug workarounds. Rollback attack detection is a security feature.
11675 The problem will only arise on OpenSSL servers when TLSv1 is not
11676 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11677 Software authors not wanting to support TLSv1 will have special reasons
11678 for their choice and can explicitly enable this option.
11680 *Bodo Moeller, Lutz Jaenicke*
11682 * Rationalise EVP so it can be extended: don't include a union of
11683 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11684 (similar to those existing for EVP_CIPHER_CTX).
11689 EVP_MD_CTX_init(&md); /* new function call */
11690 EVP_DigestInit(&md, EVP_sha1());
11691 EVP_DigestUpdate(&md, in, len);
11692 EVP_DigestFinal(&md, out, NULL);
11693 EVP_MD_CTX_cleanup(&md); /* new function call */
11697 * Make DES key schedule conform to the usual scheme, as well as
11698 correcting its structure. This means that calls to DES functions
11699 now have to pass a pointer to a des_key_schedule instead of a
11700 plain des_key_schedule (which was actually always a pointer
11703 des_key_schedule ks;
11705 des_set_key_checked(..., &ks);
11706 des_ncbc_encrypt(..., &ks, ...);
11708 (Note that a later change renames 'des_...' into 'DES_...'.)
11712 * Initial reduction of linker bloat: the use of some functions, such as
11713 PEM causes large amounts of unused functions to be linked in due to
11714 poor organisation. For example pem_all.c contains every PEM function
11715 which has a knock on effect of linking in large amounts of (unused)
11716 ASN1 code. Grouping together similar functions and splitting unrelated
11717 functions prevents this.
11721 * Cleanup of EVP macros.
11725 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11726 correct `_ecb suffix`.
11730 * Add initial OCSP responder support to ocsp application. The
11731 revocation information is handled using the text based index
11732 use by the ca application. The responder can either handle
11733 requests generated internally, supplied in files (for example
11734 via a CGI script) or using an internal minimal server.
11738 * Add configuration choices to get zlib compression for TLS.
11742 * Changes to Kerberos SSL for RFC 2712 compliance:
11743 1. Implemented real KerberosWrapper, instead of just using
11744 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11745 2. Implemented optional authenticator field of KerberosWrapper.
11747 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11748 and authenticator structs; see crypto/krb5/.
11750 Generalized Kerberos calls to support multiple Kerberos libraries.
11751 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11752 via Richard Levitte*
11754 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11755 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11756 values for each of the key sizes rather than having just
11757 parameters (and 'speed' generating keys each time).
11761 * Speed up EVP routines.
11764 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11765 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11766 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11767 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11769 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11770 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11771 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11774 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11776 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11780 * Added the OS2-EMX target.
11782 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11784 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11785 to support NCONF routines in extension code. New function CONF_set_nconf()
11786 to allow functions which take an NCONF to also handle the old LHASH
11787 structure: this means that the old CONF compatible routines can be
11788 retained (in particular wrt extensions) without having to duplicate the
11789 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11793 * Enhance the general user interface with mechanisms for inner control
11794 and with possibilities to have yes/no kind of prompts.
11798 * Change all calls to low level digest routines in the library and
11799 applications to use EVP. Add missing calls to HMAC_cleanup() and
11800 don't assume HMAC_CTX can be copied using memcpy().
11802 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11804 * Add the possibility to control engines through control names but with
11805 arbitrary arguments instead of just a string.
11806 Change the key loaders to take a UI_METHOD instead of a callback
11807 function pointer. NOTE: this breaks binary compatibility with earlier
11808 versions of OpenSSL [engine].
11809 Adapt the nCipher code for these new conditions and add a card insertion
11814 * Enhance the general user interface with mechanisms to better support
11815 dialog box interfaces, application-defined prompts, the possibility
11816 to use defaults (for example default passwords from somewhere else)
11817 and interrupts/cancellations.
11821 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11822 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11826 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11827 tidy up some unnecessarily weird code in 'sk_new()').
11829 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11831 * Change the key loading routines for ENGINEs to use the same kind
11832 callback (pem_password_cb) as all other routines that need this
11837 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11838 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11839 than this minimum value is recommended.
11843 * New random seeder for OpenVMS, using the system process statistics
11844 that are easily reachable.
11848 * Windows apparently can't transparently handle global
11849 variables defined in DLLs. Initialisations such as:
11851 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11853 won't compile. This is used by the any applications that need to
11854 declare their own ASN1 modules. This was fixed by adding the option
11855 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11856 needed for static libraries under Win32.
11860 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11861 setting of purpose and trust fields. New X509_STORE trust and
11862 purpose functions and tidy up setting in other SSL functions.
11866 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11867 structure. These are inherited by X509_STORE_CTX when it is
11868 initialised. This allows various defaults to be set in the
11869 X509_STORE structure (such as flags for CRL checking and custom
11870 purpose or trust settings) for functions which only use X509_STORE_CTX
11871 internally such as S/MIME.
11873 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11874 trust settings if they are not set in X509_STORE. This allows X509_STORE
11875 purposes and trust (in S/MIME for example) to override any set by default.
11877 Add command line options for CRL checking to smime, s_client and s_server
11882 * Initial CRL based revocation checking. If the CRL checking flag(s)
11883 are set then the CRL is looked up in the X509_STORE structure and
11884 its validity and signature checked, then if the certificate is found
11885 in the CRL the verify fails with a revoked error.
11887 Various new CRL related callbacks added to X509_STORE_CTX structure.
11889 Command line options added to 'verify' application to support this.
11891 This needs some additional work, such as being able to handle multiple
11892 CRLs with different times, extension based lookup (rather than just
11893 by subject name) and ultimately more complete V2 CRL extension
11898 * Add a general user interface API (crypto/ui/). This is designed
11899 to replace things like des_read_password and friends (backward
11900 compatibility functions using this new API are provided).
11901 The purpose is to remove prompting functions from the DES code
11902 section as well as provide for prompting through dialog boxes in
11903 a window system and the like.
11907 * Add "ex_data" support to ENGINE so implementations can add state at a
11908 per-structure level rather than having to store it globally.
11912 * Make it possible for ENGINE structures to be copied when retrieved by
11913 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11914 This causes the "original" ENGINE structure to act like a template,
11915 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11916 operational state can be localised to each ENGINE structure, despite the
11917 fact they all share the same "methods". New ENGINE structures returned in
11918 this case have no functional references and the return value is the single
11919 structural reference. This matches the single structural reference returned
11920 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11925 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11926 needs to match any other type at all we need to manually clear the
11931 * Changes to the "openssl engine" utility to include;
11932 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11933 about an ENGINE's available control commands.
11934 - executing control commands from command line arguments using the
11935 '-pre' and '-post' switches. '-post' is only used if '-t' is
11936 specified and the ENGINE is successfully initialised. The syntax for
11937 the individual commands are colon-separated, for example;
11938 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11942 * New dynamic control command support for ENGINEs. ENGINEs can now
11943 declare their own commands (numbers), names (strings), descriptions,
11944 and input types for run-time discovery by calling applications. A
11945 subset of these commands are implicitly classed as "executable"
11946 depending on their input type, and only these can be invoked through
11947 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11948 can be based on user input, config files, etc). The distinction is
11949 that "executable" commands cannot return anything other than a boolean
11950 result and can only support numeric or string input, whereas some
11951 discoverable commands may only be for direct use through
11952 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11953 pointers, or other custom uses. The "executable" commands are to
11954 support parameterisations of ENGINE behaviour that can be
11955 unambiguously defined by ENGINEs and used consistently across any
11956 OpenSSL-based application. Commands have been added to all the
11957 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11958 control over shared-library paths without source code alterations.
11962 * Changed all ENGINE implementations to dynamically allocate their
11963 ENGINEs rather than declaring them statically. Apart from this being
11964 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11965 this also allows the implementations to compile without using the
11966 internal engine_int.h header.
11970 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11971 'const' value. Any code that should be able to modify a RAND_METHOD
11972 should already have non-const pointers to it (ie. they should only
11973 modify their own ones).
11977 * Made a variety of little tweaks to the ENGINE code.
11978 - "atalla" and "ubsec" string definitions were moved from header files
11979 to C code. "nuron" string definitions were placed in variables
11980 rather than hard-coded - allowing parameterisation of these values
11981 later on via ctrl() commands.
11982 - Removed unused "#if 0"'d code.
11983 - Fixed engine list iteration code so it uses ENGINE_free() to release
11984 structural references.
11985 - Constified the RAND_METHOD element of ENGINE structures.
11986 - Constified various get/set functions as appropriate and added
11987 missing functions (including a catch-all ENGINE_cpy that duplicates
11988 all ENGINE values onto a new ENGINE except reference counts/state).
11989 - Removed NULL parameter checks in get/set functions. Setting a method
11990 or function to NULL is a way of cancelling out a previously set
11991 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11992 and doesn't justify the extra error symbols and code.
11993 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
11994 flags from engine_int.h to engine.h.
11995 - Changed prototypes for ENGINE handler functions (init(), finish(),
11996 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12000 * Implement binary inversion algorithm for BN_mod_inverse in addition
12001 to the algorithm using long division. The binary algorithm can be
12002 used only if the modulus is odd. On 32-bit systems, it is faster
12003 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12004 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12005 up to 450 bits. In 64-bit environments, the binary algorithm
12006 appears to be advantageous for much longer moduli; here we use it
12007 for moduli up to 2048 bits.
12011 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12012 could not support the combine flag in choice fields.
12016 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12017 extensions from a certificate request to the certificate.
12021 * Allow multiple 'certopt' and 'nameopt' options to be separated
12022 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12023 file: this allows the display of the certificate about to be
12024 signed to be customised, to allow certain fields to be included
12025 or excluded and extension details. The old system didn't display
12026 multicharacter strings properly, omitted fields not in the policy
12027 and couldn't display additional details such as extensions.
12031 * Function EC_POINTs_mul for multiple scalar multiplication
12032 of an arbitrary number of elliptic curve points
12033 \sum scalars[i]*points[i],
12034 optionally including the generator defined for the EC_GROUP:
12035 scalar*generator + \sum scalars[i]*points[i].
12037 EC_POINT_mul is a simple wrapper function for the typical case
12038 that the point list has just one item (besides the optional
12043 * First EC_METHODs for curves over GF(p):
12045 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12046 operations and provides various method functions that can also
12047 operate with faster implementations of modular arithmetic.
12049 EC_GFp_mont_method() reuses most functions that are part of
12050 EC_GFp_simple_method, but uses Montgomery arithmetic.
12052 *Bodo Moeller; point addition and point doubling
12053 implementation directly derived from source code provided by
12054 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12056 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12057 crypto/ec/ec_lib.c):
12059 Curves are EC_GROUP objects (with an optional group generator)
12060 based on EC_METHODs that are built into the library.
12062 Points are EC_POINT objects based on EC_GROUP objects.
12064 Most of the framework would be able to handle curves over arbitrary
12065 finite fields, but as there are no obvious types for fields other
12066 than GF(p), some functions are limited to that for now.
12070 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12071 that the file contains a complete HTTP response.
12075 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12076 change the def and num file printf format specifier from "%-40sXXX"
12077 to "%-39s XXX". The latter will always guarantee a space after the
12078 field while the former will cause them to run together if the field
12079 is 40 of more characters long.
12083 * Constify the cipher and digest 'method' functions and structures
12084 and modify related functions to take constant EVP_MD and EVP_CIPHER
12089 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12090 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12094 * Modify `EVP_Digest*()` routines so they now return values. Although the
12095 internal software routines can never fail additional hardware versions
12100 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12102 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12103 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12106 ERR_R_NESTED_ASN1_ERROR
12108 ERR_R_MISSING_ASN1_EOS
12109 were 4 .. 9, conflicting with
12110 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12112 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12113 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12115 Add new error code 'ERR_R_INTERNAL_ERROR'.
12119 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12124 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12125 sets the subject name for a new request or supersedes the
12126 subject name in a given request. Formats that can be parsed are
12127 'CN=Some Name, OU=myOU, C=IT'
12129 'CN=Some Name/OU=myOU/C=IT'.
12131 Add options '-batch' and '-verbose' to 'openssl req'.
12133 *Massimiliano Pala <madwolf@hackmasters.net>*
12135 * Introduce the possibility to access global variables through
12136 functions on platform were that's the best way to handle exporting
12137 global variables in shared libraries. To enable this functionality,
12138 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12139 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12140 is normally done by Configure or something similar).
12142 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12143 in the source file (foo.c) like this:
12145 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12146 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12148 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12149 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12151 OPENSSL_DECLARE_GLOBAL(int,foo);
12152 #define foo OPENSSL_GLOBAL_REF(foo)
12153 OPENSSL_DECLARE_GLOBAL(double,bar);
12154 #define bar OPENSSL_GLOBAL_REF(bar)
12156 The #defines are very important, and therefore so is including the
12157 header file everywhere where the defined globals are used.
12159 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12160 of ASN.1 items, but that structure is a bit different.
12162 The largest change is in util/mkdef.pl which has been enhanced with
12163 better and easier to understand logic to choose which symbols should
12164 go into the Windows .def files as well as a number of fixes and code
12165 cleanup (among others, algorithm keywords are now sorted
12166 lexicographically to avoid constant rewrites).
12170 * In BN_div() keep a copy of the sign of 'num' before writing the
12171 result to 'rm' because if rm==num the value will be overwritten
12172 and produce the wrong result if 'num' is negative: this caused
12173 problems with BN_mod() and BN_nnmod().
12177 * Function OCSP_request_verify(). This checks the signature on an
12178 OCSP request and verifies the signer certificate. The signer
12179 certificate is just checked for a generic purpose and OCSP request
12184 * Add OCSP_check_validity() function to check the validity of OCSP
12185 responses. OCSP responses are prepared in real time and may only
12186 be a few seconds old. Simply checking that the current time lies
12187 between thisUpdate and nextUpdate max reject otherwise valid responses
12188 caused by either OCSP responder or client clock inaccuracy. Instead
12189 we allow thisUpdate and nextUpdate to fall within a certain period of
12190 the current time. The age of the response can also optionally be
12191 checked. Two new options -validity_period and -status_age added to
12196 * If signature or public key algorithm is unrecognized print out its
12197 OID rather that just UNKNOWN.
12201 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12202 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12203 ID to be generated from the issuer certificate alone which can then be
12204 passed to OCSP_id_issuer_cmp().
12208 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12209 ASN1 modules to export functions returning ASN1_ITEM pointers
12210 instead of the ASN1_ITEM structures themselves. This adds several
12211 new macros which allow the underlying ASN1 function/structure to
12212 be accessed transparently. As a result code should not use ASN1_ITEM
12213 references directly (such as &X509_it) but instead use the relevant
12214 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12215 use of the new ASN1 code on platforms where exporting structures
12216 is problematical (for example in shared libraries) but exporting
12217 functions returning pointers to structures is not.
12221 * Add support for overriding the generation of SSL/TLS session IDs.
12222 These callbacks can be registered either in an SSL_CTX or per SSL.
12223 The purpose of this is to allow applications to control, if they wish,
12224 the arbitrary values chosen for use as session IDs, particularly as it
12225 can be useful for session caching in multiple-server environments. A
12226 command-line switch for testing this (and any client code that wishes
12227 to use such a feature) has been added to "s_server".
12229 *Geoff Thorpe, Lutz Jaenicke*
12231 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12232 of the form '#if defined(...) || defined(...) || ...' and
12233 '#if !defined(...) && !defined(...) && ...'. This also avoids
12234 the growing number of special cases it was previously handling.
12238 * Make all configuration macros available for application by making
12239 sure they are available in opensslconf.h, by giving them names starting
12240 with `OPENSSL_` to avoid conflicts with other packages and by making
12241 sure e_os2.h will cover all platform-specific cases together with
12243 Additionally, it is now possible to define configuration/platform-
12244 specific names (called "system identities"). In the C code, these
12245 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12246 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12247 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12252 * New option -set_serial to 'req' and 'x509' this allows the serial
12253 number to use to be specified on the command line. Previously self
12254 signed certificates were hard coded with serial number 0 and the
12255 CA options of 'x509' had to use a serial number in a file which was
12260 * New options to 'ca' utility to support V2 CRL entry extensions.
12261 Currently CRL reason, invalidity date and hold instruction are
12262 supported. Add new CRL extensions to V3 code and some new objects.
12266 * New function EVP_CIPHER_CTX_set_padding() this is used to
12267 disable standard block padding (aka PKCS#5 padding) in the EVP
12268 API, which was previously mandatory. This means that the data is
12269 not padded in any way and so the total length much be a multiple
12270 of the block size, otherwise an error occurs.
12274 * Initial (incomplete) OCSP SSL support.
12278 * New function OCSP_parse_url(). This splits up a URL into its host,
12279 port and path components: primarily to parse OCSP URLs. New -url
12280 option to ocsp utility.
12284 * New nonce behavior. The return value of OCSP_check_nonce() now
12285 reflects the various checks performed. Applications can decide
12286 whether to tolerate certain situations such as an absent nonce
12287 in a response when one was present in a request: the ocsp application
12288 just prints out a warning. New function OCSP_add1_basic_nonce()
12289 this is to allow responders to include a nonce in a response even if
12290 the request is nonce-less.
12294 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12295 skipped when using openssl x509 multiple times on a single input file,
12296 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12300 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12301 set string type: to handle setting ASN1_TIME structures. Fix ca
12302 utility to correctly initialize revocation date of CRLs.
12306 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12307 the clients preferred ciphersuites and rather use its own preferences.
12308 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12309 Internet Explorer by ensuring unchanged hash method during stepup.
12310 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12314 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12315 to aes and add a new 'exist' option to print out symbols that don't
12320 * Additional options to ocsp utility to allow flags to be set and
12321 additional certificates supplied.
12325 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12326 OCSP client a number of certificate to only verify the response
12331 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12332 handle the new API. Currently only ECB, CBC modes supported. Add new
12335 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12336 Encryption Standard (AES) Ciphersuites for Transport Layer
12337 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12338 not enabled by default and were not part of the "ALL" ciphersuite
12339 alias because they were not yet official; they could be
12340 explicitly requested by specifying the "AESdraft" ciphersuite
12341 group alias. In the final release of OpenSSL 0.9.7, the group
12342 alias is called "AES" and is part of "ALL".)
12344 *Ben Laurie, Steve Henson, Bodo Moeller*
12346 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12347 request to response.
12351 * Functions for OCSP responders. OCSP_request_onereq_count(),
12352 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12353 extract information from a certificate request. OCSP_response_create()
12354 creates a response and optionally adds a basic response structure.
12355 OCSP_basic_add1_status() adds a complete single response to a basic
12356 response and returns the OCSP_SINGLERESP structure just added (to allow
12357 extensions to be included for example). OCSP_basic_add1_cert() adds a
12358 certificate to a basic response and OCSP_basic_sign() signs a basic
12359 response with various flags. New helper functions ASN1_TIME_check()
12360 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12361 (converts ASN1_TIME to GeneralizedTime).
12365 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12366 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12367 structure from a certificate. X509_pubkey_digest() digests the public_key
12368 contents: this is used in various key identifiers.
12372 * Make sk_sort() tolerate a NULL argument.
12374 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12376 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12377 passed by the function are trusted implicitly. If any of them signed the
12378 response then it is assumed to be valid and is not verified.
12382 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12383 to data. This was previously part of the PKCS7 ASN1 code. This
12384 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12385 *Steve Henson, reported by Kenneth R. Robinette
12386 <support@securenetterm.com>*
12388 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12389 routines: without these tracing memory leaks is very painful.
12390 Fix leaks in PKCS12 and PKCS7 routines.
12394 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12395 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12396 effectively meant GeneralizedTime would never be used. Now it
12397 is initialised to -1 but X509_time_adj() now has to check the value
12398 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12399 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12400 *Steve Henson, reported by Kenneth R. Robinette
12401 <support@securenetterm.com>*
12403 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12404 result in a zero length in the ASN1_INTEGER structure which was
12405 not consistent with the structure when d2i_ASN1_INTEGER() was used
12406 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12407 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12408 where it did not print out a minus for negative ASN1_INTEGER.
12412 * Add summary printout to ocsp utility. The various functions which
12413 convert status values to strings have been renamed to:
12414 OCSP_response_status_str(), OCSP_cert_status_str() and
12415 OCSP_crl_reason_str() and are no longer static. New options
12416 to verify nonce values and to disable verification. OCSP response
12417 printout format cleaned up.
