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 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
46 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
47 used and applications should instead use the
48 L<EC_POINT_set_affine_coordinates(3)> and
49 L<EC_POINT_get_affine_coordinates(3)> functions.
53 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
54 arrays to be more easily constructed via a series of utility functions.
55 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
56 the various push functions and finally convert to a passable OSSL_PARAM
57 array using OSSL_PARAM_BLD_to_param().
61 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
62 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
63 internal keys, if they correspond to one of those built in types.
67 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
68 contain a provider side internal key.
72 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
73 They are old functions that we don't use, and that you could disable with
74 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
78 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
79 have been converted to Markdown with the goal to produce documents
80 which not only look pretty when viewed online in the browser, but
81 remain well readable inside a plain text editor.
83 To achieve this goal, a 'minimalistic' Markdown style has been applied
84 which avoids formatting elements that interfere too much with the
85 reading flow in the text file. For example, it
87 * avoids [ATX headings][] and uses [setext headings][] instead
88 (which works for `<h1>` and `<h2>` headings only).
89 * avoids [inline links][] and uses [reference links][] instead.
90 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
92 [ATX headings]: https://github.github.com/gfm/#atx-headings
93 [setext headings]: https://github.github.com/gfm/#setext-headings
94 [inline links]: https://github.github.com/gfm/#inline-link
95 [reference links]: https://github.github.com/gfm/#reference-link
96 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
97 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
101 * The test suite is changed to preserve results of each test recipe.
102 A new directory test-runs/ with subdirectories named like the
103 test recipes are created in the build tree for this purpose.
107 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
108 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
109 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
111 *David von Oheimb, Martin Peylo*
113 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
114 The legacy OCSP-focused and only partly documented API is retained.
115 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
119 * All of the low level RSA functions have been deprecated including:
121 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
122 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
123 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
124 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
125 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
126 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
127 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
128 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
129 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
130 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
131 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
132 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
133 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
134 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
135 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
136 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
137 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
138 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
139 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
140 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
141 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
142 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
143 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
144 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
145 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
146 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
147 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
148 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
150 Use of these low level functions has been informally discouraged for a long
151 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
152 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
153 L<EVP_PKEY_decrypt(3)>.
157 * X509 certificates signed using SHA1 are no longer allowed at security
159 In TLS/SSL the default security level is 1. It can be set either
160 using the cipher string with @SECLEVEL, or calling
161 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
162 a call to SSL_CTX_use_certificate() will fail if the security level is not
164 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
165 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
170 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
171 modified to use PKEY APIs. These commands are now in maintenance mode
172 and no new features will be added to them.
176 * The command line utility rsautl has been deprecated.
177 Instead use the pkeyutl program.
181 * The command line utilities genrsa and rsa have been modified to use PKEY
182 APIs These commands are now in maintenance mode and no new features will
187 * All of the low level DH functions have been deprecated including:
189 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
190 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
191 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
192 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
193 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
194 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
195 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
196 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
197 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
198 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
199 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
200 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
201 DH_meth_get_generate_params and DH_meth_set_generate_params.
203 Use of these low level functions has been informally discouraged for a long
204 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
205 and L<EVP_PKEY_derive(3)>.
209 * All of the low level DSA functions have been deprecated including:
211 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
212 DSA_get_default_method, DSA_set_method, DSA_get_method,
213 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
214 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
215 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
216 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
217 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
218 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
219 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
220 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
221 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
222 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
223 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
226 Use of these low level functions has been informally discouraged for a long
227 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
228 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
232 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
233 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
234 This means that applications don't have to look at the curve NID and
235 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
236 However, they still can, that EVP_PKEY_set_alias_type() call acts as
237 a no-op when the EVP_PKEY is already of the given type.
239 Parameter and key generation is also reworked to make it possible
240 to generate EVP_PKEY_SM2 parameters and keys without having to go
241 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
242 However, code that does the latter will still work as before.
246 * Deprecated low level ECDH and ECDSA functions. These include:
248 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
249 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
252 Use of these low level functions has been informally discouraged for a long
253 time. Instead applications should use the EVP_PKEY_derive(3),
254 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
258 * Deprecated the EC_KEY_METHOD functions. These include:
260 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
261 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
262 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
263 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
264 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
265 EC_KEY_METHOD_get_verify.
267 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
271 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
272 and EVP_PKEY_decrypt() instead.