12421 * Add additional OCSP certificate checks. These are those specified
12422 in RFC2560. This consists of two separate checks: the CA of the
12423 certificate being checked must either be the OCSP signer certificate
12424 or the issuer of the OCSP signer certificate. In the latter case the
12425 OCSP signer certificate must contain the OCSP signing extended key
12426 usage. This check is performed by attempting to match the OCSP
12427 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12428 in the OCSP_CERTID structures of the response.
12432 * Initial OCSP certificate verification added to OCSP_basic_verify()
12433 and related routines. This uses the standard OpenSSL certificate
12434 verify routines to perform initial checks (just CA validity) and
12435 to obtain the certificate chain. Then additional checks will be
12436 performed on the chain. Currently the root CA is checked to see
12437 if it is explicitly trusted for OCSP signing. This is used to set
12438 a root CA as a global signing root: that is any certificate that
12439 chains to that CA is an acceptable OCSP signing certificate.
12443 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12444 extensions from a separate configuration file.
12445 As when reading extensions from the main configuration file,
12446 the '-extensions ...' option may be used for specifying the
12449 *Massimiliano Pala <madwolf@comune.modena.it>*
12451 * New OCSP utility. Allows OCSP requests to be generated or
12452 read. The request can be sent to a responder and the output
12453 parsed, outputted or printed in text form. Not complete yet:
12454 still needs to check the OCSP response validity.
12458 * New subcommands for 'openssl ca':
12459 `openssl ca -status <serial>` prints the status of the cert with
12460 the given serial number (according to the index file).
12461 `openssl ca -updatedb` updates the expiry status of certificates
12464 *Massimiliano Pala <madwolf@comune.modena.it>*
12466 * New '-newreq-nodes' command option to CA.pl. This is like
12467 '-newreq', but calls 'openssl req' with the '-nodes' option
12468 so that the resulting key is not encrypted.
12470 *Damien Miller <djm@mindrot.org>*
12472 * New configuration for the GNU Hurd.
12474 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12476 * Initial code to implement OCSP basic response verify. This
12477 is currently incomplete. Currently just finds the signer's
12478 certificate and verifies the signature on the response.
12482 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12483 value of OPENSSLDIR. This is available via the new '-d' option
12484 to 'openssl version', and is also included in 'openssl version -a'.
12488 * Allowing defining memory allocation callbacks that will be given
12489 file name and line number information in additional arguments
12490 (a `const char*` and an int). The basic functionality remains, as
12491 well as the original possibility to just replace malloc(),
12492 realloc() and free() by functions that do not know about these
12493 additional arguments. To register and find out the current
12494 settings for extended allocation functions, the following
12495 functions are provided:
12497 CRYPTO_set_mem_ex_functions
12498 CRYPTO_set_locked_mem_ex_functions
12499 CRYPTO_get_mem_ex_functions
12500 CRYPTO_get_locked_mem_ex_functions
12502 These work the same way as CRYPTO_set_mem_functions and friends.
12503 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12504 extended allocation function is enabled.
12505 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12506 a conventional allocation function is enabled.
12508 *Richard Levitte, Bodo Moeller*
12510 * Finish off removing the remaining LHASH function pointer casts.
12511 There should no longer be any prototype-casting required when using
12512 the LHASH abstraction, and any casts that remain are "bugs". See
12513 the callback types and macros at the head of lhash.h for details
12514 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12518 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12519 If /dev/[u]random devices are not available or do not return enough
12520 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12522 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12523 /etc/entropy will be queried once each in this sequence, querying stops
12524 when enough entropy was collected without querying more sockets.
12528 * Change the Unix RAND_poll() variant to be able to poll several
12529 random devices, as specified by DEVRANDOM, until a sufficient amount
12530 of data has been collected. We spend at most 10 ms on each file
12531 (select timeout) and read in non-blocking mode. DEVRANDOM now
12532 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12533 (previously it was just the string "/dev/urandom"), so on typical
12534 platforms the 10 ms delay will never occur.
12535 Also separate out the Unix variant to its own file, rand_unix.c.
12536 For VMS, there's a currently-empty rand_vms.c.
12540 * Move OCSP client related routines to ocsp_cl.c. These
12541 provide utility functions which an application needing
12542 to issue a request to an OCSP responder and analyse the
12543 response will typically need: as opposed to those which an
12544 OCSP responder itself would need which will be added later.
12546 OCSP_request_sign() signs an OCSP request with an API similar
12547 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12548 response. OCSP_response_get1_basic() extracts basic response
12549 from response. OCSP_resp_find_status(): finds and extracts status
12550 information from an OCSP_CERTID structure (which will be created
12551 when the request structure is built). These are built from lower
12552 level functions which work on OCSP_SINGLERESP structures but
12553 won't normally be used unless the application wishes to examine
12554 extensions in the OCSP response for example.
12556 Replace nonce routines with a pair of functions.
12557 OCSP_request_add1_nonce() adds a nonce value and optionally
12558 generates a random value. OCSP_check_nonce() checks the
12559 validity of the nonce in an OCSP response.
12563 * Change function OCSP_request_add() to OCSP_request_add0_id().
12564 This doesn't copy the supplied OCSP_CERTID and avoids the
12565 need to free up the newly created id. Change return type
12566 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12567 This can then be used to add extensions to the request.
12568 Deleted OCSP_request_new(), since most of its functionality
12569 is now in OCSP_REQUEST_new() (and the case insensitive name
12570 clash) apart from the ability to set the request name which
12571 will be added elsewhere.
12575 * Update OCSP API. Remove obsolete extensions argument from
12576 various functions. Extensions are now handled using the new
12577 OCSP extension code. New simple OCSP HTTP function which
12578 can be used to send requests and parse the response.
12582 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12583 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12584 uses the special reorder version of SET OF to sort the attributes
12585 and reorder them to match the encoded order. This resolves a long
12586 standing problem: a verify on a PKCS7 structure just after signing
12587 it used to fail because the attribute order did not match the
12588 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12589 it uses the received order. This is necessary to tolerate some broken
12590 software that does not order SET OF. This is handled by encoding
12591 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12592 to produce the required SET OF.
12596 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12597 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12598 files to get correct declarations of the ASN.1 item variables.
12602 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12603 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12604 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12605 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12606 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12607 ASN1_ITEM and no wrapper functions.
12611 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12612 replace the old function pointer based I/O routines. Change most of
12613 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12617 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12618 lines, recognize more "algorithms" that can be deselected, and make
12619 it complain about algorithm deselection that isn't recognised.
12623 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12624 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12625 to use new functions. Add NO_ASN1_OLD which can be set to remove
12626 some old style ASN1 functions: this can be used to determine if old
12627 code will still work when these eventually go away.
12631 * New extension functions for OCSP structures, these follow the
12632 same conventions as certificates and CRLs.
12636 * New function X509V3_add1_i2d(). This automatically encodes and
12637 adds an extension. Its behaviour can be customised with various
12638 flags to append, replace or delete. Various wrappers added for
12639 certificates and CRLs.
12643 * Fix to avoid calling the underlying ASN1 print routine when
12644 an extension cannot be parsed. Correct a typo in the
12645 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12649 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12650 entries for variables.
12654 * Add functionality to apps/openssl.c for detecting locking
12655 problems: As the program is single-threaded, all we have
12656 to do is register a locking callback using an array for
12657 storing which locks are currently held by the program.
12661 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12662 SSL_get_ex_data_X509_STORE_idx(), which is used in
12663 ssl_verify_cert_chain() and thus can be called at any time
12664 during TLS/SSL handshakes so that thread-safety is essential.
12665 Unfortunately, the ex_data design is not at all suited
12666 for multi-threaded use, so it probably should be abolished.
12670 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12672 *Broadcom, tweaked and integrated by Geoff Thorpe*
12674 * Move common extension printing code to new function
12675 X509V3_print_extensions(). Reorganise OCSP print routines and
12676 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12680 * New function X509_signature_print() to remove duplication in some
12685 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12686 set (this was treated exactly the same as SET OF previously). This
12687 is used to reorder the STACK representing the structure to match the
12688 encoding. This will be used to get round a problem where a PKCS7
12689 structure which was signed could not be verified because the STACK
12690 order did not reflect the encoded order.
12694 * Reimplement the OCSP ASN1 module using the new code.
12698 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12699 for its ASN1 operations. The old style function pointers still exist
12700 for now but they will eventually go away.
12704 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12705 completely replaces the old ASN1 functionality with a table driven
12706 encoder and decoder which interprets an ASN1_ITEM structure describing
12707 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12708 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12709 has also been converted to the new form.
12713 * Change BN_mod_exp_recp so that negative moduli are tolerated
12714 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12715 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12716 for negative moduli.
12720 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12721 of not touching the result's sign bit.
12725 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12730 * Changed the LHASH code to use prototypes for callbacks, and created
12731 macros to declare and implement thin (optionally static) functions
12732 that provide type-safety and avoid function pointer casting for the
12733 type-specific callbacks.
12737 * Added Kerberos Cipher Suites to be used with TLS, as written in
12739 *Veers Staats <staatsvr@asc.hpc.mil>,
12740 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12742 * Reformat the FAQ so the different questions and answers can be divided
12743 in sections depending on the subject.
12747 * Have the zlib compression code load ZLIB.DLL dynamically under
12752 * New function BN_mod_sqrt for computing square roots modulo a prime
12753 (using the probabilistic Tonelli-Shanks algorithm unless
12754 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12755 be handled deterministically).
12757 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12759 * Make BN_mod_inverse faster by explicitly handling small quotients
12760 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12761 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12765 * New function BN_kronecker.
12769 * Fix BN_gcd so that it works on negative inputs; the result is
12770 positive unless both parameters are zero.
12771 Previously something reasonably close to an infinite loop was
12772 possible because numbers could be growing instead of shrinking
12773 in the implementation of Euclid's algorithm.
12777 * Fix BN_is_word() and BN_is_one() macros to take into account the
12778 sign of the number in question.
12780 Fix BN_is_word(a,w) to work correctly for w == 0.
12782 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12783 because its test if the absolute value of 'a' equals 'w'.
12784 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12785 it exists mostly for use in the implementations of BN_is_zero(),
12786 BN_is_one(), and BN_is_word().
12790 * New function BN_swap.
12794 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12795 the exponentiation functions are more likely to produce reasonable
12796 results on negative inputs.
12800 * Change BN_mod_mul so that the result is always non-negative.
12801 Previously, it could be negative if one of the factors was negative;
12802 I don't think anyone really wanted that behaviour.
12806 * Move `BN_mod_...` functions into new file crypto/bn/bn_mod.c
12807 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12808 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12809 and add new functions:
12818 BN_mod_lshift1_quick
12820 BN_mod_lshift_quick
12822 These functions always generate non-negative results.
12824 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12825 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12827 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12828 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12829 be reduced modulo m.
12831 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12834 The following entry accidentally appeared in the CHANGES file
12835 distributed with OpenSSL 0.9.7. The modifications described in
12836 it do *not* apply to OpenSSL 0.9.7.
12838 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12839 was actually never needed) and in BN_mul(). The removal in BN_mul()
12840 required a small change in bn_mul_part_recursive() and the addition
12841 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12842 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12843 bn_sub_words() and bn_add_words() except they take arrays with
12849 * In 'openssl passwd', verify passwords read from the terminal
12850 unless the '-salt' option is used (which usually means that
12851 verification would just waste user's time since the resulting
12852 hash is going to be compared with some given password hash)
12853 or the new '-noverify' option is used.
12855 This is an incompatible change, but it does not affect
12856 non-interactive use of 'openssl passwd' (passwords on the command
12857 line, '-stdin' option, '-in ...' option) and thus should not
12858 cause any problems.
12862 * Remove all references to RSAref, since there's no more need for it.
12866 * Make DSO load along a path given through an environment variable
12867 (SHLIB_PATH) with shl_load().
12871 * Constify the ENGINE code as a result of BIGNUM constification.
12872 Also constify the RSA code and most things related to it. In a
12873 few places, most notable in the depth of the ASN.1 code, ugly
12874 casts back to non-const were required (to be solved at a later
12879 * Make it so the openssl application has all engines loaded by default.
12883 * Constify the BIGNUM routines a little more.
12887 * Add the following functions:
12889 ENGINE_load_cswift()
12891 ENGINE_load_atalla()
12892 ENGINE_load_nuron()
12893 ENGINE_load_builtin_engines()
12895 That way, an application can itself choose if external engines that
12896 are built-in in OpenSSL shall ever be used or not. The benefit is
12897 that applications won't have to be linked with libdl or other dso
12898 libraries unless it's really needed.
12900 Changed 'openssl engine' to load all engines on demand.
12901 Changed the engine header files to avoid the duplication of some
12902 declarations (they differed!).
12906 * 'openssl engine' can now list capabilities.
12910 * Better error reporting in 'openssl engine'.
12914 * Never call load_dh_param(NULL) in s_server.
12918 * Add engine application. It can currently list engines by name and
12919 identity, and test if they are actually available.
12923 * Improve RPM specification file by forcing symbolic linking and making
12924 sure the installed documentation is also owned by root.root.
12926 *Damien Miller <djm@mindrot.org>*
12928 * Give the OpenSSL applications more possibilities to make use of
12929 keys (public as well as private) handled by engines.
12933 * Add OCSP code that comes from CertCo.
12937 * Add VMS support for the Rijndael code.
12941 * Added untested support for Nuron crypto accelerator.
12945 * Add support for external cryptographic devices. This code was
12946 previously distributed separately as the "engine" branch.
12948 *Geoff Thorpe, Richard Levitte*
12950 * Rework the filename-translation in the DSO code. It is now possible to
12951 have far greater control over how a "name" is turned into a filename
12952 depending on the operating environment and any oddities about the
12953 different shared library filenames on each system.
12957 * Support threads on FreeBSD-elf in Configure.
12961 * Fix for SHA1 assembly problem with MASM: it produces
12962 warnings about corrupt line number information when assembling
12963 with debugging information. This is caused by the overlapping
12966 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12969 NCONF_get_number() has no error checking at all. As a replacement,
12970 NCONF_get_number_e() is defined (`_e` for "error checking") and is
12971 promoted strongly. The old NCONF_get_number is kept around for
12972 binary backward compatibility.
12973 Make it possible for methods to load from something other than a BIO,
12974 by providing a function pointer that is given a name instead of a BIO.
12975 For example, this could be used to load configuration data from an
12980 * Fix for non blocking accept BIOs. Added new I/O special reason
12981 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12982 with non blocking I/O was not possible because no retry code was
12983 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12988 * Added the beginnings of Rijndael support.
12992 * Fix for bug in DirectoryString mask setting. Add support for
12993 X509_NAME_print_ex() in 'req' and X509_print_ex() function
12994 to allow certificate printing to more controllable, additional
12995 'certopt' option to 'x509' to allow new printing options to be
13000 * Clean old EAY MD5 hack from e_os.h.
13004 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13006 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13007 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13009 *Joe Orton, Steve Henson*
13011 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13013 * Fix additional bug revealed by the NISCC test suite:
13015 Stop bug triggering large recursion when presented with
13016 certain ASN.1 tags [CVE-2003-0851][]
13020 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13022 * Fix various bugs revealed by running the NISCC test suite:
13024 Stop out of bounds reads in the ASN1 code when presented with
13025 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13027 If verify callback ignores invalid public key errors don't try to check
13028 certificate signature with the NULL public key.
13032 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13033 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13038 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13039 extra data after the compression methods not only for TLS 1.0
13040 but also for SSL 3.0 (as required by the specification).
13042 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13044 * Change X509_certificate_type() to mark the key as exported/exportable
13045 when it's 512 *bits* long, not 512 bytes.
13049 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13051 * Countermeasure against the Klima-Pokorny-Rosa extension of
13052 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13053 a protocol version number mismatch like a decryption error
13054 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13058 * Turn on RSA blinding by default in the default implementation
13059 to avoid a timing attack. Applications that don't want it can call
13060 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13061 They would be ill-advised to do so in most cases.
13063 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13065 * Change RSA blinding code so that it works when the PRNG is not
13066 seeded (in this case, the secret RSA exponent is abused as
13067 an unpredictable seed -- if it is not unpredictable, there
13068 is no point in blinding anyway). Make RSA blinding thread-safe
13069 by remembering the creator's thread ID in rsa->blinding and
13070 having all other threads use local one-time blinding factors
13071 (this requires more computation than sharing rsa->blinding, but
13072 avoids excessive locking; and if an RSA object is not shared
13073 between threads, blinding will still be very fast).
13077 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13079 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13080 via timing by performing a MAC computation even if incorrect
13081 block cipher padding has been found. This is a countermeasure
13082 against active attacks where the attacker has to distinguish
13083 between bad padding and a MAC verification error. [CVE-2003-0078][]
13085 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13086 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13087 Martin Vuagnoux (EPFL, Ilion)*
13089 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13091 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13092 memory from its contents. This is done with a counter that will
13093 place alternating values in each byte. This can be used to solve
13094 two issues: 1) the removal of calls to memset() by highly optimizing
13095 compilers, and 2) cleansing with other values than 0, since those can
13096 be read through on certain media, for example a swap space on disk.
13100 * Bugfix: client side session caching did not work with external caching,
13101 because the session->cipher setting was not restored when reloading
13102 from the external cache. This problem was masked, when
13103 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13104 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13108 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13109 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13111 *Zeev Lieber <zeev-l@yahoo.com>*
13113 * Undo an undocumented change introduced in 0.9.6e which caused
13114 repeated calls to OpenSSL_add_all_ciphers() and
13115 OpenSSL_add_all_digests() to be ignored, even after calling
13120 * Change the default configuration reader to deal with last line not
13121 being properly terminated.
13125 * Change X509_NAME_cmp() so it applies the special rules on handling
13126 DN values that are of type PrintableString, as well as RDNs of type
13127 emailAddress where the value has the type ia5String.
13129 *stefank@valicert.com via Richard Levitte*
13131 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13132 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13133 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13134 the bitwise-OR of the two for use by the majority of applications
13135 wanting this behaviour, and update the docs. The documented
13136 behaviour and actual behaviour were inconsistent and had been
13137 changing anyway, so this is more a bug-fix than a behavioural
13140 *Geoff Thorpe, diagnosed by Nadav Har'El*
13142 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13143 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13147 * Fix initialization code race conditions in
13148 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13149 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13150 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13151 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13152 ssl2_get_cipher_by_char(),
13153 ssl3_get_cipher_by_char().
13155 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13157 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13158 the cached sessions are flushed, as the remove_cb() might use ex_data
13159 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13160 (see [openssl.org #212]).
13162 *Geoff Thorpe, Lutz Jaenicke*
13164 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13165 length, instead of the encoding length to d2i_ASN1_OBJECT.
13169 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13171 * [In 0.9.6g-engine release:]
13172 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13174 *Lynn Gazis <lgazis@rainbow.com>*
13176 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13178 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13179 and get fix the header length calculation.
13180 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13181 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13183 * Use proper error handling instead of 'assertions' in buffer
13184 overflow checks added in 0.9.6e. This prevents DoS (the
13185 assertions could call abort()).
13187 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13189 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13191 * Add various sanity checks to asn1_get_length() to reject
13192 the ASN1 length bytes if they exceed sizeof(long), will appear
13193 negative or the content length exceeds the length of the
13196 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13198 * Fix cipher selection routines: ciphers without encryption had no flags
13199 for the cipher strength set and where therefore not handled correctly
13200 by the selection routines (PR #130).
13204 * Fix EVP_dsa_sha macro.
13209 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13210 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13211 that was added in OpenSSL 0.9.6d.
13213 As the countermeasure turned out to be incompatible with some
13214 broken SSL implementations, the new option is part of SSL_OP_ALL.
13215 SSL_OP_ALL is usually employed when compatibility with weird SSL
13216 implementations is desired (e.g. '-bugs' option to 's_client' and
13217 's_server'), so the new option is automatically set in many
13222 * Changes in security patch:
13224 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13225 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13226 Air Force Materiel Command, USAF, under agreement number
13229 * Add various sanity checks to asn1_get_length() to reject
13230 the ASN1 length bytes if they exceed sizeof(long), will appear
13231 negative or the content length exceeds the length of the
13232 supplied buffer. [CVE-2002-0659][]
13234 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13236 * Assertions for various potential buffer overflows, not known to
13237 happen in practice.
13239 *Ben Laurie (CHATS)*
13241 * Various temporary buffers to hold ASCII versions of integers were
13242 too small for 64 bit platforms. [CVE-2002-0655][]
13243 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13245 * Remote buffer overflow in SSL3 protocol - an attacker could
13246 supply an oversized session ID to a client. [CVE-2002-0656][]
13248 *Ben Laurie (CHATS)*
13250 * Remote buffer overflow in SSL2 protocol - an attacker could
13251 supply an oversized client master key. [CVE-2002-0656][]
13253 *Ben Laurie (CHATS)*
13255 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13257 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13258 encoded as NULL) with id-dsa-with-sha1.
13260 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13262 * Check various `X509_...()` return values in apps/req.c.
13264 *Nils Larsch <nla@trustcenter.de>*
13266 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13267 an end-of-file condition would erroneously be flagged, when the CRLF
13268 was just at the end of a processed block. The bug was discovered when
13269 processing data through a buffering memory BIO handing the data to a
13270 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13271 <ptsekov@syntrex.com> and Nedelcho Stanev.
13275 * Implement a countermeasure against a vulnerability recently found
13276 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13277 before application data chunks to avoid the use of known IVs
13278 with data potentially chosen by the attacker.
13282 * Fix length checks in ssl3_get_client_hello().
13286 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13287 to prevent ssl3_read_internal() from incorrectly assuming that
13288 ssl3_read_bytes() found application data while handshake
13289 processing was enabled when in fact s->s3->in_read_app_data was
13290 merely automatically cleared during the initial handshake.
13292 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13294 * Fix object definitions for Private and Enterprise: they were not
13295 recognized in their shortname (=lowercase) representation. Extend
13296 obj_dat.pl to issue an error when using undefined keywords instead
13297 of silently ignoring the problem (Svenning Sorensen
13298 <sss@sss.dnsalias.net>).
13302 * Fix DH_generate_parameters() so that it works for 'non-standard'
13303 generators, i.e. generators other than 2 and 5. (Previously, the
13304 code did not properly initialise the 'add' and 'rem' values to
13305 BN_generate_prime().)
13307 In the new general case, we do not insist that 'generator' is
13308 actually a primitive root: This requirement is rather pointless;
13309 a generator of the order-q subgroup is just as good, if not
13314 * Map new X509 verification errors to alerts. Discovered and submitted by
13315 Tom Wu <tom@arcot.com>.
13319 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13320 returning non-zero before the data has been completely received
13321 when using non-blocking I/O.
13323 *Bodo Moeller; problem pointed out by John Hughes*
13325 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13327 *Ben Laurie, Lutz Jaenicke*
13329 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13330 Yoram Zahavi <YoramZ@gilian.com>).
13334 * Add information about CygWin 1.3 and on, and preserve proper
13335 configuration for the versions before that.
13337 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13339 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13340 check whether we deal with a copy of a session and do not delete from
13341 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13342 <izhar@checkpoint.com>.
13346 * Do not store session data into the internal session cache, if it
13347 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13348 flag is set). Proposed by Aslam <aslam@funk.com>.
13352 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13357 * [In 0.9.6d-engine release:]
13358 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13360 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13362 * Add the configuration target linux-s390x.
13364 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13366 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13367 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13368 variable as an indication that a ClientHello message has been
13369 received. As the flag value will be lost between multiple
13370 invocations of ssl3_accept when using non-blocking I/O, the
13371 function may not be aware that a handshake has actually taken
13372 place, thus preventing a new session from being added to the
13375 To avoid this problem, we now set s->new_session to 2 instead of
13376 using a local variable.
13378 *Lutz Jaenicke, Bodo Moeller*
13380 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13381 if the SSL_R_LENGTH_MISMATCH error is detected.
13383 *Geoff Thorpe, Bodo Moeller*
13385 * New 'shared_ldflag' column in Configure platform table.
13389 * Fix EVP_CIPHER_mode macro.
13391 *"Dan S. Camper" <dan@bti.net>*
13393 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13394 type, we must throw them away by setting rr->length to 0.
13396 *D P Chang <dpc@qualys.com>*
13398 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13400 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13401 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13402 worked incorrectly for those cases where range = `10..._2` and
13403 `3*range` is two bits longer than range.)
13407 * Only add signing time to PKCS7 structures if it is not already
13412 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13413 OBJ_ld_ce should be OBJ_id_ce.
13414 Also some ip-pda OIDs in crypto/objects/objects.txt were
13415 incorrect (cf. RFC 3039).
13417 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13419 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13420 returns early because it has nothing to do.
13422 *Andy Schneider <andy.schneider@bjss.co.uk>*
13424 * [In 0.9.6c-engine release:]
13425 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13427 *Andy Schneider <andy.schneider@bjss.co.uk>*
13429 * [In 0.9.6c-engine release:]
13430 Add support for Cryptographic Appliance's keyserver technology.
13431 (Use engine 'keyclient')
13433 *Cryptographic Appliances and Geoff Thorpe*
13435 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13436 is called via tools/c89.sh because arguments have to be
13437 rearranged (all '-L' options must appear before the first object
13440 *Richard Shapiro <rshapiro@abinitio.com>*
13442 * [In 0.9.6c-engine release:]
13443 Add support for Broadcom crypto accelerator cards, backported
13446 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13448 * [In 0.9.6c-engine release:]
13449 Add support for SureWare crypto accelerator cards from
13450 Baltimore Technologies. (Use engine 'sureware')
13452 *Baltimore Technologies and Mark Cox*
13454 * [In 0.9.6c-engine release:]
13455 Add support for crypto accelerator cards from Accelerated
13456 Encryption Processing, www.aep.ie. (Use engine 'aep')
13458 *AEP Inc. and Mark Cox*
13460 * Add a configuration entry for gcc on UnixWare.
13462 *Gary Benson <gbenson@redhat.com>*
13464 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13465 messages are stored in a single piece (fixed-length part and
13466 variable-length part combined) and fix various bugs found on the way.
13470 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13471 instead. BIO_gethostbyname() does not know what timeouts are
13472 appropriate, so entries would stay in cache even when they have
13474 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13476 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13477 faced with a pathologically small ClientHello fragment that does
13478 not contain client_version: Instead of aborting with an error,
13479 simply choose the highest available protocol version (i.e.,
13480 TLS 1.0 unless it is disabled). In practice, ClientHello
13481 messages are never sent like this, but this change gives us
13482 strictly correct behaviour at least for TLS.
13486 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13487 never resets s->method to s->ctx->method when called from within
13488 one of the SSL handshake functions.
13490 *Bodo Moeller; problem pointed out by Niko Baric*
13492 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13493 (sent using the client's version number) if client_version is
13494 smaller than the protocol version in use. Also change
13495 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13496 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13497 the client will at least see that alert.
13501 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13506 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13507 client receives HelloRequest while in a handshake.
13509 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13511 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13512 should end in 'break', not 'goto end' which circumvents various
13513 cleanups done in state SSL_ST_OK. But session related stuff
13514 must be disabled for SSL_ST_OK in the case that we just sent a
13517 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13518 before just sending a HelloRequest.
13520 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13522 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13523 reveal whether illegal block cipher padding was found or a MAC
13524 verification error occurred. (Neither SSLerr() codes nor alerts
13525 are directly visible to potential attackers, but the information
13526 may leak via logfiles.)
13528 Similar changes are not required for the SSL 2.0 implementation
13529 because the number of padding bytes is sent in clear for SSL 2.0,
13530 and the extra bytes are just ignored. However ssl/s2_pkt.c
13531 failed to verify that the purported number of padding bytes is in
13536 * Add OpenUNIX-8 support including shared libraries
13537 (Boyd Lynn Gerber <gerberb@zenez.com>).
13541 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13542 'wristwatch attack' using huge encoding parameters (cf.
13543 James H. Manger's CRYPTO 2001 paper). Note that the
13544 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13545 encoding parameters and hence was not vulnerable.
13549 * BN_sqr() bug fix.
13551 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13553 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13554 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13555 followed by modular reduction.
13557 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13559 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13560 equivalent based on BN_pseudo_rand() instead of BN_rand().
13564 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13565 This function was broken, as the check for a new client hello message
13566 to handle SGC did not allow these large messages.
13567 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13571 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13575 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13576 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13580 * Rework the configuration and shared library support for Tru64 Unix.
13581 The configuration part makes use of modern compiler features and
13582 still retains old compiler behavior for those that run older versions
13583 of the OS. The shared library support part includes a variant that
13584 uses the RPATH feature, and is available through the special
13585 configuration target "alpha-cc-rpath", which will never be selected
13588 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13590 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13591 with the same message size as in ssl3_get_certificate_request().
13592 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13593 messages might inadvertently be reject as too long.
13595 *Petr Lampa <lampa@fee.vutbr.cz>*
13597 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13601 * Modified SSL library such that the verify_callback that has been set
13602 specifically for an SSL object with SSL_set_verify() is actually being
13603 used. Before the change, a verify_callback set with this function was
13604 ignored and the verify_callback() set in the SSL_CTX at the time of
13605 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13606 to allow the necessary settings.
13610 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13611 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13612 done automatically (in contradiction to the requirements of the C
13613 standard). This made problems when used from OpenSSH.
13617 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13618 dh->length and always used
13620 BN_rand_range(priv_key, dh->p).
13622 BN_rand_range() is not necessary for Diffie-Hellman, and this
13623 specific range makes Diffie-Hellman unnecessarily inefficient if
13624 dh->length (recommended exponent length) is much smaller than the
13625 length of dh->p. We could use BN_rand_range() if the order of
13626 the subgroup was stored in the DH structure, but we only have
13631 BN_rand(priv_key, l, ...)
13633 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13640 RSA_eay_public_encrypt
13641 RSA_eay_private_decrypt
13642 RSA_eay_private_encrypt (signing)
13643 RSA_eay_public_decrypt (signature verification)
13645 (default implementations for RSA_public_encrypt,
13646 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13647 always reject numbers >= n.
13651 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13652 to synchronize access to 'locking_thread'. This is necessary on
13653 systems where access to 'locking_thread' (an 'unsigned long'
13654 variable) is not atomic.
13658 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13659 *before* setting the 'crypto_lock_rand' flag. The previous code had
13660 a race condition if 0 is a valid thread ID.
13662 *Travis Vitek <vitek@roguewave.com>*
13664 * Add support for shared libraries under Irix.
13666 *Albert Chin-A-Young <china@thewrittenword.com>*
13668 * Add configuration option to build on Linux on both big-endian and
13669 little-endian MIPS.
13671 *Ralf Baechle <ralf@uni-koblenz.de>*
13673 * Add the possibility to create shared libraries on HP-UX.
13677 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13679 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13680 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13681 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13682 PRNG state recovery was possible based on the output of
13683 one PRNG request appropriately sized to gain knowledge on
13684 'md' followed by enough consecutive 1-byte PRNG requests
13685 to traverse all of 'state'.
13687 1. When updating 'md_local' (the current thread's copy of 'md')
13688 during PRNG output generation, hash all of the previous
13689 'md_local' value, not just the half used for PRNG output.
13691 2. Make the number of bytes from 'state' included into the hash
13692 independent from the number of PRNG bytes requested.
13694 The first measure alone would be sufficient to avoid
13695 Markku-Juhani's attack. (Actually it had never occurred
13696 to me that the half of 'md_local' used for chaining was the
13697 half from which PRNG output bytes were taken -- I had always
13698 assumed that the secret half would be used.) The second
13699 measure makes sure that additional data from 'state' is never
13700 mixed into 'md_local' in small portions; this heuristically
13701 further strengthens the PRNG.
13705 * Fix crypto/bn/asm/mips3.s.
13709 * When only the key is given to "enc", the IV is undefined. Print out
13710 an error message in this case.
13714 * Handle special case when X509_NAME is empty in X509 printing routines.
13718 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13719 positive and less than q.
13723 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13724 used: it isn't thread safe and the add_lock_callback should handle
13727 *Paul Rose <Paul.Rose@bridge.com>*
13729 * Verify that incoming data obeys the block size in
13730 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13736 *Ulf Möller, Bodo Möller*
13738 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13739 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13740 when fixing the server behaviour for backwards-compatible 'client
13741 hello' messages. (Note that the attack is impractical against
13742 SSL 3.0 and TLS 1.0 anyway because length and version checking
13743 means that the probability of guessing a valid ciphertext is
13744 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13747 Before 0.9.5, the countermeasure (hide the error by generating a
13748 random 'decryption result') did not work properly because
13749 ERR_clear_error() was missing, meaning that SSL_get_error() would
13750 detect the supposedly ignored error.
13752 Both problems are now fixed.
13756 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13757 (previously it was 1024).
13761 * Fix for compatibility mode trust settings: ignore trust settings
13762 unless some valid trust or reject settings are present.
13766 * Fix for blowfish EVP: its a variable length cipher.
13770 * Fix various bugs related to DSA S/MIME verification. Handle missing
13771 parameters in DSA public key structures and return an error in the
13772 DSA routines if parameters are absent.
13776 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13777 in the current directory if neither $RANDFILE nor $HOME was set.
13778 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13779 caused some confusion to Windows users who haven't defined $HOME.
13780 Thus RAND_file_name() is changed again: e_os.h can define a
13781 DEFAULT_HOME, which will be used if $HOME is not set.
13782 For Windows, we use "C:"; on other platforms, we still require
13783 environment variables.
13785 * Move 'if (!initialized) RAND_poll()' into regions protected by
13786 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13787 having multiple threads call RAND_poll() concurrently.
13791 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13792 combination of a flag and a thread ID variable.
13793 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13794 flag), *other* threads can enter ssleay_add_bytes without obeying
13795 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13796 that they do not hold after the first thread unsets add_do_not_lock).
13800 * Change bctest again: '-x' expressions are not available in all
13801 versions of 'test'.
13805 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13807 * Fix a couple of memory leaks in PKCS7_dataDecode()
13809 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13811 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13812 the default extension for executables, if any. Also, make the perl
13813 scripts that use symlink() to test if it really exists and use "cp"
13814 if it doesn't. All this made OpenSSL compilable and installable in
13819 * Fix for asn1_GetSequence() for indefinite length constructed data.
13820 If SEQUENCE is length is indefinite just set c->slen to the total
13821 amount of data available.
13823 *Steve Henson, reported by shige@FreeBSD.org*
13825 *This change does not apply to 0.9.7.*
13827 * Change bctest to avoid here-documents inside command substitution
13828 (workaround for FreeBSD /bin/sh bug).
13829 For compatibility with Ultrix, avoid shell functions (introduced
13830 in the bctest version that searches along $PATH).
13834 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13835 with des_encrypt() defined on some operating systems, like Solaris
13840 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13841 On the Importance of Eliminating Errors in Cryptographic
13842 Computations, J. Cryptology 14 (2001) 2, 101-119,
13843 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13847 * MIPS assembler BIGNUM division bug fix.
13851 * Disabled incorrect Alpha assembler code.
13855 * Fix PKCS#7 decode routines so they correctly update the length
13856 after reading an EOC for the EXPLICIT tag.
13860 *This change does not apply to 0.9.7.*
13862 * Fix bug in PKCS#12 key generation routines. This was triggered
13863 if a 3DES key was generated with a 0 initial byte. Include
13864 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13865 (but broken) behaviour.
13869 * Enhance bctest to search for a working bc along $PATH and print
13872 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13874 * Fix memory leaks in err.c: free err_data string if necessary;
13875 don't write to the wrong index in ERR_set_error_data.
13879 * Implement ssl23_peek (analogous to ssl23_read), which previously
13884 * Replace rdtsc with `_emit` statements for VC++ version 5.
13886 *Jeremy Cooper <jeremy@baymoo.org>*
13888 * Make it possible to reuse SSLv2 sessions.
13892 * In copy_email() check for >= 0 as a return value for
13893 X509_NAME_get_index_by_NID() since 0 is a valid index.
13895 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13897 * Avoid coredump with unsupported or invalid public keys by checking if
13898 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13899 PKCS7_verify() fails with non detached data.
13903 * Don't use getenv in library functions when run as setuid/setgid.
13904 New function OPENSSL_issetugid().
13908 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13909 due to incorrect handling of multi-threading:
13911 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13913 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13915 3. Count how many times MemCheck_off() has been called so that
13916 nested use can be treated correctly. This also avoids
13917 inband-signalling in the previous code (which relied on the
13918 assumption that thread ID 0 is impossible).
13922 * Add "-rand" option also to s_client and s_server.
13926 * Fix CPU detection on Irix 6.x.
13927 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13928 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13930 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13935 *This change does not apply to 0.9.7.*
13937 * Use the cached encoding of an X509_NAME structure rather than
13938 copying it. This is apparently the reason for the libsafe "errors"
13939 but the code is actually correct.
13943 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13944 Bleichenbacher's DSA attack.
13945 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13946 to be set and top=0 forces the highest bit to be set; top=-1 is new
13947 and leaves the highest bit random.
13949 *Ulf Moeller, Bodo Moeller*
13951 * In the `NCONF_...`-based implementations for `CONF_...` queries
13952 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13953 a temporary CONF structure with the data component set to NULL
13954 (which gives segmentation faults in lh_retrieve).
13955 Instead, use NULL for the CONF pointer in CONF_get_string and
13956 CONF_get_number (which may use environment variables) and directly
13957 return NULL from CONF_get_section.
13961 * Fix potential buffer overrun for EBCDIC.
13965 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13966 keyUsage if basicConstraints absent for a CA.
13970 * Make SMIME_write_PKCS7() write mail header values with a format that
13971 is more generally accepted (no spaces before the semicolon), since
13972 some programs can't parse those values properly otherwise. Also make
13973 sure BIO's that break lines after each write do not create invalid
13978 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13979 macros previously used would not encode an empty SEQUENCE OF
13980 and break the signature.
13984 *This change does not apply to 0.9.7.*
13986 * Zero the premaster secret after deriving the master secret in
13991 * Add some EVP_add_digest_alias registrations (as found in
13992 OpenSSL_add_all_digests()) to SSL_library_init()
13993 aka OpenSSL_add_ssl_algorithms(). This provides improved
13994 compatibility with peers using X.509 certificates
13995 with unconventional AlgorithmIdentifier OIDs.
13999 * Fix for Irix with NO_ASM.
14001 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14003 * ./config script fixes.
14005 *Ulf Moeller, Richard Levitte*
14007 * Fix 'openssl passwd -1'.
14011 * Change PKCS12_key_gen_asc() so it can cope with non null
14012 terminated strings whose length is passed in the passlen
14013 parameter, for example from PEM callbacks. This was done
14014 by adding an extra length parameter to asc2uni().
14016 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14018 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14019 call failed, free the DSA structure.
14023 * Fix to uni2asc() to cope with zero length Unicode strings.
14024 These are present in some PKCS#12 files.
14028 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14029 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14030 when writing a 32767 byte record.
14032 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14034 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14035 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14037 (RSA objects have a reference count access to which is protected
14038 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14039 so they are meant to be shared between threads.)
14040 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14041 "Reddie, Steven" <Steven.Reddie@ca.com>*
14043 * Fix a deadlock in CRYPTO_mem_leaks().
14047 * Use better test patterns in bntest.
14051 * rand_win.c fix for Borland C.
14055 * BN_rshift bugfix for n == 0.
14059 * Add a 'bctest' script that checks for some known 'bc' bugs
14060 so that 'make test' does not abort just because 'bc' is broken.
14064 * Store verify_result within SSL_SESSION also for client side to
14065 avoid potential security hole. (Re-used sessions on the client side
14066 always resulted in verify_result==X509_V_OK, not using the original
14067 result of the server certificate verification.)
14071 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14072 SSL3_RT_APPLICATION_DATA, return 0.
14073 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14078 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14079 releases, have been re-implemented by renaming the previous
14080 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14081 and ssl3_read_internal, respectively, and adding 'peek' parameters
14082 to them. The new ssl[23]_{read,peek} functions are calls to
14083 ssl[23]_read_internal with the 'peek' flag set appropriately.
14084 A 'peek' parameter has also been added to ssl3_read_bytes, which
14085 does the actual work for ssl3_read_internal.
14089 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14090 the method-specific "init()" handler. Also clean up ex_data after
14091 calling the method-specific "finish()" handler. Previously, this was
14092 happening the other way round.
14096 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14097 The previous value, 12, was not always sufficient for BN_mod_exp().
14101 * Make sure that shared libraries get the internal name engine with
14102 the full version number and not just 0. This should mark the
14103 shared libraries as not backward compatible. Of course, this should
14104 be changed again when we can guarantee backward binary compatibility.
14108 * Fix typo in get_cert_by_subject() in by_dir.c
14110 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14112 * Rework the system to generate shared libraries:
14114 - Make note of the expected extension for the shared libraries and
14115 if there is a need for symbolic links from for example libcrypto.so.0
14116 to libcrypto.so.0.9.7. There is extended info in Configure for
14119 - Make as few rebuilds of the shared libraries as possible.