273 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
274 and EVP_PKEY_encrypt() instead.
278 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
279 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
280 a new formulation to include all the things it can be used for,
281 as well as words of caution.
285 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
286 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
290 * All of the low level HMAC functions have been deprecated including:
292 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
293 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
296 Use of these low level functions has been informally discouraged for a long
297 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
298 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
299 and L<EVP_MAC_final(3)>.
303 * Over two thousand fixes were made to the documentation, including:
304 - Common options (such as -rand/-writerand, TLS version control, etc)
305 were refactored and point to newly-enhanced descriptions in openssl.pod.
306 - Added style conformance for all options (with help from Richard Levitte),
307 documented all reported missing options, added a CI build to check
308 that all options are documented and that no unimplemented options
310 - Documented some internals, such as all use of environment variables.
311 - Addressed all internal broken L<> references.
315 * All of the low level CMAC functions have been deprecated including:
317 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
318 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
320 Use of these low level functions has been informally discouraged for a long
321 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
322 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
323 and L<EVP_MAC_final(3)>.
327 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
328 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
331 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
332 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
333 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
334 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
335 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
336 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
337 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
338 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
339 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
340 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
342 Use of these low level functions has been informally discouraged
343 for a long time. Applications should use the EVP_DigestInit_ex(3),
344 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
348 * Corrected the documentation of the return values from the `EVP_DigestSign*`
349 set of functions. The documentation mentioned negative values for some
350 errors, but this was never the case, so the mention of negative values
353 Code that followed the documentation and thereby check with something
354 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
358 * All of the low level cipher functions have been deprecated including:
360 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
361 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
362 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
363 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
364 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
365 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
366 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
367 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
368 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
369 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
370 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
371 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
372 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
373 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
374 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
375 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
376 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
377 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
378 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
379 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
380 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
381 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
382 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
383 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
384 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
385 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
386 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
387 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
388 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
390 Use of these low level functions has been informally discouraged for
391 a long time. Applications should use the high level EVP APIs, e.g.
392 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
393 equivalently named decrypt functions instead.
395 *Matt Caswell and Paul Dale*
397 * Removed include/openssl/opensslconf.h.in and replaced it with
398 include/openssl/configuration.h.in, which differs in not including
399 <openssl/macros.h>. A short header include/openssl/opensslconf.h
400 was added to include both.
402 This allows internal hacks where one might need to modify the set
403 of configured macros, for example this if deprecated symbols are
404 still supposed to be available internally:
406 #include <openssl/configuration.h>
408 #undef OPENSSL_NO_DEPRECATED
409 #define OPENSSL_SUPPRESS_DEPRECATED
411 #include <openssl/macros.h>
413 This should not be used by applications that use the exported
414 symbols, as that will lead to linking errors.
418 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
419 used in exponentiation with 512-bit moduli. No EC algorithms are
420 affected. Analysis suggests that attacks against 2-prime RSA1024,
421 3-prime RSA1536, and DSA1024 as a result of this defect would be very
422 difficult to perform and are not believed likely. Attacks against DH512
423 are considered just feasible. However, for an attack the target would
424 have to re-use the DH512 private key, which is not recommended anyway.
425 Also applications directly using the low level API BN_mod_exp may be
426 affected if they use BN_FLG_CONSTTIME.
431 * Most memory-debug features have been deprecated, and the functionality
432 replaced with no-ops.
436 * Added documentation for the STACK API. OpenSSL only defines the STACK
437 functions where they are used.
441 * Introduced a new method type and API, OSSL_SERIALIZER, to
442 represent generic serializers. An implementation is expected to
443 be able to serialize an object associated with a given name (such
444 as an algorithm name for an asymmetric key) into forms given by
445 implementation properties.