14121 - Still avoid linking the OpenSSL programs with the shared libraries.
14123 - When installing, install the shared libraries separately from the
14128 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14130 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14131 and not in SSL_clear because the latter is also used by the
14132 accept/connect functions; previously, the settings made by
14133 SSL_set_read_ahead would be lost during the handshake.
14135 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14137 * Correct util/mkdef.pl to be selective about disabled algorithms.
14138 Previously, it would create entries for disabled algorithms no
14143 * Added several new manual pages for SSL_* function.
14147 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14149 * In ssl23_get_client_hello, generate an error message when faced
14150 with an initial SSL 3.0/TLS record that is too small to contain the
14151 first two bytes of the ClientHello message, i.e. client_version.
14152 (Note that this is a pathologic case that probably has never happened
14153 in real life.) The previous approach was to use the version number
14154 from the record header as a substitute; but our protocol choice
14155 should not depend on that one because it is not authenticated
14156 by the Finished messages.
14160 * More robust randomness gathering functions for Windows.
14162 *Jeffrey Altman <jaltman@columbia.edu>*
14164 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14165 not set then we don't setup the error code for issuer check errors
14166 to avoid possibly overwriting other errors which the callback does
14167 handle. If an application does set the flag then we assume it knows
14168 what it is doing and can handle the new informational codes
14173 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14174 a general "ANY" type, as such it should be able to decode anything
14175 including tagged types. However it didn't check the class so it would
14176 wrongly interpret tagged types in the same way as their universal
14177 counterpart and unknown types were just rejected. Changed so that the
14178 tagged and unknown types are handled in the same way as a SEQUENCE:
14179 that is the encoding is stored intact. There is also a new type
14180 "V_ASN1_OTHER" which is used when the class is not universal, in this
14181 case we have no idea what the actual type is so we just lump them all
14186 * On VMS, stdout may very well lead to a file that is written to
14187 in a record-oriented fashion. That means that every write() will
14188 write a separate record, which will be read separately by the
14189 programs trying to read from it. This can be very confusing.
14191 The solution is to put a BIO filter in the way that will buffer
14192 text until a linefeed is reached, and then write everything a
14193 line at a time, so every record written will be an actual line,
14194 not chunks of lines and not (usually doesn't happen, but I've
14195 seen it once) several lines in one record. BIO_f_linebuffer() is
14198 Currently, it's a VMS-only method, because that's where it has
14199 been tested well enough.
14203 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14204 it can return incorrect results.
14205 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14206 but it was in 0.9.6-beta[12].)
14210 * Disable the check for content being present when verifying detached
14211 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14212 include zero length content when signing messages.
14216 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14217 BIO_ctrl (for BIO pairs).
14221 * Add DSO method for VMS.
14225 * Bug fix: Montgomery multiplication could produce results with the
14230 * Add RPM specification openssl.spec and modify it to build three
14231 packages. The default package contains applications, application
14232 documentation and run-time libraries. The devel package contains
14233 include files, static libraries and function documentation. The
14234 doc package contains the contents of the doc directory. The original
14235 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14239 * Add a large number of documentation files for many SSL routines.
14241 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14243 * Add a configuration entry for Sony News 4.
14245 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14247 * Don't set the two most significant bits to one when generating a
14248 random number < q in the DSA library.
14252 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14253 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14254 the underlying transport is blocking) if a handshake took place.
14255 (The default behaviour is needed by applications such as s_client
14256 and s_server that use select() to determine when to use SSL_read;
14257 but for applications that know in advance when to expect data, it
14258 just makes things more complicated.)
14262 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14267 * Add a few more EBCDIC conditionals that make `req` and `x509`
14268 work better on such systems.
14270 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14272 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14273 Update PKCS12_parse() so it copies the friendlyName and the
14274 keyid to the certificates aux info.
14278 * Fix bug in PKCS7_verify() which caused an infinite loop
14279 if there was more than one signature.
14281 *Sven Uszpelkat <su@celocom.de>*
14283 * Major change in util/mkdef.pl to include extra information
14284 about each symbol, as well as presenting variables as well
14285 as functions. This change means that there's n more need
14286 to rebuild the .num files when some algorithms are excluded.
14290 * Allow the verify time to be set by an application,
14291 rather than always using the current time.
14295 * Phase 2 verify code reorganisation. The certificate
14296 verify code now looks up an issuer certificate by a
14297 number of criteria: subject name, authority key id
14298 and key usage. It also verifies self signed certificates
14299 by the same criteria. The main comparison function is
14300 X509_check_issued() which performs these checks.
14302 Lot of changes were necessary in order to support this
14303 without completely rewriting the lookup code.
14305 Authority and subject key identifier are now cached.
14307 The LHASH 'certs' is X509_STORE has now been replaced
14308 by a STACK_OF(X509_OBJECT). This is mainly because an
14309 LHASH can't store or retrieve multiple objects with
14310 the same hash value.
14312 As a result various functions (which were all internal
14313 use only) have changed to handle the new X509_STORE
14314 structure. This will break anything that messed round
14315 with X509_STORE internally.
14317 The functions X509_STORE_add_cert() now checks for an
14318 exact match, rather than just subject name.
14320 The X509_STORE API doesn't directly support the retrieval
14321 of multiple certificates matching a given criteria, however
14322 this can be worked round by performing a lookup first
14323 (which will fill the cache with candidate certificates)
14324 and then examining the cache for matches. This is probably
14325 the best we can do without throwing out X509_LOOKUP
14326 entirely (maybe later...).
14328 The X509_VERIFY_CTX structure has been enhanced considerably.
14330 All certificate lookup operations now go via a get_issuer()
14331 callback. Although this currently uses an X509_STORE it
14332 can be replaced by custom lookups. This is a simple way
14333 to bypass the X509_STORE hackery necessary to make this
14334 work and makes it possible to use more efficient techniques
14335 in future. A very simple version which uses a simple
14336 STACK for its trusted certificate store is also provided
14337 using X509_STORE_CTX_trusted_stack().
14339 The verify_cb() and verify() callbacks now have equivalents
14340 in the X509_STORE_CTX structure.
14342 X509_STORE_CTX also has a 'flags' field which can be used
14343 to customise the verify behaviour.
14347 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14348 excludes S/MIME capabilities.
14352 * When a certificate request is read in keep a copy of the
14353 original encoding of the signed data and use it when outputting
14354 again. Signatures then use the original encoding rather than
14355 a decoded, encoded version which may cause problems if the
14356 request is improperly encoded.
14360 * For consistency with other BIO_puts implementations, call
14361 buffer_write(b, ...) directly in buffer_puts instead of calling
14364 In BIO_puts, increment b->num_write as in BIO_write.
14366 *Peter.Sylvester@EdelWeb.fr*
14368 * Fix BN_mul_word for the case where the word is 0. (We have to use
14369 BN_zero, we may not return a BIGNUM with an array consisting of
14370 words set to zero.)
14374 * Avoid calling abort() from within the library when problems are
14375 detected, except if preprocessor symbols have been defined
14376 (such as REF_CHECK, BN_DEBUG etc.).
14380 * New openssl application 'rsautl'. This utility can be
14381 used for low level RSA operations. DER public key
14382 BIO/fp routines also added.
14386 * New Configure entry and patches for compiling on QNX 4.
14388 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14390 * A demo state-machine implementation was sponsored by
14391 Nuron (<http://www.nuron.com/>) and is now available in
14392 demos/state_machine.
14396 * New options added to the 'dgst' utility for signature
14397 generation and verification.
14401 * Unrecognized PKCS#7 content types are now handled via a
14402 catch all ASN1_TYPE structure. This allows unsupported
14403 types to be stored as a "blob" and an application can
14404 encode and decode it manually.
14408 * Fix various signed/unsigned issues to make a_strex.c
14409 compile under VC++.
14411 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14413 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14414 length if passed a buffer. ASN1_INTEGER_to_BN failed
14415 if passed a NULL BN and its argument was negative.
14417 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14419 * Modification to PKCS#7 encoding routines to output definite
14420 length encoding. Since currently the whole structures are in
14421 memory there's not real point in using indefinite length
14422 constructed encoding. However if OpenSSL is compiled with
14423 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14427 * Added BIO_vprintf() and BIO_vsnprintf().
14431 * Added more prefixes to parse for in the strings written
14432 through a logging bio, to cover all the levels that are available
14433 through syslog. The prefixes are now:
14435 PANIC, EMERG, EMR => LOG_EMERG
14436 ALERT, ALR => LOG_ALERT
14437 CRIT, CRI => LOG_CRIT
14438 ERROR, ERR => LOG_ERR
14439 WARNING, WARN, WAR => LOG_WARNING
14440 NOTICE, NOTE, NOT => LOG_NOTICE
14441 INFO, INF => LOG_INFO
14442 DEBUG, DBG => LOG_DEBUG
14444 and as before, if none of those prefixes are present at the
14445 beginning of the string, LOG_ERR is chosen.
14447 On Win32, the `LOG_*` levels are mapped according to this:
14449 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14450 LOG_WARNING => EVENTLOG_WARNING_TYPE
14451 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14455 * Made it possible to reconfigure with just the configuration
14456 argument "reconf" or "reconfigure". The command line arguments
14457 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14458 and are retrieved from there when reconfiguring.
14464 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14466 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14470 * The obj_dat.pl script was messing up the sorting of object
14471 names. The reason was that it compared the quoted version
14472 of strings as a result "OCSP" > "OCSP Signing" because
14473 " > SPACE. Changed script to store unquoted versions of
14474 names and add quotes on output. It was also omitting some
14475 names from the lookup table if they were given a default
14476 value (that is if SN is missing it is given the same
14477 value as LN and vice versa), these are now added on the
14478 grounds that if an object has a name we should be able to
14479 look it up. Finally added warning output when duplicate
14480 short or long names are found.
14484 * Changes needed for Tandem NSK.
14486 *Scott Uroff <scott@xypro.com>*
14488 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14489 RSA_padding_check_SSLv23(), special padding was never detected
14490 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14491 version rollback attacks was not effective.
14493 In s23_clnt.c, don't use special rollback-attack detection padding
14494 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14495 client; similarly, in s23_srvr.c, don't do the rollback check if
14496 SSL 2.0 is the only protocol enabled in the server.
14500 * Make it possible to get hexdumps of unprintable data with 'openssl
14501 asn1parse'. By implication, the functions ASN1_parse_dump() and
14502 BIO_dump_indent() are added.
14506 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14507 these print out strings and name structures based on various
14508 flags including RFC2253 support and proper handling of
14509 multibyte characters. Added options to the 'x509' utility
14510 to allow the various flags to be set.
14514 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14515 Also change the functions X509_cmp_current_time() and
14516 X509_gmtime_adj() work with an ASN1_TIME structure,
14517 this will enable certificates using GeneralizedTime in validity
14518 dates to be checked.
14522 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14523 negative public key encodings) on by default,
14524 NO_NEG_PUBKEY_BUG can be set to disable it.
14528 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14529 content octets. An i2c_ASN1_OBJECT is unnecessary because
14530 the encoding can be trivially obtained from the structure.
14534 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14535 not read locks (`CRYPTO_r_[un]lock`).
14539 * A first attempt at creating official support for shared
14540 libraries through configuration. I've kept it so the
14541 default is static libraries only, and the OpenSSL programs
14542 are always statically linked for now, but there are
14543 preparations for dynamic linking in place.
14544 This has been tested on Linux and Tru64.
14548 * Randomness polling function for Win9x, as described in:
14549 Peter Gutmann, Software Generation of Practically Strong
14554 * Fix so PRNG is seeded in req if using an already existing
14559 * New options to smime application. -inform and -outform
14560 allow alternative formats for the S/MIME message including
14561 PEM and DER. The -content option allows the content to be
14562 specified separately. This should allow things like Netscape
14563 form signing output easier to verify.
14567 * Fix the ASN1 encoding of tags using the 'long form'.
14571 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14572 STRING types. These convert content octets to and from the
14573 underlying type. The actual tag and length octets are
14574 already assumed to have been read in and checked. These
14575 are needed because all other string types have virtually
14576 identical handling apart from the tag. By having versions
14577 of the ASN1 functions that just operate on content octets
14578 IMPLICIT tagging can be handled properly. It also allows
14579 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14580 and ASN1_INTEGER are identical apart from the tag.
14584 * Change the handling of OID objects as follows:
14586 - New object identifiers are inserted in objects.txt, following
14587 the syntax given in objects.README.
14588 - objects.pl is used to process obj_mac.num and create a new
14590 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14593 This is currently kind of a hack, and the perl code in objects.pl
14594 isn't very elegant, but it works as I intended. The simplest way
14595 to check that it worked correctly is to look in obj_dat.h and
14596 check the array nid_objs and make sure the objects haven't moved
14597 around (this is important!). Additions are OK, as well as
14598 consistent name changes.
14602 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14606 * Addition of the command line parameter '-rand file' to 'openssl req'.
14607 The given file adds to whatever has already been seeded into the
14608 random pool through the RANDFILE configuration file option or
14609 environment variable, or the default random state file.
14613 * mkstack.pl now sorts each macro group into lexical order.
14614 Previously the output order depended on the order the files
14615 appeared in the directory, resulting in needless rewriting
14620 * Patches to make OpenSSL compile under Win32 again. Mostly
14621 work arounds for the VC++ problem that it treats func() as
14622 func(void). Also stripped out the parts of mkdef.pl that
14623 added extra typesafe functions: these no longer exist.
14627 * Reorganisation of the stack code. The macros are now all
14628 collected in safestack.h . Each macro is defined in terms of
14629 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14630 DEBUG_SAFESTACK is now handled in terms of function casts,
14631 this has the advantage of retaining type safety without the
14632 use of additional functions. If DEBUG_SAFESTACK is not defined
14633 then the non typesafe macros are used instead. Also modified the
14634 mkstack.pl script to handle the new form. Needs testing to see
14635 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14636 the default if no major problems. Similar behaviour for ASN1_SET_OF
14637 and PKCS12_STACK_OF.
14641 * When some versions of IIS use the 'NET' form of private key the
14642 key derivation algorithm is different. Normally MD5(password) is
14643 used as a 128 bit RC4 key. In the modified case
14644 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14645 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14646 as the old Netscape_RSA functions except they have an additional
14647 'sgckey' parameter which uses the modified algorithm. Also added
14648 an -sgckey command line option to the rsa utility. Thanks to
14649 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14650 algorithm to openssl-dev.
14654 * The evp_local.h macros were using 'c.##kname' which resulted in
14655 invalid expansion on some systems (SCO 5.0.5 for example).
14656 Corrected to 'c.kname'.
14658 *Phillip Porch <root@theporch.com>*
14660 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14661 a STACK of email addresses from a certificate or request, these look
14662 in the subject name and the subject alternative name extensions and
14663 omit any duplicate addresses.
14667 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14668 This makes DSA verification about 2 % faster.
14672 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14673 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14674 plus overhead for 1024 bit moduli).
14675 This makes exponentiations about 0.5 % faster for 1024 bit
14676 exponents (as measured by "openssl speed rsa2048").
14680 * Rename memory handling macros to avoid conflicts with other
14682 Malloc => OPENSSL_malloc
14683 Malloc_locked => OPENSSL_malloc_locked
14684 Realloc => OPENSSL_realloc
14685 Free => OPENSSL_free
14689 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14690 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14694 * CygWin32 support.
14696 *John Jarvie <jjarvie@newsguy.com>*
14698 * The type-safe stack code has been rejigged. It is now only compiled
14699 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14700 by default all type-specific stack functions are "#define"d back to
14701 standard stack functions. This results in more streamlined output
14702 but retains the type-safety checking possibilities of the original
14707 * The STACK code has been cleaned up, and certain type declarations
14708 that didn't make a lot of sense have been brought in line. This has
14709 also involved a cleanup of sorts in safestack.h to more correctly
14710 map type-safe stack functions onto their plain stack counterparts.
14711 This work has also resulted in a variety of "const"ifications of
14712 lots of the code, especially `_cmp` operations which should normally
14713 be prototyped with "const" parameters anyway.
14717 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14718 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14719 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14720 where all of 'md' is used each time the PRNG is used, but 'state'
14721 is used only indexed by a cyclic counter. As entropy may not be
14722 well distributed from the beginning, 'md' is important as a
14723 chaining variable. However, the output function chains only half
14724 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14725 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14726 in all of 'state' being rewritten, with the new values depending
14727 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14731 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14732 the handshake is continued after ssl_verify_cert_chain();
14733 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14734 can lead to 'unexplainable' connection aborts later.
14736 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14738 * Major EVP API cipher revision.
14739 Add hooks for extra EVP features. This allows various cipher
14740 parameters to be set in the EVP interface. Support added for variable
14741 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14742 setting of RC2 and RC5 parameters.
14744 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14747 Remove lots of duplicated code from the EVP library. For example *every*
14748 cipher init() function handles the 'iv' in the same way according to the
14749 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14750 for CFB and OFB modes they zero ctx->num.
14752 New functionality allows removal of S/MIME code RC2 hack.
14754 Most of the routines have the same form and so can be declared in terms
14757 By shifting this to the top level EVP_CipherInit() it can be removed from
14758 all individual ciphers. If the cipher wants to handle IVs or keys
14759 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14762 Change lots of functions like EVP_EncryptUpdate() to now return a
14763 value: although software versions of the algorithms cannot fail
14764 any installed hardware versions can.
14768 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14769 this option is set, tolerate broken clients that send the negotiated
14770 protocol version number instead of the requested protocol version
14775 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14776 i.e. non-zero for export ciphersuites, zero otherwise.
14777 Previous versions had this flag inverted, inconsistent with
14778 rsa_tmp_cb (..._TMP_RSA_CB).
14780 *Bodo Moeller; problem reported by Amit Chopra*
14782 * Add missing DSA library text string. Work around for some IIS
14783 key files with invalid SEQUENCE encoding.
14787 * Add a document (doc/standards.txt) that list all kinds of standards
14788 and so on that are implemented in OpenSSL.
14792 * Enhance c_rehash script. Old version would mishandle certificates
14793 with the same subject name hash and wouldn't handle CRLs at all.
14794 Added -fingerprint option to crl utility, to support new c_rehash
14799 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14803 * Fix for SSL server purpose checking. Server checking was
14804 rejecting certificates which had extended key usage present
14805 but no ssl client purpose.
14807 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14809 * Make PKCS#12 code work with no password. The PKCS#12 spec
14810 is a little unclear about how a blank password is handled.
14811 Since the password in encoded as a BMPString with terminating
14812 double NULL a zero length password would end up as just the
14813 double NULL. However no password at all is different and is
14814 handled differently in the PKCS#12 key generation code. NS
14815 treats a blank password as zero length. MSIE treats it as no
14816 password on export: but it will try both on import. We now do
14817 the same: PKCS12_parse() tries zero length and no password if
14818 the password is set to "" or NULL (NULL is now a valid password:
14819 it wasn't before) as does the pkcs12 application.
14823 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14824 perror when PEM_read_bio_X509_REQ fails, the error message must
14825 be obtained from the error queue.
14829 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14830 it in ERR_remove_state if appropriate, and change ERR_get_state
14831 accordingly to avoid race conditions (this is necessary because
14832 thread_hash is no longer constant once set).
14836 * Bugfix for linux-elf makefile.one.
14840 * RSA_get_default_method() will now cause a default
14841 RSA_METHOD to be chosen if one doesn't exist already.
14842 Previously this was only set during a call to RSA_new()
14843 or RSA_new_method(NULL) meaning it was possible for
14844 RSA_get_default_method() to return NULL.
14848 * Added native name translation to the existing DSO code
14849 that will convert (if the flag to do so is set) filenames
14850 that are sufficiently small and have no path information
14851 into a canonical native form. Eg. "blah" converted to
14852 "libblah.so" or "blah.dll" etc.
14856 * New function ERR_error_string_n(e, buf, len) which is like
14857 ERR_error_string(e, buf), but writes at most 'len' bytes
14858 including the 0 terminator. For ERR_error_string_n, 'buf'
14861 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14863 * CONF library reworked to become more general. A new CONF
14864 configuration file reader "class" is implemented as well as a
14865 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14866 old `CONF_*` functions are still there, but are reimplemented to
14867 work in terms of the new functions. Also, a set of functions
14868 to handle the internal storage of the configuration data is
14869 provided to make it easier to write new configuration file
14870 reader "classes" (I can definitely see something reading a
14871 configuration file in XML format, for example), called `_CONF_*`,
14872 or "the configuration storage API"...
14874 The new configuration file reading functions are:
14876 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14877 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14879 NCONF_default, NCONF_WIN32
14881 NCONF_dump_fp, NCONF_dump_bio
14883 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14884 NCONF_new creates a new CONF object. This works in the same way
14885 as other interfaces in OpenSSL, like the BIO interface.
14886 `NCONF_dump_*` dump the internal storage of the configuration file,
14887 which is useful for debugging. All other functions take the same
14888 arguments as the old `CONF_*` functions with the exception of the
14889 first that must be a `CONF *` instead of a `LHASH *`.
14891 To make it easier to use the new classes with the old `CONF_*` functions,
14892 the function CONF_set_default_method is provided.
14896 * Add '-tls1' option to 'openssl ciphers', which was already
14897 mentioned in the documentation but had not been implemented.
14898 (This option is not yet really useful because even the additional
14899 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14903 * Initial DSO code added into libcrypto for letting OpenSSL (and
14904 OpenSSL-based applications) load shared libraries and bind to
14905 them in a portable way.
14907 *Geoff Thorpe, with contributions from Richard Levitte*
14909 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
14911 * Make sure _lrotl and _lrotr are only used with MSVC.
14913 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14914 (the default implementation of RAND_status).
14916 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14917 to '-clrext' (= clear extensions), as intended and documented.
14918 *Bodo Moeller; inconsistency pointed out by Michael Attili
14919 <attili@amaxo.com>*
14921 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14922 was larger than the MD block size.
14924 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14926 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14927 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14928 using the passed key: if the passed key was a private key the result
14929 of X509_print(), for example, would be to print out all the private key
14934 * des_quad_cksum() byte order bug fix.
14935 *Ulf Möller, using the problem description in krb4-0.9.7, where
14936 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14938 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14941 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14943 * For easily testing in shell scripts whether some command
14944 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14945 returns with exit code 0 iff no command of the given name is available.
14946 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14947 the output goes to stdout and nothing is printed to stderr.
14948 Additional arguments are always ignored.
14950 Since for each cipher there is a command of the same name,
14951 the 'no-cipher' compilation switches can be tested this way.
14953 ('openssl no-XXX' is not able to detect pseudo-commands such
14954 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14958 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14962 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14963 is set; it will be thrown away anyway because each handshake creates
14965 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14966 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14967 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14968 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14972 * New s_client option -ign_eof: EOF at stdin is ignored, and
14973 'Q' and 'R' lose their special meanings (quit/renegotiate).
14974 This is part of what -quiet does; unlike -quiet, -ign_eof
14975 does not suppress any output.
14979 * Add compatibility options to the purpose and trust code. The
14980 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14981 accepts a certificate or CA, this was the previous behaviour,
14982 with all the associated security issues.
14984 X509_TRUST_COMPAT is the old trust behaviour: only and
14985 automatically trust self signed roots in certificate store. A
14986 new trust setting X509_TRUST_DEFAULT is used to specify that
14987 a purpose has no associated trust setting and it should instead
14988 use the value in the default purpose.
14992 * Fix the PKCS#8 DSA private key code so it decodes keys again
14993 and fix a memory leak.
14997 * In util/mkerr.pl (which implements 'make errors'), preserve
14998 reason strings from the previous version of the .c file, as
14999 the default to have only downcase letters (and digits) in
15000 automatically generated reasons codes is not always appropriate.
15004 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15005 using strerror. Previously, ERR_reason_error_string() returned
15006 library names as reason strings for SYSerr; but SYSerr is a special
15007 case where small numbers are errno values, not library numbers.
15011 * Add '-dsaparam' option to 'openssl dhparam' application. This
15012 converts DSA parameters into DH parameters. (When creating parameters,
15013 DSA_generate_parameters is used.)
15017 * Include 'length' (recommended exponent length) in C code generated
15018 by 'openssl dhparam -C'.
15022 * The second argument to set_label in perlasm was already being used
15023 so couldn't be used as a "file scope" flag. Moved to third argument
15028 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15029 instead of RAND_bytes for encryption IVs and salts.
15033 * Include RAND_status() into RAND_METHOD instead of implementing
15034 it only for md_rand.c Otherwise replacing the PRNG by calling
15035 RAND_set_rand_method would be impossible.
15039 * Don't let DSA_generate_key() enter an infinite loop if the random
15040 number generation fails.
15044 * New 'rand' application for creating pseudo-random output.
15048 * Added configuration support for Linux/IA64
15050 *Rolf Haberrecker <rolf@suse.de>*
15052 * Assembler module support for Mingw32.
15056 * Shared library support for HPUX (in shlib/).
15058 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15060 * Shared library support for Solaris gcc.
15062 *Lutz Behnke <behnke@trustcenter.de>*
15064 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15066 * PKCS7_encrypt() was adding text MIME headers twice because they
15067 were added manually and by SMIME_crlf_copy().
15071 * In bntest.c don't call BN_rand with zero bits argument.
15073 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15075 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15076 case was implemented. This caused BN_div_recp() to fail occasionally.
15080 * Add an optional second argument to the set_label() in the perl
15081 assembly language builder. If this argument exists and is set
15082 to 1 it signals that the assembler should use a symbol whose
15083 scope is the entire file, not just the current function. This
15084 is needed with MASM which uses the format label:: for this scope.
15086 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15088 * Change the ASN1 types so they are typedefs by default. Before
15089 almost all types were #define'd to ASN1_STRING which was causing
15090 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15095 * Change names of new functions to the new get1/get0 naming
15096 convention: After 'get1', the caller owns a reference count
15097 and has to call `..._free`; 'get0' returns a pointer to some
15098 data structure without incrementing reference counters.
15099 (Some of the existing 'get' functions increment a reference
15100 counter, some don't.)
15101 Similarly, 'set1' and 'add1' functions increase reference
15102 counters or duplicate objects.
15106 * Allow for the possibility of temp RSA key generation failure:
15107 the code used to assume it always worked and crashed on failure.
15111 * Fix potential buffer overrun problem in BIO_printf().
15112 *Ulf Möller, using public domain code by Patrick Powell; problem
15113 pointed out by David Sacerdote <das33@cornell.edu>*
15115 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15116 RAND_egd() and RAND_status(). In the command line application,
15117 the EGD socket can be specified like a seed file using RANDFILE
15122 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15123 Some CAs (e.g. Verisign) distribute certificates in this form.
15127 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15128 list to exclude them. This means that no special compilation option
15129 is needed to use anonymous DH: it just needs to be included in the
15134 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15135 EVP_MD_type. The old functionality is available in a new macro called
15136 EVP_MD_md(). Change code that uses it and update docs.
15140 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15141 where the `void *` argument is replaced by a function pointer argument.
15142 Previously `void *` was abused to point to functions, which works on
15143 many platforms, but is not correct. As these functions are usually
15144 called by macros defined in OpenSSL header files, most source code
15145 should work without changes.
15149 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15150 sections with information on -D... compiler switches used for
15151 compiling the library so that applications can see them. To enable
15152 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15153 must be defined. E.g.,
15154 #define OPENSSL_ALGORITHM_DEFINES
15155 #include <openssl/opensslconf.h>
15156 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15158 *Richard Levitte, Ulf and Bodo Möller*
15160 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15165 * Change the 'other' type in certificate aux info to a STACK_OF
15166 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15167 the required ASN1 format: arbitrary types determined by an OID.
15171 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15172 argument to 'req'. This is not because the function is newer or
15173 better than others it just uses the work 'NEW' in the certificate
15174 request header lines. Some software needs this.
15178 * Reorganise password command line arguments: now passwords can be
15179 obtained from various sources. Delete the PEM_cb function and make
15180 it the default behaviour: i.e. if the callback is NULL and the
15181 usrdata argument is not NULL interpret it as a null terminated pass
15182 phrase. If usrdata and the callback are NULL then the pass phrase
15183 is prompted for as usual.
15187 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15188 the support is automatically enabled. The resulting binaries will
15189 autodetect the card and use it if present.
15191 *Ben Laurie and Compaq Inc.*
15193 * Work around for Netscape hang bug. This sends certificate request
15194 and server done in one record. Since this is perfectly legal in the
15195 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15196 the bugs/SSLv3 entry for more info.
15200 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15204 * Add -rand argument to smime and pkcs12 applications and read/write
15209 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15213 * Add command line password options to the remaining applications.
15217 * Bug fix for BN_div_recp() for numerators with an even number of
15222 * More tests in bntest.c, and changed test_bn output.
15226 * ./config recognizes MacOS X now.
15230 * Bug fix for BN_div() when the first words of num and divisor are
15231 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15235 * Add support for various broken PKCS#8 formats, and command line
15236 options to produce them.
15240 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15241 get temporary BIGNUMs from a BN_CTX.
15245 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15250 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15251 include a #define from the old name to the new. The original intent
15252 was that statically linked binaries could for example just call
15253 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15254 link with digests. This never worked because SSLeay_add_all_digests()
15255 and SSLeay_add_all_ciphers() were in the same source file so calling
15256 one would link with the other. They are now in separate source files.
15260 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15264 * Use a less unusual form of the Miller-Rabin primality test (it used
15265 a binary algorithm for exponentiation integrated into the Miller-Rabin
15266 loop, our standard modexp algorithms are faster).
15270 * Support for the EBCDIC character set completed.
15272 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15274 * Source code cleanups: use const where appropriate, eliminate casts,
15275 use `void *` instead of `char *` in lhash.
15279 * Bugfix: ssl3_send_server_key_exchange was not restartable
15280 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15281 this the server could overwrite ephemeral keys that the client
15286 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15287 using 50 iterations of the Rabin-Miller test.
15289 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15290 iterations of the Rabin-Miller test as required by the appendix
15291 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15292 As BN_is_prime_fasttest includes trial division, DSA parameter
15293 generation becomes much faster.
15295 This implies a change for the callback functions in DSA_is_prime
15296 and DSA_generate_parameters: The callback function is called once
15297 for each positive witness in the Rabin-Miller test, not just
15298 occasionally in the inner loop; and the parameters to the
15299 callback function now provide an iteration count for the outer
15300 loop rather than for the current invocation of the inner loop.
15301 DSA_generate_parameters additionally can call the callback
15302 function with an 'iteration count' of -1, meaning that a
15303 candidate has passed the trial division test (when q is generated
15304 from an application-provided seed, trial division is skipped).
15308 * New function BN_is_prime_fasttest that optionally does trial
15309 division before starting the Rabin-Miller test and has
15310 an additional BN_CTX * argument (whereas BN_is_prime always
15311 has to allocate at least one BN_CTX).
15312 'callback(1, -1, cb_arg)' is called when a number has passed the
15313 trial division stage.
15317 * Fix for bug in CRL encoding. The validity dates weren't being handled
15322 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15326 * New function BN_pseudo_rand().
15330 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15331 bignum version of BN_from_montgomery() with the working code from
15332 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15337 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15338 made it impossible to use the same SSL_SESSION data structure in
15339 SSL2 clients in multiple threads.
15343 * The return value of RAND_load_file() no longer counts bytes obtained
15344 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15345 to seed the PRNG (previously an explicit byte count was required).
15347 *Ulf Möller, Bodo Möller*
15349 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15350 used `char *` instead of `void *` and had casts all over the place.
15354 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15358 * Retain source code compatibility for BN_prime_checks macro:
15359 BN_is_prime(..., BN_prime_checks, ...) now uses
15360 BN_prime_checks_for_size to determine the appropriate number of
15361 Rabin-Miller iterations.
15365 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15366 DH_CHECK_P_NOT_SAFE_PRIME.
15367 (Check if this is true? OpenPGP calls them "strong".)
15371 * Merge the functionality of "dh" and "gendh" programs into a new program
15372 "dhparam". The old programs are retained for now but will handle DH keys
15373 (instead of parameters) in future.
15377 * Make the ciphers, s_server and s_client programs check the return values
15378 when a new cipher list is set.
15382 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15383 ciphers. Before when the 56bit ciphers were enabled the sorting was
15386 The syntax for the cipher sorting has been extended to support sorting by
15387 cipher-strength (using the strength_bits hard coded in the tables).
15388 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15390 Fix a bug in the cipher-command parser: when supplying a cipher command
15391 string with an "undefined" symbol (neither command nor alphanumeric
15392 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15393 an error is flagged.
15395 Due to the strength-sorting extension, the code of the
15396 ssl_create_cipher_list() function was completely rearranged. I hope that
15397 the readability was also increased :-)
15399 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15401 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15402 for the first serial number and places 2 in the serial number file. This
15403 avoids problems when the root CA is created with serial number zero and
15404 the first user certificate has the same issuer name and serial number
15409 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15410 the new code. Add documentation for this stuff.
15414 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15415 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15416 structures and behave in an analogous way to the X509v3 functions:
15417 they shouldn't be called directly but wrapper functions should be used
15420 So we also now have some wrapper functions that call the X509at functions
15421 when passed certificate requests. (TO DO: similar things can be done with
15422 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15423 things. Some of these need some d2i or i2d and print functionality
15424 because they handle more complex structures.)
15428 * Add missing #ifndefs that caused missing symbols when building libssl
15429 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15430 NO_RSA in `ssl/s2*.c`.
15432 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15434 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15435 has a return value which indicates the quality of the random data
15436 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15437 error queue. New function RAND_pseudo_bytes() generates output that is
15438 guaranteed to be unique but not unpredictable. RAND_add is like
15439 RAND_seed, but takes an extra argument for an entropy estimate
15440 (RAND_seed always assumes full entropy).
15444 * Do more iterations of Rabin-Miller probable prime test (specifically,
15445 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15446 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15447 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15448 false-positive rate of at most 2^-80 for random input.
15452 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15456 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15457 in the 0.9.5 release), this returns the chain
15458 from an X509_CTX structure with a dup of the stack and all
15459 the X509 reference counts upped: so the stack will exist
15460 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15463 Also make SSL_SESSION_print() print out the verify return
15468 * Add manpage for the pkcs12 command. Also change the default
15469 behaviour so MAC iteration counts are used unless the new
15470 -nomaciter option is used. This improves file security and
15471 only older versions of MSIE (4.0 for example) need it.
15475 * Honor the no-xxx Configure options when creating .DEF files.
15479 * Add PKCS#10 attributes to field table: challengePassword,
15480 unstructuredName and unstructuredAddress. These are taken from
15481 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15482 international characters are used.
15484 More changes to X509_ATTRIBUTE code: allow the setting of types
15485 based on strings. Remove the 'loc' parameter when adding
15486 attributes because these will be a SET OF encoding which is sorted
15491 * Initial changes to the 'req' utility to allow request generation
15492 automation. This will allow an application to just generate a template
15493 file containing all the field values and have req construct the
15496 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15497 used all over the place including certificate requests and PKCS#7
15498 structures. They are currently handled manually where necessary with
15499 some primitive wrappers for PKCS#7. The new functions behave in a
15500 manner analogous to the X509 extension functions: they allow
15501 attributes to be looked up by NID and added.
15503 Later something similar to the X509V3 code would be desirable to
15504 automatically handle the encoding, decoding and printing of the
15505 more complex types. The string types like challengePassword can
15506 be handled by the string table functions.
15508 Also modified the multi byte string table handling. Now there is
15509 a 'global mask' which masks out certain types. The table itself
15510 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15511 is useful when for example there is only one permissible type
15512 (as in countryName) and using the mask might result in no valid
15517 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15518 SSL_get_peer_finished to allow applications to obtain the latest
15519 Finished messages sent to the peer or expected from the peer,
15520 respectively. (SSL_get_peer_finished is usually the Finished message
15521 actually received from the peer, otherwise the protocol will be aborted.)
15523 As the Finished message are message digests of the complete handshake
15524 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15525 be used for external authentication procedures when the authentication
15526 provided by SSL/TLS is not desired or is not enough.
15530 * Enhanced support for Alpha Linux is added. Now ./config checks if
15531 the host supports BWX extension and if Compaq C is present on the
15532 $PATH. Just exploiting of the BWX extension results in 20-30%
15533 performance kick for some algorithms, e.g. DES and RC4 to mention
15534 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15539 * Add support for MS "fast SGC". This is arguably a violation of the
15540 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15541 weak crypto and after checking the certificate is SGC a second one
15542 with strong crypto. MS SGC stops the first handshake after receiving
15543 the server certificate message and sends a second client hello. Since
15544 a server will typically do all the time consuming operations before
15545 expecting any further messages from the client (server key exchange
15546 is the most expensive) there is little difference between the two.
15548 To get OpenSSL to support MS SGC we have to permit a second client
15549 hello message after we have sent server done. In addition we have to
15550 reset the MAC if we do get this second client hello.
15554 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15555 if a DER encoded private key is RSA or DSA traditional format. Changed
15556 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15557 format DER encoded private key. Newer code should use PKCS#8 format which
15558 has the key type encoded in the ASN1 structure. Added DER private key
15559 support to pkcs8 application.
15563 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15564 ciphersuites has been selected (as required by the SSL 3/TLS 1
15565 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15566 is set, we interpret this as a request to violate the specification
15567 (the worst that can happen is a handshake failure, and 'correct'
15568 behaviour would result in a handshake failure anyway).
15572 * In SSL_CTX_add_session, take into account that there might be multiple
15573 SSL_SESSION structures with the same session ID (e.g. when two threads
15574 concurrently obtain them from an external cache).
15575 The internal cache can handle only one SSL_SESSION with a given ID,
15576 so if there's a conflict, we now throw out the old one to achieve
15581 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15582 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15583 some routines that use cipher OIDs: some ciphers do not have OIDs
15584 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15589 * Simplify the trust setting structure and code. Now we just have
15590 two sequences of OIDs for trusted and rejected settings. These will
15591 typically have values the same as the extended key usage extension
15592 and any application specific purposes.
15594 The trust checking code now has a default behaviour: it will just
15595 check for an object with the same NID as the passed id. Functions can
15596 be provided to override either the default behaviour or the behaviour
15597 for a given id. SSL client, server and email already have functions
15598 in place for compatibility: they check the NID and also return "trusted"
15599 if the certificate is self signed.
15603 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15604 traditional format into an EVP_PKEY structure.
15608 * Add a password callback function PEM_cb() which either prompts for
15609 a password if usr_data is NULL or otherwise assumes it is a null
15610 terminated password. Allow passwords to be passed on command line
15611 environment or config files in a few more utilities.
15615 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15616 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15617 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15618 Update documentation.
15622 * Support for ASN1 "NULL" type. This could be handled before by using
15623 ASN1_TYPE but there wasn't any function that would try to read a NULL
15624 and produce an error if it couldn't. For compatibility we also have
15625 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15626 don't allocate anything because they don't need to.
15630 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15633 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15635 * Rebuild of the memory allocation routines used by OpenSSL code and
15636 possibly others as well. The purpose is to make an interface that
15637 provide hooks so anyone can build a separate set of allocation and
15638 deallocation routines to be used by OpenSSL, for example memory
15639 pool implementations, or something else, which was previously hard
15640 since Malloc(), Realloc() and Free() were defined as macros having
15641 the values malloc, realloc and free, respectively (except for Win32
15642 compilations). The same is provided for memory debugging code.
15643 OpenSSL already comes with functionality to find memory leaks, but
15644 this gives people a chance to debug other memory problems.
15646 With these changes, a new set of functions and macros have appeared:
15648 CRYPTO_set_mem_debug_functions() [F]
15649 CRYPTO_get_mem_debug_functions() [F]
15650 CRYPTO_dbg_set_options() [F]
15651 CRYPTO_dbg_get_options() [F]
15652 CRYPTO_malloc_debug_init() [M]
15654 The memory debug functions are NULL by default, unless the library
15655 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15656 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15657 gives the standard debugging functions that come with OpenSSL) or
15658 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15659 provided by the library user) must be used. When the standard
15660 debugging functions are used, CRYPTO_dbg_set_options can be used to
15661 request additional information:
15662 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15663 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15665 Also, things like CRYPTO_set_mem_functions will always give the
15666 expected result (the new set of functions is used for allocation
15667 and deallocation) at all times, regardless of platform and compiler
15670 To finish it up, some functions that were never use in any other
15671 way than through macros have a new API and new semantic:
15673 CRYPTO_dbg_malloc()
15674 CRYPTO_dbg_realloc()
15677 All macros of value have retained their old syntax.
15679 *Richard Levitte and Bodo Moeller*
15681 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15682 ordering of SMIMECapabilities wasn't in "strength order" and there
15683 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15688 * Some ASN1 types with illegal zero length encoding (INTEGER,
15689 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15691 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15693 * Merge in my S/MIME library for OpenSSL. This provides a simple
15694 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15695 functionality to handle multipart/signed properly) and a utility
15696 called 'smime' to call all this stuff. This is based on code I
15697 originally wrote for Celo who have kindly allowed it to be
15698 included in OpenSSL.