447 Serializers are primarily used from inside libcrypto, through
448 calls to functions like EVP_PKEY_print_private(),
449 PEM_write_bio_PrivateKey() and similar.
451 Serializers are specified in such a way that they can be made to
452 directly handle the provider side portion of an object, if this
453 provider side part comes from the same provider as the serializer
454 itself, but can also be made to handle objects in parametrized
455 form (as an OSSL_PARAM array of data). This allows a provider to
456 offer generic serializers as a service for any other provider.
460 * Added a .pragma directive to the syntax of configuration files, to
461 allow varying behavior in a supported and predictable manner.
462 Currently added pragma:
466 This allows dollar signs to be a keyword character unless it's
467 followed by a opening brace or parenthesis. This is useful for
468 platforms where dollar signs are commonly used in names, such as
469 volume names and system directory names on VMS.
473 * Added functionality to create an EVP_PKEY from user data. This
474 is effectively the same as creating a RSA, DH or DSA object and
475 then assigning them to an EVP_PKEY, but directly using algorithm
476 agnostic EVP functions. A benefit is that this should be future
477 proof for public key algorithms to come.
481 * Change the interpretation of the '--api' configuration option to
482 mean that this is a desired API compatibility level with no
483 further meaning. The previous interpretation, that this would
484 also mean to remove all deprecated symbols up to and including
485 the given version, no requires that 'no-deprecated' is also used
486 in the configuration.
488 When building applications, the desired API compatibility level
489 can be set with the OPENSSL_API_COMPAT macro like before. For
490 API compatibility version below 3.0, the old style numerical
491 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
492 For version 3.0 and on, the value is expected to be the decimal
493 value calculated from the major and minor version like this:
495 MAJOR * 10000 + MINOR * 100
499 -DOPENSSL_API_COMPAT=30000 For 3.0
500 -DOPENSSL_API_COMPAT=30200 For 3.2
502 To hide declarations that are deprecated up to and including the
503 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
504 given when building the application as well.
508 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
509 access to certificate and CRL stores via URIs and OSSL_STORE
512 This adds the following functions:
514 - X509_LOOKUP_store()
515 - X509_STORE_load_file()
516 - X509_STORE_load_path()
517 - X509_STORE_load_store()
518 - SSL_add_store_cert_subjects_to_stack()
519 - SSL_CTX_set_default_verify_store()
520 - SSL_CTX_load_verify_file()
521 - SSL_CTX_load_verify_dir()
522 - SSL_CTX_load_verify_store()
524 Also, the following functions are now deprecated:
526 - X509_STORE_load_locations() (use X509_STORE_load_file(),
527 X509_STORE_load_path() or X509_STORE_load_store() instead)
528 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
529 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
533 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
534 The presence of this system service is determined at run-time.
538 * Added functionality to create an EVP_PKEY context based on data
539 for methods from providers. This takes an algorithm name and a
540 property query string and simply stores them, with the intent
541 that any operation that uses this context will use those strings
542 to fetch the needed methods implicitly, thereby making the port
543 of application written for pre-3.0 OpenSSL easier.
547 * The undocumented function NCONF_WIN32() has been deprecated; for
548 conversion details see the HISTORY section of doc/man5/config.pod
552 * Introduced the new functions EVP_DigestSignInit_ex() and
553 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
554 EVP_DigestVerifyUpdate() have been converted to functions. See the man
555 pages for further details.
559 * Over two thousand fixes were made to the documentation, including:
560 adding missing command flags, better style conformance, documentation
563 *Rich Salz, Richard Levitte*
565 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
566 X25519, X448, Ed25519 and Ed448.
570 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
575 * Deprecated the public definition of ERR_STATE as well as the function
576 ERR_get_state(). This is done in preparation of making ERR_STATE an
581 * Added ERR functionality to give callers access to the stored function
582 names that have replaced the older function code based functions.
584 New functions are ERR_get_error_func(), ERR_peek_error_func(),
585 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
586 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
587 and ERR_peek_last_error_all().
589 These functions have become deprecated: ERR_get_error_line_data(),
590 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
591 ERR_func_error_string().
595 * Extended testing to be verbose for failing tests only. The make variables
596 VERBOSE_FAILURE or VF can be used to enable this:
598 $ make VF=1 test # Unix
599 $ mms /macro=(VF=1) test ! OpenVMS
600 $ nmake VF=1 test # Windows
604 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
605 used even when parsing explicit parameters, when loading a serialized key
606 or calling `EC_GROUP_new_from_ecpkparameters()`/
607 `EC_GROUP_new_from_ecparameters()`.
608 This prevents bypass of security hardening and performance gains,
609 especially for curves with specialized EC_METHODs.
610 By default, if a key encoded with explicit parameters is loaded and later
611 serialized, the output is still encoded with explicit parameters, even if
612 internally a "named" EC_GROUP is used for computation.
616 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
617 this change, EC_GROUP_set_generator would accept order and/or cofactor as
618 NULL. After this change, only the cofactor parameter can be NULL. It also
619 does some minimal sanity checks on the passed order.
624 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
625 An attack is simple, if the first CMS_recipientInfo is valid but the
626 second CMS_recipientInfo is chosen ciphertext. If the second
627 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
628 encryption key will be replaced by garbage, and the message cannot be
629 decoded, but if the RSA decryption fails, the correct encryption key is
630 used and the recipient will not notice the attack.
631 As a work around for this potential attack the length of the decrypted
632 key must be equal to the cipher default key length, in case the
633 certifiate is not given and all recipientInfo are tried out.
634 The old behaviour can be re-enabled in the CMS code by setting the
635 CMS_DEBUG_DECRYPT flag.
639 * Early start up entropy quality from the DEVRANDOM seed source has been
640 improved for older Linux systems. The RAND subsystem will wait for
641 /dev/random to be producing output before seeding from /dev/urandom.
642 The seeded state is stored for future library initialisations using
643 a system global shared memory segment. The shared memory identifier
644 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
645 the desired value. The default identifier is 114.
649 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
650 when primes for RSA keys are computed.
651 Since we previously always generated primes == 2 (mod 3) for RSA keys,
652 the 2-prime and 3-prime RSA modules were easy to distinguish, since
653 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
654 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
655 This avoids possible fingerprinting of newly generated RSA modules.
659 * Correct the extended master secret constant on EBCDIC systems. Without this
660 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
661 negotiate EMS will fail. Unfortunately this also means that TLS connections
662 between EBCDIC systems with this fix, and EBCDIC systems without this
663 fix will fail if they negotiate EMS.
667 * Changed the library initialisation so that the config file is now loaded
668 by default. This was already the case for libssl. It now occurs for both
669 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
670 OPENSSL_init_crypto() to suppress automatic loading of a config file.
674 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
675 where the former acts as a replacement for ERR_put_error(), and the
676 latter replaces the combination ERR_put_error()+ERR_add_error_data().
677 ERR_raise_data() adds more flexibility by taking a format string and
678 an arbitrary number of arguments following it, to be processed with
683 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
684 to check if a named provider is loaded and available. When called, it
685 will also activate all fallback providers if such are still present.
689 * Enforce a minimum DH modulus size of 512 bits.
693 * Changed DH parameters to generate the order q subgroup instead of 2q.
694 Previously generated DH parameters are still accepted by DH_check
695 but DH_generate_key works around that by clearing bit 0 of the
696 private key for those. This avoids leaking bit 0 of the private key.
700 * Significantly reduce secure memory usage by the randomness pools.
704 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
709 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
710 algorithms. An implementation of a key exchange algorithm can be obtained
711 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
712 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
713 the older EVP_PKEY_derive_init() function. See the man pages for the new
714 functions for further details.
718 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
722 * Removed the function names from error messages and deprecated the
725 * Removed NextStep support and the macro OPENSSL_UNISTD
729 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
730 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
731 Also removed "export var as function" capability; we do not export
732 variables, only functions.
736 * RC5_32_set_key has been changed to return an int type, with 0 indicating
737 an error and 1 indicating success. In previous versions of OpenSSL this
738 was a void type. If a key was set longer than the maximum possible this
743 * Support SM2 signing and verification schemes with X509 certificate.