15702 * Add variants des_set_key_checked and des_set_key_unchecked of
15703 des_set_key (aka des_key_sched). Global variable des_check_key
15704 decides which of these is called by des_set_key; this way
15705 des_check_key behaves as it always did, but applications and
15706 the library itself, which was buggy for des_check_key == 1,
15707 have a cleaner way to pick the version they need.
15711 * New function PKCS12_newpass() which changes the password of a
15716 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15717 dynamic mix. In both cases the ids can be used as an index into the
15718 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15719 functions so they accept a list of the field values and the
15720 application doesn't need to directly manipulate the X509_TRUST
15725 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15730 * Modify the way the V3 extension code looks up extensions. This now
15731 works in a similar way to the object code: we have some "standard"
15732 extensions in a static table which is searched with OBJ_bsearch()
15733 and the application can add dynamic ones if needed. The file
15734 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15735 updated whenever a new extension is added to the core code and kept
15736 in ext_nid order. There is a simple program 'tabtest.c' which checks
15737 this. New extensions are not added too often so this file can readily
15738 be maintained manually.
15740 There are two big advantages in doing things this way. The extensions
15741 can be looked up immediately and no longer need to be "added" using
15742 X509V3_add_standard_extensions(): this function now does nothing.
15743 Side note: I get *lots* of email saying the extension code doesn't
15744 work because people forget to call this function.
15745 Also no dynamic allocation is done unless new extensions are added:
15746 so if we don't add custom extensions there is no need to call
15747 X509V3_EXT_cleanup().
15751 * Modify enc utility's salting as follows: make salting the default. Add a
15752 magic header, so unsalted files fail gracefully instead of just decrypting
15753 to garbage. This is because not salting is a big security hole, so people
15754 should be discouraged from doing it.
15758 * Fixes and enhancements to the 'x509' utility. It allowed a message
15759 digest to be passed on the command line but it only used this
15760 parameter when signing a certificate. Modified so all relevant
15761 operations are affected by the digest parameter including the
15762 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15763 DSA key was used because it didn't fix the digest.
15767 * Initial certificate chain verify code. Currently tests the untrusted
15768 certificates for consistency with the verify purpose (which is set
15769 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15771 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15772 this is because it will reject chains with invalid extensions whereas
15773 every previous version of OpenSSL and SSLeay made no checks at all.
15775 Trust code: checks the root CA for the relevant trust settings. Trust
15776 settings have an initial value consistent with the verify purpose: e.g.
15777 if the verify purpose is for SSL client use it expects the CA to be
15778 trusted for SSL client use. However the default value can be changed to
15779 permit custom trust settings: one example of this would be to only trust
15780 certificates from a specific "secure" set of CAs.
15782 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15783 which should be used for version portability: especially since the
15784 verify structure is likely to change more often now.
15786 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15787 to set them. If not set then assume SSL clients will verify SSL servers
15790 Two new options to the verify program: -untrusted allows a set of
15791 untrusted certificates to be passed in and -purpose which sets the
15792 intended purpose of the certificate. If a purpose is set then the
15793 new chain verify code is used to check extension consistency.
15797 * Support for the authority information access extension.
15801 * Modify RSA and DSA PEM read routines to transparently handle
15802 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15803 public keys in a format compatible with certificate
15804 SubjectPublicKeyInfo structures. Unfortunately there were already
15805 functions called *_PublicKey_* which used various odd formats so
15806 these are retained for compatibility: however the DSA variants were
15807 never in a public release so they have been deleted. Changed dsa/rsa
15808 utilities to handle the new format: note no releases ever handled public
15809 keys so we should be OK.
15811 The primary motivation for this change is to avoid the same fiasco
15812 that dogs private keys: there are several incompatible private key
15813 formats some of which are standard and some OpenSSL specific and
15814 require various evil hacks to allow partial transparent handling and
15815 even then it doesn't work with DER formats. Given the option anything
15816 other than PKCS#8 should be dumped: but the other formats have to
15817 stay in the name of compatibility.
15819 With public keys and the benefit of hindsight one standard format
15820 is used which works with EVP_PKEY, RSA or DSA structures: though
15821 it clearly returns an error if you try to read the wrong kind of key.
15823 Added a -pubkey option to the 'x509' utility to output the public key.
15824 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15825 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15826 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15827 that do the same as the `EVP_PKEY_assign_*()` except they up the
15828 reference count of the added key (they don't "swallow" the
15833 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15834 CRLs would fail if the file contained no certificates or no CRLs:
15835 added a new function to read in both types and return the number
15836 read: this means that if none are read it will be an error. The
15837 DER versions of the certificate and CRL reader would always fail
15838 because it isn't possible to mix certificates and CRLs in DER format
15839 without choking one or the other routine. Changed this to just read
15840 a certificate: this is the best we can do. Also modified the code
15841 in apps/verify.c to take notice of return codes: it was previously
15842 attempting to read in certificates from NULL pointers and ignoring
15843 any errors: this is one reason why the cert and CRL reader seemed
15844 to work. It doesn't check return codes from the default certificate
15845 routines: these may well fail if the certificates aren't installed.
15849 * Code to support otherName option in GeneralName.
15853 * First update to verify code. Change the verify utility
15854 so it warns if it is passed a self signed certificate:
15855 for consistency with the normal behaviour. X509_verify
15856 has been modified to it will now verify a self signed
15857 certificate if *exactly* the same certificate appears
15858 in the store: it was previously impossible to trust a
15859 single self signed certificate. This means that:
15860 openssl verify ss.pem
15861 now gives a warning about a self signed certificate but
15862 openssl verify -CAfile ss.pem ss.pem
15867 * For servers, store verify_result in SSL_SESSION data structure
15868 (and add it to external session representation).
15869 This is needed when client certificate verifications fails,
15870 but an application-provided verification callback (set by
15871 SSL_CTX_set_cert_verify_callback) allows accepting the session
15872 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15873 but returns 1): When the session is reused, we have to set
15874 ssl->verify_result to the appropriate error code to avoid
15877 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15879 * Fix a bug in the new PKCS#7 code: it didn't consider the
15880 case in PKCS7_dataInit() where the signed PKCS7 structure
15881 didn't contain any existing data because it was being created.
15883 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15885 * Add a salt to the key derivation routines in enc.c. This
15886 forms the first 8 bytes of the encrypted file. Also add a
15887 -S option to allow a salt to be input on the command line.
15891 * New function X509_cmp(). Oddly enough there wasn't a function
15892 to compare two certificates. We do this by working out the SHA1
15893 hash and comparing that. X509_cmp() will be needed by the trust
15898 * SSL_get1_session() is like SSL_get_session(), but increments
15899 the reference count in the SSL_SESSION returned.
15901 *Geoff Thorpe <geoff@eu.c2.net>*
15903 * Fix for 'req': it was adding a null to request attributes.
15904 Also change the X509_LOOKUP and X509_INFO code to handle
15905 certificate auxiliary information.
15909 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15914 * Add the possibility to add extra information to the memory leak
15915 detecting output, to form tracebacks, showing from where each
15916 allocation was originated: CRYPTO_push_info("constant string") adds
15917 the string plus current file name and line number to a per-thread
15918 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15919 is like calling CYRPTO_pop_info() until the stack is empty.
15920 Also updated memory leak detection code to be multi-thread-safe.
15924 * Add options -text and -noout to pkcs7 utility and delete the
15925 encryption options which never did anything. Update docs.
15929 * Add options to some of the utilities to allow the pass phrase
15930 to be included on either the command line (not recommended on
15931 OSes like Unix) or read from the environment. Update the
15932 manpages and fix a few bugs.
15936 * Add a few manpages for some of the openssl commands.
15940 * Fix the -revoke option in ca. It was freeing up memory twice,
15941 leaking and not finding already revoked certificates.
15945 * Extensive changes to support certificate auxiliary information.
15946 This involves the use of X509_CERT_AUX structure and X509_AUX
15947 functions. An X509_AUX function such as PEM_read_X509_AUX()
15948 can still read in a certificate file in the usual way but it
15949 will also read in any additional "auxiliary information". By
15950 doing things this way a fair degree of compatibility can be
15951 retained: existing certificates can have this information added
15952 using the new 'x509' options.
15954 Current auxiliary information includes an "alias" and some trust
15955 settings. The trust settings will ultimately be used in enhanced
15956 certificate chain verification routines: currently a certificate
15957 can only be trusted if it is self signed and then it is trusted
15962 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
15963 The problem was that one of the replacement routines had not been working
15964 since SSLeay releases. For now the offending routine has been replaced
15965 with non-optimised assembler. Even so, this now gives around 95%
15966 performance improvement for 1024 bit RSA signs.
15970 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15971 handling. Most clients have the effective key size in bits equal to
15972 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15973 A few however don't do this and instead use the size of the decrypted key
15974 to determine the RC2 key length and the AlgorithmIdentifier to determine
15975 the effective key length. In this case the effective key length can still
15976 be 40 bits but the key length can be 168 bits for example. This is fixed
15977 by manually forcing an RC2 key into the EVP_PKEY structure because the
15978 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15979 the key length and effective key length are equal.
15983 * Add a bunch of functions that should simplify the creation of
15984 X509_NAME structures. Now you should be able to do:
15985 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15986 and have it automatically work out the correct field type and fill in
15987 the structures. The more adventurous can try:
15988 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15989 and it will (hopefully) work out the correct multibyte encoding.
15993 * Change the 'req' utility to use the new field handling and multibyte
15994 copy routines. Before the DN field creation was handled in an ad hoc
15995 way in req, ca, and x509 which was rather broken and didn't support
15996 BMPStrings or UTF8Strings. Since some software doesn't implement
15997 BMPStrings or UTF8Strings yet, they can be enabled using the config file
15998 using the dirstring_type option. See the new comment in the default
15999 openssl.cnf for more info.
16003 * Make crypto/rand/md_rand.c more robust:
16004 - Assure unique random numbers after fork().
16005 - Make sure that concurrent threads access the global counter and
16006 md serializably so that we never lose entropy in them
16007 or use exactly the same state in multiple threads.
16008 Access to the large state is not always serializable because
16009 the additional locking could be a performance killer, and
16010 md should be large enough anyway.
16014 * New file apps/app_rand.c with commonly needed functionality
16015 for handling the random seed file.
16017 Use the random seed file in some applications that previously did not:
16019 dsaparam -genkey (which also ignored its '-rand' option),
16022 x509 (when signing).
16023 Except on systems with /dev/urandom, it is crucial to have a random
16024 seed file at least for key creation, DSA signing, and for DH exchanges;
16025 for RSA signatures we could do without one.
16027 gendh and gendsa (unlike genrsa) used to read only the first byte
16028 of each file listed in the '-rand' option. The function as previously
16029 found in genrsa is now in app_rand.c and is used by all programs
16030 that support '-rand'.
16034 * In RAND_write_file, use mode 0600 for creating files;
16035 don't just chmod when it may be too late.
16039 * Report an error from X509_STORE_load_locations
16040 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16044 * New function ASN1_mbstring_copy() this copies a string in either
16045 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16046 into an ASN1_STRING type. A mask of permissible types is passed
16047 and it chooses the "minimal" type to use or an error if not type
16052 * Add function equivalents to the various macros in asn1.h. The old
16053 macros are retained with an `M_` prefix. Code inside the library can
16054 use the `M_` macros. External code (including the openssl utility)
16055 should *NOT* in order to be "shared library friendly".
16059 * Add various functions that can check a certificate's extensions
16060 to see if it usable for various purposes such as SSL client,
16061 server or S/MIME and CAs of these types. This is currently
16062 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16063 verification. Also added a -purpose flag to x509 utility to
16064 print out all the purposes.
16068 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16073 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16074 for, obtain and decode and extension and obtain its critical flag.
16075 This allows all the necessary extension code to be handled in a
16076 single function call.
16080 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16081 platforms. See crypto/rc4/rc4_enc.c for further details.
16085 * New -noout option to asn1parse. This causes no output to be produced
16086 its main use is when combined with -strparse and -out to extract data
16087 from a file (which may not be in ASN.1 format).
16091 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16092 when producing the local key id.
16094 *Richard Levitte <levitte@stacken.kth.se>*
16096 * New option -dhparam in s_server. This allows a DH parameter file to be
16097 stated explicitly. If it is not stated then it tries the first server
16098 certificate file. The previous behaviour hard coded the filename
16103 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16104 a public key to be input or output. For example:
16105 openssl rsa -in key.pem -pubout -out pubkey.pem
16106 Also added necessary DSA public key functions to handle this.
16110 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16111 in the message. This was handled by allowing
16112 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16114 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16116 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16117 to the end of the strings whereas this didn't. This would cause problems
16118 if strings read with d2i_ASN1_bytes() were later modified.
16120 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16122 * Fix for base64 decode bug. When a base64 bio reads only one line of
16123 data and it contains EOF it will end up returning an error. This is
16124 caused by input 46 bytes long. The cause is due to the way base64
16125 BIOs find the start of base64 encoded data. They do this by trying a
16126 trial decode on each line until they find one that works. When they
16127 do a flag is set and it starts again knowing it can pass all the
16128 data directly through the decoder. Unfortunately it doesn't reset
16129 the context it uses. This means that if EOF is reached an attempt
16130 is made to pass two EOFs through the context and this causes the
16131 resulting error. This can also cause other problems as well. As is
16132 usual with these problems it takes *ages* to find and the fix is
16133 trivial: move one line.
16135 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16137 * Ugly workaround to get s_client and s_server working under Windows. The
16138 old code wouldn't work because it needed to select() on sockets and the
16139 tty (for keypresses and to see if data could be written). Win32 only
16140 supports select() on sockets so we select() with a 1s timeout on the
16141 sockets and then see if any characters are waiting to be read, if none
16142 are present then we retry, we also assume we can always write data to
16143 the tty. This isn't nice because the code then blocks until we've
16144 received a complete line of data and it is effectively polling the
16145 keyboard at 1s intervals: however it's quite a bit better than not
16146 working at all :-) A dedicated Windows application might handle this
16147 with an event loop for example.
16151 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16152 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16153 will be called when RSA_sign() and RSA_verify() are used. This is useful
16154 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16155 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16156 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16157 This necessitated the support of an extra signature type NID_md5_sha1
16158 for SSL signatures and modifications to the SSL library to use it instead
16159 of calling RSA_public_decrypt() and RSA_private_encrypt().
16163 * Add new -verify -CAfile and -CApath options to the crl program, these
16164 will lookup a CRL issuers certificate and verify the signature in a
16165 similar way to the verify program. Tidy up the crl program so it
16166 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16167 less strict. It will now permit CRL extensions even if it is not
16168 a V2 CRL: this will allow it to tolerate some broken CRLs.
16172 * Initialize all non-automatic variables each time one of the openssl
16173 sub-programs is started (this is necessary as they may be started
16174 multiple times from the "OpenSSL>" prompt).
16176 *Lennart Bang, Bodo Moeller*
16178 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16179 removing all other RSA functionality (this is what NO_RSA does). This
16180 is so (for example) those in the US can disable those operations covered
16181 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16186 * Non-copying interface to BIO pairs.
16187 (still largely untested)
16191 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16192 ASCII string. This was handled independently in various places before.
16196 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16197 UTF8 strings a character at a time.
16201 * Use client_version from client hello to select the protocol
16202 (s23_srvr.c) and for RSA client key exchange verification
16203 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16207 * Add various utility functions to handle SPKACs, these were previously
16208 handled by poking round in the structure internals. Added new function
16209 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16210 print, verify and generate SPKACs. Based on an original idea from
16211 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16215 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16219 * Allow the config file extension section to be overwritten on the
16220 command line. Based on an original idea from Massimiliano Pala
16221 <madwolf@comune.modena.it>. The new option is called -extensions
16222 and can be applied to ca, req and x509. Also -reqexts to override
16223 the request extensions in req and -crlexts to override the crl extensions
16228 * Add new feature to the SPKAC handling in ca. Now you can include
16229 the same field multiple times by preceding it by "XXXX." for example:
16232 this is the same syntax as used in the req config file.
16236 * Allow certificate extensions to be added to certificate requests. These
16237 are specified in a 'req_extensions' option of the req section of the
16238 config file. They can be printed out with the -text option to req but
16239 are otherwise ignored at present.
16243 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16244 data read consists of only the final block it would not decrypted because
16245 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16246 A misplaced 'break' also meant the decrypted final block might not be
16247 copied until the next read.
16251 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16252 a few extra parameters to the DH structure: these will be useful if
16253 for example we want the value of 'q' or implement X9.42 DH.
16257 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16258 provides hooks that allow the default DSA functions or functions on a
16259 "per key" basis to be replaced. This allows hardware acceleration and
16260 hardware key storage to be handled without major modification to the
16261 library. Also added low level modexp hooks and CRYPTO_EX structure and
16262 associated functions.
16266 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16267 as "read only": it can't be written to and the buffer it points to will
16268 not be freed. Reading from a read only BIO is much more efficient than
16269 a normal memory BIO. This was added because there are several times when
16270 an area of memory needs to be read from a BIO. The previous method was
16271 to create a memory BIO and write the data to it, this results in two
16272 copies of the data and an O(n^2) reading algorithm. There is a new
16273 function BIO_new_mem_buf() which creates a read only memory BIO from
16274 an area of memory. Also modified the PKCS#7 routines to use read only
16279 * Bugfix: ssl23_get_client_hello did not work properly when called in
16280 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16281 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16282 but a retry condition occurred while trying to read the rest.
16286 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16287 NID_pkcs7_encrypted by default: this was wrong since this should almost
16288 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16289 the encrypted data type: this is a more sensible place to put it and it
16290 allows the PKCS#12 code to be tidied up that duplicated this
16295 * Changed obj_dat.pl script so it takes its input and output files on
16296 the command line. This should avoid shell escape redirection problems
16301 * Initial support for certificate extension requests, these are included
16302 in things like Xenroll certificate requests. Included functions to allow
16303 extensions to be obtained and added.
16307 * -crlf option to s_client and s_server for sending newlines as
16308 CRLF (as required by many protocols).
16312 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16314 * Install libRSAglue.a when OpenSSL is built with RSAref.
16316 *Ralf S. Engelschall*
16318 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16320 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16322 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16327 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16328 DH parameters/keys (q is lost during that conversion, but the resulting
16329 DH parameters contain its length).
16331 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16332 much faster than DH_generate_parameters (which creates parameters
16333 where `p = 2*q + 1`), and also the smaller q makes DH computations
16334 much more efficient (160-bit exponentiation instead of 1024-bit
16335 exponentiation); so this provides a convenient way to support DHE
16336 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16337 utter importance to use
16338 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16340 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16341 when such DH parameters are used, because otherwise small subgroup
16342 attacks may become possible!
16346 * Avoid memory leak in i2d_DHparams.
16350 * Allow the -k option to be used more than once in the enc program:
16351 this allows the same encrypted message to be read by multiple recipients.
16355 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16356 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16357 it will always use the numerical form of the OID, even if it has a short
16362 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16363 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16364 otherwise bn_mod_exp was called. In the case of hardware keys for example
16365 no private key components need be present and it might store extra data
16366 in the RSA structure, which cannot be accessed from bn_mod_exp.
16367 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16368 private key operations.
16372 * Added support for SPARC Linux.
16376 * pem_password_cb function type incompatibly changed from
16377 typedef int pem_password_cb(char *buf, int size, int rwflag);
16379 ....(char *buf, int size, int rwflag, void *userdata);
16380 so that applications can pass data to their callbacks:
16381 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16382 additional void * argument, which is just handed through whenever
16383 the password callback is called.
16385 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16387 New function SSL_CTX_set_default_passwd_cb_userdata.
16389 Compatibility note: As many C implementations push function arguments
16390 onto the stack in reverse order, the new library version is likely to
16391 interoperate with programs that have been compiled with the old
16392 pem_password_cb definition (PEM_whatever takes some data that
16393 happens to be on the stack as its last argument, and the callback
16394 just ignores this garbage); but there is no guarantee whatsoever that
16397 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16398 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16399 problems not only on Windows, but also on some Unix platforms.
16400 To avoid problematic command lines, these definitions are now in an
16401 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16402 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16406 * MIPS III/IV assembler module is reimplemented.