747 * Use SHA256 as the default digest for TS query in the ts app.
751 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
752 This checks that the salt length is at least 128 bits, the derived key
753 length is at least 112 bits, and that the iteration count is at least 1000.
754 For backwards compatibility these checks are disabled by default in the
755 default provider, but are enabled by default in the fips provider.
756 To enable or disable these checks use the control
757 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
761 * Default cipher lists/suites are now available via a function, the
762 #defines are deprecated.
766 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
767 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
768 for Windows Store apps easier. Also, the "no-uplink" option has been added.
772 * Join the directories crypto/x509 and crypto/x509v3
776 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
777 This changes the size when using the genpkey app when no size is given. It
778 fixes an omission in earlier changes that changed all RSA, DSA and DH
779 generation apps to use 2048 bits by default.
783 * Added command 'openssl kdf' that uses the EVP_KDF API.
787 * Added command 'openssl mac' that uses the EVP_MAC API.
791 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
792 as default directories. Also added the command 'openssl info'
793 for scripting purposes.
797 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
798 deprecated. These undocumented functions were never integrated into the EVP
799 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
800 Bi-directional IGE mode. These modes were never formally standardised and
801 usage of these functions is believed to be very small. In particular
802 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
803 is ever used. The security implications are believed to be minimal, but
804 this issue was never fixed for backwards compatibility reasons. New code
805 should not use these modes.
809 * Add prediction resistance to the DRBG reseeding process.
813 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
814 mandated by IEEE Std 1619-2018.
818 * Added newline escaping functionality to a filename when using openssl dgst.
819 This output format is to replicate the output format found in the `*sum`
820 checksum programs. This aims to preserve backward compatibility.
822 *Matt Eaton, Richard Levitte, and Paul Dale*
824 * Removed the heartbeat message in DTLS feature, as it has very
825 little usage and doesn't seem to fulfill a valuable purpose.
826 The configuration option is now deprecated.
830 * Changed the output of 'openssl {digestname} < file' to display the
831 digest name in its output.
835 * Added a new generic trace API which provides support for enabling
836 instrumentation through trace output. This feature is mainly intended
837 as an aid for developers and is disabled by default. To utilize it,
838 OpenSSL needs to be configured with the `enable-trace` option.
840 If the tracing API is enabled, the application can activate trace output
841 by registering BIOs as trace channels for a number of tracing and debugging
844 The 'openssl' application has been expanded to enable any of the types
845 available via environment variables defined by the user, and serves as
846 one possible example on how to use this functionality.
848 *Richard Levitte & Matthias St. Pierre*
850 * Added build tests for C++. These are generated files that only do one
851 thing, to include one public OpenSSL head file each. This tests that
852 the public header files can be usefully included in a C++ application.
854 This test isn't enabled by default. It can be enabled with the option
855 'enable-buildtest-c++'.
859 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
863 * Add KMAC to EVP_MAC.
867 * Added property based algorithm implementation selection framework to
872 * Added SCA hardening for modular field inversion in EC_GROUP through
873 a new dedicated field_inv() pointer in EC_METHOD.
874 This also addresses a leakage affecting conversions from projective
875 to affine coordinates.
877 *Billy Bob Brumley, Nicola Tuveri*
879 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
880 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
881 those algorithms that were already supported through the EVP_PKEY API
882 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
883 and scrypt are now wrappers that call EVP_KDF.
887 * Build devcrypto engine as a dynamic engine.
891 * Add keyed BLAKE2 to EVP_MAC.
895 * Fix a bug in the computation of the endpoint-pair shared secret used
896 by DTLS over SCTP. This breaks interoperability with older versions
897 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
898 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
899 interoperability with such broken implementations. However, enabling
900 this switch breaks interoperability with correct implementations.
902 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
903 re-used X509_PUBKEY object if the second PUBKEY is malformed.
907 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
911 * Change the license to the Apache License v2.0.
915 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
917 - Major releases (indicated by incrementing the MAJOR release number)
918 may introduce incompatible API/ABI changes.
919 - Minor releases (indicated by incrementing the MINOR release number)
920 may introduce new features but retain API/ABI compatibility.
921 - Patch releases (indicated by incrementing the PATCH number)
922 are intended for bug fixes and other improvements of existing
923 features only (like improving performance or adding documentation)
924 and retain API/ABI compatibility.
928 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
932 * Remove the 'dist' target and add a tarball building script. The
933 'dist' target has fallen out of use, and it shouldn't be
934 necessary to configure just to create a source distribution.
938 * Recreate the OS390-Unix config target. It no longer relies on a
939 special script like it did for OpenSSL pre-1.1.0.
943 * Instead of having the source directories listed in Configure, add
944 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
949 * Add GMAC to EVP_MAC.
953 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
957 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
958 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
959 to facilitate the continued use of MACs through raw private keys in
960 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
964 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
965 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
969 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
970 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
971 are retained for backwards compatibility.
975 * AES-XTS mode now enforces that its two keys are different to mitigate
976 the attacked described in "Efficient Instantiations of Tweakable
977 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
978 Details of this attack can be obtained from:
979 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
983 * Rename the object files, i.e. give them other names than in previous
984 versions. Their names now include the name of the final product, as
985 well as its type mnemonic (bin, lib, shlib).
989 * Added new option for 'openssl list', '-objects', which will display the
990 list of built in objects, i.e. OIDs with names.
994 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
995 improves application performance by removing data copies and providing
996 applications with zero-copy system calls such as sendfile and splice.
1003 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1005 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1007 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1008 while reading in libssl then we would report an error back to the
1009 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1010 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1011 therefore give a hint as to what went wrong.
1015 * Check that ed25519 and ed448 are allowed by the security level. Previously
1016 signature algorithms not using an MD were not being checked that they were
1017 allowed by the security level.
1021 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1022 was not quite right. The behaviour was not consistent between resumption
1023 and normal handshakes, and also not quite consistent with historical
1024 behaviour. The behaviour in various scenarios has been clarified and
1025 it has been updated to make it match historical behaviour as closely as
1030 * *[VMS only]* The header files that the VMS compilers include automatically,
1031 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1032 that the C++ compiler doesn't understand. This is a shortcoming in the
1033 compiler, but can be worked around with `__cplusplus` guards.
1035 C++ applications that use OpenSSL libraries must be compiled using the
1036 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1037 functions. Otherwise, only functions with symbols of less than 31
1038 characters can be used, as the linker will not be able to successfully
1039 resolve symbols with longer names.
1043 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1044 The presence of this system service is determined at run-time.
1048 * Added newline escaping functionality to a filename when using openssl dgst.
1049 This output format is to replicate the output format found in the `*sum`
1050 checksum programs. This aims to preserve backward compatibility.
1052 *Matt Eaton, Richard Levitte, and Paul Dale*
1054 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1059 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1061 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1062 number generator (RNG). This was intended to include protection in the
1063 event of a fork() system call in order to ensure that the parent and child
1064 processes did not share the same RNG state. However this protection was not
1065 being used in the default case.
1067 A partial mitigation for this issue is that the output from a high
1068 precision timer is mixed into the RNG state so the likelihood of a parent
1069 and child process sharing state is significantly reduced.
1071 If an application already calls OPENSSL_init_crypto() explicitly using
1072 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1075 *Matthias St. Pierre*
1077 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1078 used even when parsing explicit parameters, when loading a serialized key
1079 or calling `EC_GROUP_new_from_ecpkparameters()`/
1080 `EC_GROUP_new_from_ecparameters()`.
1081 This prevents bypass of security hardening and performance gains,
1082 especially for curves with specialized EC_METHODs.
1083 By default, if a key encoded with explicit parameters is loaded and later
1084 serialized, the output is still encoded with explicit parameters, even if
1085 internally a "named" EC_GROUP is used for computation.
1089 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1090 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1091 NULL. After this change, only the cofactor parameter can be NULL. It also
1092 does some minimal sanity checks on the passed order.