16410 * More DES library cleanups: remove references to srand/rand and
16411 delete an unused file.
16415 * Add support for the free Netwide assembler (NASM) under Win32,
16416 since not many people have MASM (ml) and it can be hard to obtain.
16417 This is currently experimental but it seems to work OK and pass all
16418 the tests. Check out INSTALL.W32 for info.
16422 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16423 without temporary keys kept an extra copy of the server key,
16424 and connections with temporary keys did not free everything in case
16429 * New function RSA_check_key and new openssl rsa option -check
16430 for verifying the consistency of RSA keys.
16432 *Ulf Moeller, Bodo Moeller*
16434 * Various changes to make Win32 compile work:
16435 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16436 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16437 comparison" warnings.
16438 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16442 * Add a debugging option to PKCS#5 v2 key generation function: when
16443 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16444 derived keys are printed to stderr.
16448 * Copy the flags in ASN1_STRING_dup().
16450 *Roman E. Pavlov <pre@mo.msk.ru>*
16452 * The x509 application mishandled signing requests containing DSA
16453 keys when the signing key was also DSA and the parameters didn't match.
16455 It was supposed to omit the parameters when they matched the signing key:
16456 the verifying software was then supposed to automatically use the CA's
16457 parameters if they were absent from the end user certificate.
16459 Omitting parameters is no longer recommended. The test was also
16460 the wrong way round! This was probably due to unusual behaviour in
16461 EVP_cmp_parameters() which returns 1 if the parameters match.
16462 This meant that parameters were omitted when they *didn't* match and
16463 the certificate was useless. Certificates signed with 'ca' didn't have
16466 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16468 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16469 The interface is as follows:
16470 Applications can use
16471 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16472 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16473 "off" is now the default.
16474 The library internally uses
16475 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16476 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16477 to disable memory-checking temporarily.
16479 Some inconsistent states that previously were possible (and were
16480 even the default) are now avoided.
16482 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16483 with each memory chunk allocated; this is occasionally more helpful
16484 than just having a counter.
16486 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16488 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16493 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16494 which largely parallels "options", but is for changing API behaviour,
16495 whereas "options" are about protocol behaviour.
16496 Initial "mode" flags are:
16498 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16499 a single record has been written.
16500 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16501 retries use the same buffer location.
16502 (But all of the contents must be
16507 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16510 * Fix problems with no-hmac etc.
16512 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16514 * New functions RSA_get_default_method(), RSA_set_method() and
16515 RSA_get_method(). These allows replacement of RSA_METHODs without having
16516 to mess around with the internals of an RSA structure.
16520 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16521 Also really enable memory leak checks in openssl.c and in some
16524 *Chad C. Mulligan, Bodo Moeller*
16526 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16527 up the length of negative integers. This has now been simplified to just
16528 store the length when it is first determined and use it later, rather
16529 than trying to keep track of where data is copied and updating it to
16531 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16533 * Add a new function PKCS7_signatureVerify. This allows the verification
16534 of a PKCS#7 signature but with the signing certificate passed to the
16535 function itself. This contrasts with PKCS7_dataVerify which assumes the
16536 certificate is present in the PKCS#7 structure. This isn't always the
16537 case: certificates can be omitted from a PKCS#7 structure and be
16538 distributed by "out of band" means (such as a certificate database).
16542 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16543 function prototypes in pem.h, also change util/mkdef.pl to add the
16544 necessary function names.
16548 * mk1mf.pl (used by Windows builds) did not properly read the
16549 options set by Configure in the top level Makefile, and Configure
16550 was not even able to write more than one option correctly.
16551 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16555 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16556 file to be loaded from a BIO or FILE pointer. The BIO version will
16557 for example allow memory BIOs to contain config info.
16561 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16562 Whoever hopes to achieve shared-library compatibility across versions
16563 must use this, not the compile-time macro.
16564 (Exercise 0.9.4: Which is the minimum library version required by
16566 Note: All this applies only to multi-threaded programs, others don't
16571 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16572 through a BIO pair triggered the default case, i.e.
16573 SSLerr(...,SSL_R_UNKNOWN_STATE).
16577 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16578 can use the SSL library even if none of the specific BIOs is
16583 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16584 for the encoded length.
16586 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16588 * Add initial documentation of the X509V3 functions.
16592 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16593 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16594 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16595 secure PKCS#8 private key format with a high iteration count.
16599 * Fix determination of Perl interpreter: A perl or perl5
16600 *directory* in $PATH was also accepted as the interpreter.
16602 *Ralf S. Engelschall*
16604 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16605 wrong with it but it was very old and did things like calling
16606 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16607 unusual formatting.
16611 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16612 to use the new extension code.
16616 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16617 with macros. This should make it easier to change their form, add extra
16618 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16623 * Add to configuration table a new entry that can specify an alternative
16624 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16625 according to Mark Crispin <MRC@Panda.COM>.
16629 * DES CBC did not update the IV. Weird.
16633 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16634 Changing the behaviour of the former might break existing programs --
16635 where IV updating is needed, des_ncbc_encrypt can be used.
16638 * When bntest is run from "make test" it drives bc to check its
16639 calculations, as well as internally checking them. If an internal check
16640 fails, it needs to cause bc to give a non-zero result or make test carries
16641 on without noticing the failure. Fixed.
16645 * DES library cleanups.
16649 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16650 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16651 ciphers. NOTE: although the key derivation function has been verified
16652 against some published test vectors it has not been extensively tested
16653 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16658 * Instead of "mkdir -p", which is not fully portable, use new
16659 Perl script "util/mkdir-p.pl".
16663 * Rewrite the way password based encryption (PBE) is handled. It used to
16664 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16665 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16666 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16667 the 'parameter' field of the AlgorithmIdentifier is passed to the
16668 underlying key generation function so it must do its own ASN1 parsing.
16669 This has also changed the EVP_PBE_CipherInit() function which now has a
16670 'parameter' argument instead of literal salt and iteration count values
16671 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16675 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16676 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16677 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16678 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16679 value was just used as a "magic string" and not used directly its
16680 value doesn't matter.
16684 * Introduce some semblance of const correctness to BN. Shame C doesn't
16689 * "linux-sparc64" configuration (ultrapenguin).
16691 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16692 "linux-sparc" configuration.
16694 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16696 * config now generates no-xxx options for missing ciphers.
16700 * Support the EBCDIC character set (work in progress).
16701 File ebcdic.c not yet included because it has a different license.
16703 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16705 * Support BS2000/OSD-POSIX.
16707 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16709 * Make callbacks for key generation use `void *` instead of `char *`.
16713 * Make S/MIME samples compile (not yet tested).
16717 * Additional typesafe stacks.
16721 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16725 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16727 * New configuration variant "sco5-gcc".
16729 * Updated some demos.
16731 *Sean O Riordain, Wade Scholine*
16733 * Add missing BIO_free at exit of pkcs12 application.
16737 * Fix memory leak in conf.c.
16741 * Updates for Win32 to assembler version of MD5.
16745 * Set #! path to perl in apps/der_chop to where we found it
16746 instead of using a fixed path.
16750 * SHA library changes for irix64-mips4-cc.
16754 * Improvements for VMS support.
16758 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16760 * Bignum library bug fix. IRIX 6 passes "make test" now!
16761 This also avoids the problems with SC4.2 and unpatched SC5.
16763 *Andy Polyakov <appro@fy.chalmers.se>*
16765 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16766 These are required because of the typesafe stack would otherwise break
16767 existing code. If old code used a structure member which used to be STACK
16768 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16769 sk_num or sk_value it would produce an error because the num, data members
16770 are not present in STACK_OF. Now it just produces a warning. sk_set
16771 replaces the old method of assigning a value to sk_value
16772 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16773 that does this will no longer work (and should use sk_set instead) but
16774 this could be regarded as a "questionable" behaviour anyway.
16778 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16779 correctly handle encrypted S/MIME data.
16783 * Change type of various DES function arguments from des_cblock
16784 (which means, in function argument declarations, pointer to char)
16785 to des_cblock * (meaning pointer to array with 8 char elements),
16786 which allows the compiler to do more typechecking; it was like
16787 that back in SSLeay, but with lots of ugly casts.
16789 Introduce new type const_des_cblock.
16793 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16794 problems: find RecipientInfo structure that matches recipient certificate
16795 and initialise the ASN1 structures properly based on passed cipher.
16799 * Belatedly make the BN tests actually check the results.
16803 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16804 to and from BNs: it was completely broken. New compilation option
16805 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16806 key elements as negative integers.
16810 * Reorganize and speed up MD5.
16812 *Andy Polyakov <appro@fy.chalmers.se>*
16816 *Richard Levitte <richard@levitte.org>*
16818 * New option -out to asn1parse to allow the parsed structure to be
16819 output to a file. This is most useful when combined with the -strparse
16820 option to examine the output of things like OCTET STRINGS.
16824 * Make SSL library a little more fool-proof by not requiring any longer
16825 that `SSL_set_{accept,connect}_state` be called before
16826 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16827 in many applications because usually everything *appeared* to work as
16828 intended anyway -- now it really works as intended).
16832 * Move openssl.cnf out of lib/.
16836 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16837 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16838 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16840 *Ralf S. Engelschall*
16842 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16843 handle PKCS#7 enveloped data properly.
16845 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16847 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16848 copying pointers. The cert_st handling is changed by this in
16849 various ways (and thus what used to be known as ctx->default_cert
16850 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16851 any longer when s->cert does not give us what we need).
16852 ssl_cert_instantiate becomes obsolete by this change.
16853 As soon as we've got the new code right (possibly it already is?),
16854 we have solved a couple of bugs of the earlier code where s->cert
16855 was used as if it could not have been shared with other SSL structures.
16857 Note that using the SSL API in certain dirty ways now will result
16858 in different behaviour than observed with earlier library versions:
16859 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16860 does not influence s as it used to.
16862 In order to clean up things more thoroughly, inside SSL_SESSION
16863 we don't use CERT any longer, but a new structure SESS_CERT
16864 that holds per-session data (if available); currently, this is
16865 the peer's certificate chain and, for clients, the server's certificate
16866 and temporary key. CERT holds only those values that can have
16867 meaningful defaults in an SSL_CTX.
16871 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16872 from the internal representation. Various PKCS#7 fixes: remove some
16873 evil casts and set the enc_dig_alg field properly based on the signing
16878 * Allow PKCS#12 password to be set from the command line or the
16879 environment. Let 'ca' get its config file name from the environment
16880 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16885 * Allow certificate policies extension to use an IA5STRING for the
16886 organization field. This is contrary to the PKIX definition but
16887 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16892 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16893 without disallowing inline assembler and the like for non-pedantic builds.
16897 * Support Borland C++ builder.
16899 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16905 * SHA-1 cleanups and performance enhancements.
16907 *Andy Polyakov <appro@fy.chalmers.se>*
16909 * Sparc v8plus assembler for the bignum library.
16911 *Andy Polyakov <appro@fy.chalmers.se>*
16913 * Accept any -xxx and +xxx compiler options in Configure.
16917 * Update HPUX configuration.
16921 * Add missing `sk_<type>_unshift()` function to safestack.h
16923 *Ralf S. Engelschall*
16925 * New function SSL_CTX_use_certificate_chain_file that sets the
16926 "extra_cert"s in addition to the certificate. (This makes sense
16927 only for "PEM" format files, as chains as a whole are not
16932 * Support verify_depth from the SSL API.
16933 x509_vfy.c had what can be considered an off-by-one-error:
16934 Its depth (which was not part of the external interface)
16935 was actually counting the number of certificates in a chain;
16936 now it really counts the depth.
16940 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16941 instead of X509err, which often resulted in confusing error
16942 messages since the error codes are not globally unique
16943 (e.g. an alleged error in ssl3_accept when a certificate
16944 didn't match the private key).
16946 * New function SSL_CTX_set_session_id_context that allows to set a default
16947 value (so that you don't need SSL_set_session_id_context for each
16948 connection using the SSL_CTX).
16952 * OAEP decoding bug fix.
16956 * Support INSTALL_PREFIX for package builders, as proposed by
16961 * New Configure options "threads" and "no-threads". For systems
16962 where the proper compiler options are known (currently Solaris
16963 and Linux), "threads" is the default.
16967 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16971 * Install various scripts to $(OPENSSLDIR)/misc, not to
16972 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16973 such as /usr/local/bin.
16977 * "make linux-shared" to build shared libraries.
16979 *Niels Poppe <niels@netbox.org>*
16981 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
16985 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16986 extension adding in x509 utility.
16990 * Remove NOPROTO sections and error code comments.
16994 * Partial rewrite of the DEF file generator to now parse the ANSI
16999 * New Configure options --prefix=DIR and --openssldir=DIR.
17003 * Complete rewrite of the error code script(s). It is all now handled
17004 by one script at the top level which handles error code gathering,
17005 header rewriting and C source file generation. It should be much better
17006 than the old method: it now uses a modified version of Ulf's parser to
17007 read the ANSI prototypes in all header files (thus the old K&R definitions
17008 aren't needed for error creation any more) and do a better job of
17009 translating function codes into names. The old 'ASN1 error code embedded
17010 in a comment' is no longer necessary and it doesn't use .err files which
17011 have now been deleted. Also the error code call doesn't have to appear all
17012 on one line (which resulted in some large lines...).
17016 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17020 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17021 0 (which usually indicates a closed connection), but continue reading.
17025 * Fix some race conditions.
17029 * Add support for CRL distribution points extension. Add Certificate
17030 Policies and CRL distribution points documentation.
17034 * Move the autogenerated header file parts to crypto/opensslconf.h.
17038 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17039 8 of keying material. Merlin has also confirmed interop with this fix
17040 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17042 *Merlin Hughes <merlin@baltimore.ie>*
17044 * Fix lots of warnings.
17046 *Richard Levitte <levitte@stacken.kth.se>*
17048 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17049 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17051 *Richard Levitte <levitte@stacken.kth.se>*
17053 * Fix problems with sizeof(long) == 8.
17055 *Andy Polyakov <appro@fy.chalmers.se>*
17057 * Change functions to ANSI C.
17061 * Fix typos in error codes.
17063 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17065 * Remove defunct assembler files from Configure.
17069 * SPARC v8 assembler BIGNUM implementation.
17071 *Andy Polyakov <appro@fy.chalmers.se>*
17073 * Support for Certificate Policies extension: both print and set.
17074 Various additions to support the r2i method this uses.
17078 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17079 return a const string when you are expecting an allocated buffer.
17083 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17084 types DirectoryString and DisplayText.
17088 * Add code to allow r2i extensions to access the configuration database,
17089 add an LHASH database driver and add several ctx helper functions.
17093 * Fix an evil bug in bn_expand2() which caused various BN functions to
17094 fail when they extended the size of a BIGNUM.
17098 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17099 support typesafe stack.
17103 * Fix typo in SSL_[gs]et_options().
17105 *Nils Frostberg <nils@medcom.se>*
17107 * Delete various functions and files that belonged to the (now obsolete)
17108 old X509V3 handling code.
17112 * New Configure option "rsaref".
17116 * Don't auto-generate pem.h.
17120 * Introduce type-safe ASN.1 SETs.
17124 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17126 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17128 * Introduce type-safe STACKs. This will almost certainly break lots of code
17129 that links with OpenSSL (well at least cause lots of warnings), but fear
17130 not: the conversion is trivial, and it eliminates loads of evil casts. A
17131 few STACKed things have been converted already. Feel free to convert more.
17132 In the fullness of time, I'll do away with the STACK type altogether.
17136 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17137 specified in `<certfile>` by updating the entry in the index.txt file.
17138 This way one no longer has to edit the index.txt file manually for
17139 revoking a certificate. The -revoke option does the gory details now.
17141 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17143 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17144 `-text` option at all and this way the `-noout -text` combination was
17145 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17147 *Ralf S. Engelschall*
17149 * Make sure a corresponding plain text error message exists for the
17150 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17151 verify callback function determined that a certificate was revoked.
17153 *Ralf S. Engelschall*
17155 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17156 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17157 all available ciphers including rc5, which was forgotten until now.
17158 In order to let the testing shell script know which algorithms
17159 are available, a new (up to now undocumented) command
17160 `openssl list-cipher-commands` is used.
17164 * Bugfix: s_client occasionally would sleep in select() when
17165 it should have checked SSL_pending() first.
17169 * New functions DSA_do_sign and DSA_do_verify to provide access to
17170 the raw DSA values prior to ASN.1 encoding.
17174 * Tweaks to Configure
17176 *Niels Poppe <niels@netbox.org>*
17178 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17183 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17187 * New config option to avoid instructions that are illegal on the 80386.
17188 The default code is faster, but requires at least a 486.
17192 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17193 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17194 same as SSL2_VERSION anyway.
17198 * New "-showcerts" option for s_client.
17202 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17203 application. Various cleanups and fixes.
17207 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17208 modify error routines to work internally. Add error codes and PBE init
17209 to library startup routines.
17213 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17214 packing functions to asn1 and evp. Changed function names and error
17215 codes along the way.
17219 * PKCS12 integration: and so it begins... First of several patches to
17220 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17221 objects to objects.h
17225 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17226 and display support for Thawte strong extranet extension.
17230 * Add LinuxPPC support.
17232 *Jeff Dubrule <igor@pobox.org>*
17234 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17235 bn_div_words in alpha.s.
17237 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17239 * Make sure the RSA OAEP test is skipped under -DRSAref because
17240 OAEP isn't supported when OpenSSL is built with RSAref.
17242 *Ulf Moeller <ulf@fitug.de>*
17244 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17245 so they no longer are missing under -DNOPROTO.
17247 *Soren S. Jorvang <soren@t.dk>*
17249 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17251 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17252 doesn't work when the session is reused. Coming soon!
17256 * Fix a security hole, that allows sessions to be reused in the wrong
17257 context thus bypassing client cert protection! All software that uses
17258 client certs and session caches in multiple contexts NEEDS PATCHING to
17259 allow session reuse! A fuller solution is in the works.
17261 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17263 * Some more source tree cleanups (removed obsolete files
17264 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17265 permission on "config" script to be executable) and a fix for the INSTALL
17268 *Ulf Moeller <ulf@fitug.de>*
17270 * Remove some legacy and erroneous uses of malloc, free instead of
17273 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17275 * Make rsa_oaep_test return non-zero on error.
17277 *Ulf Moeller <ulf@fitug.de>*
17279 * Add support for native Solaris shared libraries. Configure
17280 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17281 if someone would make that last step automatic.
17283 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17285 * ctx_size was not built with the right compiler during "make links". Fixed.
17289 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17290 except NULL ciphers". This means the default cipher list will no longer
17291 enable NULL ciphers. They need to be specifically enabled e.g. with
17292 the string "DEFAULT:eNULL".
17296 * Fix to RSA private encryption routines: if p < q then it would
17297 occasionally produce an invalid result. This will only happen with
17298 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17302 * Be less restrictive and allow also `perl util/perlpath.pl
17303 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17304 because this way one can also use an interpreter named `perl5' (which is
17305 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17306 installed as `perl').
17308 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17310 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17312 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17314 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17315 advapi32.lib to Win32 build and change the pem test comparison
17316 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17317 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17318 and crypto/des/ede_cbcm_enc.c.
17322 * DES quad checksum was broken on big-endian architectures. Fixed.
17326 * Comment out two functions in bio.h that aren't implemented. Fix up the
17327 Win32 test batch file so it (might) work again. The Win32 test batch file
17328 is horrible: I feel ill....
17332 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17333 in e_os.h. Audit of header files to check ANSI and non ANSI
17334 sections: 10 functions were absent from non ANSI section and not exported
17335 from Windows DLLs. Fixed up libeay.num for new functions.
17339 * Make `openssl version' output lines consistent.
17341 *Ralf S. Engelschall*
17343 * Fix Win32 symbol export lists for BIO functions: Added
17344 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17345 to ms/libeay{16,32}.def.
17347 *Ralf S. Engelschall*
17349 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17350 fine under Unix and passes some trivial tests I've now added. But the
17351 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17352 added to make sure no one expects that this stuff really works in the
17353 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17354 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17357 *Ralf S. Engelschall*
17359 * Fix the generation of two part addresses in perl.
17361 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17363 * Add config entry for Linux on MIPS.
17365 *John Tobey <jtobey@channel1.com>*
17367 * Make links whenever Configure is run, unless we are on Windoze.
17371 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17372 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17377 * Add a useful kludge to allow package maintainers to specify compiler and
17378 other platforms details on the command line without having to patch the
17379 Configure script every time: One now can use
17380 `perl Configure <id>:<details>`,
17381 i.e. platform ids are allowed to have details appended
17382 to them (separated by colons). This is treated as there would be a static
17383 pre-configured entry in Configure's %table under key `<id>` with value
17384 `<details>` and `perl Configure <id>` is called. So, when you want to
17385 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17386 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17387 now, which overrides the FreeBSD-elf entry on-the-fly.
17389 *Ralf S. Engelschall*
17391 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17395 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17396 on the `perl Configure ...' command line. This way one can compile
17397 OpenSSL libraries with Position Independent Code (PIC) which is needed
17398 for linking it into DSOs.
17400 *Ralf S. Engelschall*
17402 * Remarkably, export ciphers were totally broken and no-one had noticed!
17407 * Cleaned up the LICENSE document: The official contact for any license
17408 questions now is the OpenSSL core team under openssl-core@openssl.org.
17409 And add a paragraph about the dual-license situation to make sure people
17410 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17411 to the OpenSSL toolkit.
17413 *Ralf S. Engelschall*
17415 * General source tree makefile cleanups: Made `making xxx in yyy...'
17416 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17417 Additionally cleaned up the `make links' target: Remove unnecessary
17418 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17419 to speed processing and no longer clutter the display with confusing
17420 stuff. Instead only the actually done links are displayed.
17422 *Ralf S. Engelschall*
17424 * Permit null encryption ciphersuites, used for authentication only. It used
17425 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17426 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17431 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17432 signed attributes when verifying signatures (this would break them),
17433 the detached data encoding was wrong and public keys obtained using
17434 X509_get_pubkey() weren't freed.
17438 * Add text documentation for the BUFFER functions. Also added a work around
17439 to a Win95 console bug. This was triggered by the password read stuff: the
17440 last character typed gets carried over to the next fread(). If you were
17441 generating a new cert request using 'req' for example then the last
17442 character of the passphrase would be CR which would then enter the first
17447 * Added the new 'Includes OpenSSL Cryptography Software' button as
17448 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17449 button and can be used by applications based on OpenSSL to show the
17450 relationship to the OpenSSL project.
17452 *Ralf S. Engelschall*
17454 * Remove confusing variables in function signatures in files
17455 ssl/ssl_lib.c and ssl/ssl.h.
17457 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17459 * Don't install bss_file.c under PREFIX/include/
17461 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17463 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17464 functions that return function pointers and has support for NT specific
17465 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17466 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17467 unsigned to signed types: this was killing the Win32 compile.
17471 * Add new certificate file to stack functions,
17472 SSL_add_dir_cert_subjects_to_stack() and
17473 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17474 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17475 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17476 This means that Apache-SSL and similar packages don't have to mess around
17477 to add as many CAs as they want to the preferred list.
17481 * Experiment with doxygen documentation. Currently only partially applied to
17483 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17484 openssl.doxy as the configuration file.
17488 * Get rid of remaining C++-style comments which strict C compilers hate.
17490 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17492 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17493 compiled in by default: it has problems with large keys.
17497 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17498 DH private keys and/or callback functions which directly correspond to
17499 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17500 is needed for applications which have to configure certificates on a
17501 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17503 For the RSA certificate situation is makes no difference, but
17504 for the DSA certificate situation this fixes the "no shared cipher"
17505 problem where the OpenSSL cipher selection procedure failed because the
17506 temporary keys were not overtaken from the context and the API provided
17507 no way to reconfigure them.
17508 The new functions now let applications reconfigure the stuff and they
17509 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17510 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17511 non-public-API function ssl_cert_instantiate() is used as a helper
17512 function and also to reduce code redundancy inside ssl_rsa.c.
17514 *Ralf S. Engelschall*
17516 * Move s_server -dcert and -dkey options out of the undocumented feature
17517 area because they are useful for the DSA situation and should be
17518 recognized by the users.
17520 *Ralf S. Engelschall*
17522 * Fix the cipher decision scheme for export ciphers: the export bits are
17523 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17524 SSL_EXP_MASK. So, the original variable has to be used instead of the
17525 already masked variable.
17527 *Richard Levitte <levitte@stacken.kth.se>*
17529 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17531 *Richard Levitte <levitte@stacken.kth.se>*
17533 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17534 from `int` to `unsigned int` because it is a length and initialized by
17535 EVP_DigestFinal() which expects an `unsigned int *`.
17537 *Richard Levitte <levitte@stacken.kth.se>*
17539 * Don't hard-code path to Perl interpreter on shebang line of Configure
17540 script. Instead use the usual Shell->Perl transition trick.
17542 *Ralf S. Engelschall*
17544 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17545 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17546 -noout -modulus' as it's already the case for `openssl rsa -noout
17547 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17548 currently the public key is printed (a decision which was already done by
17549 `openssl dsa -modulus' in the past) which serves a similar purpose.
17550 Additionally the NO_RSA no longer completely removes the whole -modulus
17551 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17554 *Ralf S. Engelschall*
17556 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17557 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17559 *Arne Ansper <arne@ats.cyber.ee>*
17561 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17562 to be added. Now both 'req' and 'ca' can use new objects defined in the
17567 * Add cool BIO that does syslog (or event log on NT).
17569 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17571 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17572 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17573 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17574 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17578 * Add preliminary config info for new extension code.
17582 * Make RSA_NO_PADDING really use no padding.
17584 *Ulf Moeller <ulf@fitug.de>*
17586 * Generate errors when private/public key check is done.
17590 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17591 for some CRL extensions and new objects added.
17595 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17596 key usage extension and fuller support for authority key id.
17600 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17601 padding method for RSA, which is recommended for new applications in PKCS
17602 #1 v2.0 (RFC 2437, October 1998).
17603 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17604 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17605 against Bleichbacher's attack on RSA.
17606 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17609 * Updates to the new SSL compression code
17611 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17613 * Fix so that the version number in the master secret, when passed
17614 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17615 (because the server will not accept higher), that the version number
17616 is 0x03,0x01, not 0x03,0x00
17618 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17620 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17621 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17622 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17626 * Support for RAW extensions where an arbitrary extension can be
17627 created by including its DER encoding. See apps/openssl.cnf for
17632 * Make sure latest Perl versions don't interpret some generated C array
17633 code as Perl array code in the crypto/err/err_genc.pl script.
17635 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17637 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17638 not many people have the assembler. Various Win32 compilation fixes and
17639 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17640 build instructions.
17644 * Modify configure script 'Configure' to automatically create crypto/date.h
17645 file under Win32 and also build pem.h from pem.org. New script
17646 util/mkfiles.pl to create the MINFO file on environments that can't do a
17647 'make files': perl util/mkfiles.pl >MINFO should work.
17651 * Major rework of DES function declarations, in the pursuit of correctness
17652 and purity. As a result, many evil casts evaporated, and some weirdness,
17653 too. You may find this causes warnings in your code. Zapping your evil
17654 casts will probably fix them. Mostly.
17658 * Fix for a typo in asn1.h. Bug fix to object creation script
17659 obj_dat.pl. It considered a zero in an object definition to mean
17660 "end of object": none of the objects in objects.h have any zeros
17661 so it wasn't spotted.
17663 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17665 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17666 Masking (CBCM). In the absence of test vectors, the best I have been able
17667 to do is check that the decrypt undoes the encrypt, so far. Send me test
17668 vectors if you have them.
17672 * Correct calculation of key length for export ciphers (too much space was
17673 allocated for null ciphers). This has not been tested!
17677 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17678 message is now correct (it understands "crypto" and "ssl" on its
17679 command line). There is also now an "update" option. This will update
17680 the util/ssleay.num and util/libeay.num files with any new functions.
17682 perl util/mkdef.pl crypto ssl update
17683 it will update them.
17687 * Overhauled the Perl interface:
17688 - ported BN stuff to OpenSSL's different BN library
17689 - made the perl/ source tree CVS-aware
17690 - renamed the package from SSLeay to OpenSSL (the files still contain
17691 their history because I've copied them in the repository)
17692 - removed obsolete files (the test scripts will be replaced
17693 by better Test::Harness variants in the future)
17695 *Ralf S. Engelschall*
17697 * First cut for a very conservative source tree cleanup:
17698 1. merge various obsolete readme texts into doc/ssleay.txt
17699 where we collect the old documents and readme texts.
17700 2. remove the first part of files where I'm already sure that we no
17701 longer need them because of three reasons: either they are just temporary
17702 files which were left by Eric or they are preserved original files where
17703 I've verified that the diff is also available in the CVS via "cvs diff
17704 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17705 the crypto/md/ stuff).
17707 *Ralf S. Engelschall*
17709 * More extension code. Incomplete support for subject and issuer alt
17710 name, issuer and authority key id. Change the i2v function parameters
17711 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17712 what that's for :-) Fix to ASN1 macro which messed up
17713 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17717 * Preliminary support for ENUMERATED type. This is largely copied from the
17722 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17724 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17726 * Make sure `make rehash` target really finds the `openssl` program.
17728 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17730 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17731 like to hear about it if this slows down other processors.
17735 * Add CygWin32 platform information to Configure script.
17737 *Alan Batie <batie@aahz.jf.intel.com>*
17739 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17741 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17743 * New program nseq to manipulate netscape certificate sequences
17747 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17752 * Fixes to BN code. Previously the default was to define BN_RECURSION
17753 but the BN code had some problems that would cause failures when
17754 doing certificate verification and some other functions.
17756 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17758 * Add ASN1 and PEM code to support netscape certificate sequences.
17762 * Add ASN1 and PEM code to support netscape certificate sequences.
17766 * Add several PKIX and private extended key usage OIDs.
17770 * Modify the 'ca' program to handle the new extension code. Modify
17771 openssl.cnf for new extension format, add comments.
17775 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17776 and add a sample to openssl.cnf so req -x509 now adds appropriate
17781 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17782 error code, add initial support to X509_print() and x509 application.
17786 * Takes a deep breath and start adding X509 V3 extension support code. Add
17787 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17788 stuff is currently isolated and isn't even compiled yet.
17792 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17793 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17794 Removed the versions check from X509 routines when loading extensions:
17795 this allows certain broken certificates that don't set the version
17796 properly to be processed.
17800 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17801 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17802 can still be regenerated with "make depend".
17806 * Spelling mistake in C version of CAST-128.
17808 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17810 * Changes to the error generation code. The perl script err-code.pl
17811 now reads in the old error codes and retains the old numbers, only
17812 adding new ones if necessary. It also only changes the .err files if new
17813 codes are added. The makefiles have been modified to only insert errors
17814 when needed (to avoid needlessly modifying header files). This is done
17815 by only inserting errors if the .err file is newer than the auto generated
17816 C file. To rebuild all the error codes from scratch (the old behaviour)
17817 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17818 or delete all the .err files.
17822 * CAST-128 was incorrectly implemented for short keys. The C version has
17823 been fixed, but is untested. The assembler versions are also fixed, but
17824 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17825 to regenerate it if needed.
17826 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17827 Hagino <itojun@kame.net>*
17829 * File was opened incorrectly in randfile.c.
17831 *Ulf Möller <ulf@fitug.de>*
17833 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17834 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17835 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17836 al: it's just almost always a UTCTime. Note this patch adds new error
17837 codes so do a "make errors" if there are problems.
17841 * Correct Linux 1 recognition in config.
17843 *Ulf Möller <ulf@fitug.de>*
17845 * Remove pointless MD5 hash when using DSA keys in ca.
17847 *Anonymous <nobody@replay.com>*
17849 * Generate an error if given an empty string as a cert directory. Also
17850 generate an error if handed NULL (previously returned 0 to indicate an
17851 error, but didn't set one).
17853 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17855 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17859 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17860 parameters. This was causing a warning which killed off the Win32 compile.
17864 * Remove C++ style comments from crypto/bn/bn_local.h.
17866 *Neil Costigan <neil.costigan@celocom.com>*
17868 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17869 based on a text string, looking up short and long names and finally
17870 "dot" format. The "dot" format stuff didn't work. Added new function
17871 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17872 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17873 OID is not part of the table.
17877 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17878 X509_LOOKUP_by_alias().
17882 * Sort openssl functions by name.
17886 * Get the gendsa program working (hopefully) and add it to app list. Remove
17887 encryption from sample DSA keys (in case anyone is interested the password
17892 * Make *all* `*_free` functions accept a NULL pointer.
17894 *Frans Heymans <fheymans@isaserver.be>*
17896 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17899 *Anonymous <nobody@replay.com>*
17901 * s_server should send the CAfile as acceptable CAs, not its own cert.
17903 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17905 * Don't blow it for numeric -newkey arguments to apps/req.
17907 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17909 * Temp key "for export" tests were wrong in s3_srvr.c.
17911 *Anonymous <nobody@replay.com>*
17913 * Add prototype for temp key callback functions
17914 SSL_CTX_set_tmp_{rsa,dh}_callback().
17918 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17919 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17923 * X509_name_add_entry() freed the wrong thing after an error.
17925 *Arne Ansper <arne@ats.cyber.ee>*
17927 * rsa_eay.c would attempt to free a NULL context.
17929 *Arne Ansper <arne@ats.cyber.ee>*
17931 * BIO_s_socket() had a broken should_retry() on Windoze.
17933 *Arne Ansper <arne@ats.cyber.ee>*
17935 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17937 *Arne Ansper <arne@ats.cyber.ee>*
17939 * Make sure the already existing X509_STORE->depth variable is initialized
17940 in X509_STORE_new(), but document the fact that this variable is still
17941 unused in the certificate verification process.
17943 *Ralf S. Engelschall*
17945 * Fix the various library and apps files to free up pkeys obtained from
17946 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17950 * Fix reference counting in X509_PUBKEY_get(). This makes
17951 demos/maurice/example2.c work, amongst others, probably.
17953 *Steve Henson and Ben Laurie*
17955 * First cut of a cleanup for apps/. First the `ssleay` program is now named
17956 `openssl` and second, the shortcut symlinks for the `openssl <command>`
17957 are no longer created. This way we have a single and consistent command
17958 line interface `openssl <command>`, similar to `cvs <command>`.
17960 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17962 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17963 BIT STRING wrapper always have zero unused bits.
17967 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17971 * Make the top-level INSTALL documentation easier to understand.
17975 * Makefiles updated to exit if an error occurs in a sub-directory
17976 make (including if user presses ^C) [Paul Sutton]
17978 * Make Montgomery context stuff explicit in RSA data structure.
17982 * Fix build order of pem and err to allow for generated pem.h.
17986 * Fix renumbering bug in X509_NAME_delete_entry().
17990 * Enhanced the err-ins.pl script so it makes the error library number
17991 global and can add a library name. This is needed for external ASN1 and
17992 other error libraries.
17996 * Fixed sk_insert which never worked properly.
18000 * Fix ASN1 macros so they can handle indefinite length constructed
18001 EXPLICIT tags. Some non standard certificates use these: they can now
18006 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18007 into a single doc/ssleay.txt bundle. This way the information is still
18008 preserved but no longer messes up this directory. Now it's new room for
18009 the new set of documentation files.
18011 *Ralf S. Engelschall*
18013 * SETs were incorrectly DER encoded. This was a major pain, because they
18014 shared code with SEQUENCEs, which aren't coded the same. This means that
18015 almost everything to do with SETs or SEQUENCEs has either changed name or
18016 number of arguments.
18018 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18020 * Fix test data to work with the above.
18024 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18025 was already fixed by Eric for 0.9.1 it seems.
18027 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18029 * Autodetect FreeBSD3.
18033 * Fix various bugs in Configure. This affects the following platforms:
18037 unixware-2.0-pentium
18042 * Eliminate generated files from CVS. Reorder tests to regenerate files
18043 before they are needed.
18047 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18051 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18053 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18054 changed SSLeay to OpenSSL in version strings.
18056 *Ralf S. Engelschall*
18058 * Some fixups to the top-level documents.
18062 * Fixed the nasty bug where rsaref.h was not found under compile-time
18063 because the symlink to include/ was missing.
18065 *Ralf S. Engelschall*
18067 * Incorporated the popular no-RSA/DSA-only patches
18068 which allow to compile a RSA-free SSLeay.
18070 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18072 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18073 when "ssleay" is still not found.
18075 *Ralf S. Engelschall*
18077 * Added more platforms to Configure: Cray T3E, HPUX 11,
18079 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18081 * Updated the README file.
18083 *Ralf S. Engelschall*
18085 * Added various .cvsignore files in the CVS repository subdirs
18086 to make a "cvs update" really silent.
18088 *Ralf S. Engelschall*
18090 * Recompiled the error-definition header files and added
18091 missing symbols to the Win32 linker tables.
18093 *Ralf S. Engelschall*
18095 * Cleaned up the top-level documents;
18096 o new files: CHANGES and LICENSE
18097 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18098 o merged COPYRIGHT into LICENSE
18099 o removed obsolete TODO file
18100 o renamed MICROSOFT to INSTALL.W32
18102 *Ralf S. Engelschall*
18104 * Removed dummy files from the 0.9.1b source tree:
18105 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18106 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18107 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18108 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18109 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18111 *Ralf S. Engelschall*
18113 * Added various platform portability fixes.
18117 * The Genesis of the OpenSSL rpject:
18118 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18119 Young and Tim J. Hudson created while they were working for C2Net until
18122 *The OpenSSL Project*
18124 ### Changes between 0.9.0b and 0.9.1b [not released]
18126 * Updated a few CA certificates under certs/
18130 * Changed some BIGNUM api stuff.
18134 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18135 DGUX x86, Linux Alpha, etc.
18139 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18140 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18145 * Add -strparse option to asn1pars program which parses nested
18148 *Dr Stephen Henson <shenson@bigfoot.com>*
18150 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18154 * DSA fix for "ca" program.
18158 * Added "-genkey" option to "dsaparam" program.
18162 * Added RIPE MD160 (rmd160) message digest.
18166 * Added -a (all) option to "ssleay version" command.
18170 * Added PLATFORM define which is the id given to Configure.
18174 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18178 * Extended the ASN.1 parser routines.
18182 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18186 * Added a BN_CTX to the BN library.
18190 * Fixed the weak key values in DES library
18194 * Changed API in EVP library for cipher aliases.
18198 * Added support for RC2/64bit cipher.
18202 * Converted the lhash library to the crypto/mem.c functions.
18206 * Added more recognized ASN.1 object ids.
18210 * Added more RSA padding checks for SSL/TLS.
18214 * Added BIO proxy/filter functionality.
18218 * Added extra_certs to SSL_CTX which can be used
18219 send extra CA certificates to the client in the CA cert chain sending
18220 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18224 * Now Fortezza is denied in the authentication phase because
18225 this is key exchange mechanism is not supported by SSLeay at all.
18229 * Additional PKCS1 checks.
18233 * Support the string "TLSv1" for all TLS v1 ciphers.
18237 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18238 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18242 * Fixed a few memory leaks.
18246 * Fixed various code and comment typos.
18250 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18251 bytes sent in the client random.
18253 *Edward Bishop <ebishop@spyglass.com>*
18257 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18258 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18259 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18260 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18261 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18262 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18263 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18264 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18265 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18266 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18267 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18268 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18269 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18270 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18271 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18272 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18273 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18274 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18275 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18276 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18277 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18278 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18279 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18280 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18281 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18282 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18283 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18284 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18285 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18286 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18287 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18288 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18289 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18290 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18291 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18292 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18293 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18294 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18295 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18296 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18297 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18298 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18299 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18300 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18301 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18302 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18303 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18304 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18305 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18306 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18307 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18308 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18309 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18310 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18311 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18312 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18313 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18314 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18315 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18316 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18317 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18318 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18319 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18320 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18321 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18322 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18323 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18324 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18325 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18326 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18327 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18328 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18329 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18330 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18331 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18332 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18333 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18334 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18335 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18336 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18337 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18338 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18339 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18340 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18341 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18342 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18343 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18344 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18345 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18346 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18347 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18348 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18349 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18350 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18351 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18352 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18353 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18354 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18355 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18356 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18357 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18358 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18359 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18360 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18361 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18362 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18363 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18364 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18365 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18366 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18367 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18368 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18369 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18370 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18371 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18372 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18373 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18374 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18375 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18376 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18377 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18378 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18379 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18380 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18381 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18382 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18383 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18384 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18385 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18386 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18387 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18388 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18389 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18390 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18391 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18392 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18393 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18394 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18395 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18396 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18397 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18398 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18399 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18400 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18401 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18402 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18403 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18404 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18405 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18406 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18407 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18408 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18409 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18410 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18411 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18412 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18413 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18414 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18415 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18416 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655