1097 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1098 An attack is simple, if the first CMS_recipientInfo is valid but the
1099 second CMS_recipientInfo is chosen ciphertext. If the second
1100 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1101 encryption key will be replaced by garbage, and the message cannot be
1102 decoded, but if the RSA decryption fails, the correct encryption key is
1103 used and the recipient will not notice the attack.
1104 As a work around for this potential attack the length of the decrypted
1105 key must be equal to the cipher default key length, in case the
1106 certifiate is not given and all recipientInfo are tried out.
1107 The old behaviour can be re-enabled in the CMS code by setting the
1108 CMS_DEBUG_DECRYPT flag.
1113 * Early start up entropy quality from the DEVRANDOM seed source has been
1114 improved for older Linux systems. The RAND subsystem will wait for
1115 /dev/random to be producing output before seeding from /dev/urandom.
1116 The seeded state is stored for future library initialisations using
1117 a system global shared memory segment. The shared memory identifier
1118 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1119 the desired value. The default identifier is 114.
1123 * Correct the extended master secret constant on EBCDIC systems. Without this
1124 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1125 negotiate EMS will fail. Unfortunately this also means that TLS connections
1126 between EBCDIC systems with this fix, and EBCDIC systems without this
1127 fix will fail if they negotiate EMS.
1131 * Use Windows installation paths in the mingw builds
1133 Mingw isn't a POSIX environment per se, which means that Windows
1134 paths should be used for installation.
1139 * Changed DH_check to accept parameters with order q and 2q subgroups.
1140 With order 2q subgroups the bit 0 of the private key is not secret
1141 but DH_generate_key works around that by clearing bit 0 of the
1142 private key for those. This avoids leaking bit 0 of the private key.
1146 * Significantly reduce secure memory usage by the randomness pools.
1150 * Revert the DEVRANDOM_WAIT feature for Linux systems
1152 The DEVRANDOM_WAIT feature added a select() call to wait for the
1153 /dev/random device to become readable before reading from the
1154 /dev/urandom device.
1156 It turned out that this change had negative side effects on
1157 performance which were not acceptable. After some discussion it
1158 was decided to revert this feature and leave it up to the OS
1159 resp. the platform maintainer to ensure a proper initialization
1160 during early boot time.
1162 *Matthias St. Pierre*
1164 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1166 * Add build tests for C++. These are generated files that only do one
1167 thing, to include one public OpenSSL head file each. This tests that
1168 the public header files can be usefully included in a C++ application.
1170 This test isn't enabled by default. It can be enabled with the option
1171 'enable-buildtest-c++'.
1175 * Enable SHA3 pre-hashing for ECDSA and DSA.
1179 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1180 This changes the size when using the genpkey app when no size is given. It
1181 fixes an omission in earlier changes that changed all RSA, DSA and DH
1182 generation apps to use 2048 bits by default.
1186 * Reorganize the manual pages to consistently have RETURN VALUES,
1187 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1188 util/fix-doc-nits accordingly.
1190 *Paul Yang, Joshua Lock*
1192 * Add the missing accessor EVP_PKEY_get0_engine()
1196 * Have apps like 's_client' and 's_server' output the signature scheme
1197 along with other cipher suite parameters when debugging.
1201 * Make OPENSSL_config() error agnostic again.
1205 * Do the error handling in RSA decryption constant time.
1209 * Prevent over long nonces in ChaCha20-Poly1305.
1211 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1212 for every encryption operation. RFC 7539 specifies that the nonce value
1213 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1214 and front pads the nonce with 0 bytes if it is less than 12
1215 bytes. However it also incorrectly allows a nonce to be set of up to 16
1216 bytes. In this case only the last 12 bytes are significant and any
1217 additional leading bytes are ignored.
1219 It is a requirement of using this cipher that nonce values are
1220 unique. Messages encrypted using a reused nonce value are susceptible to
1221 serious confidentiality and integrity attacks. If an application changes
1222 the default nonce length to be longer than 12 bytes and then makes a
1223 change to the leading bytes of the nonce expecting the new value to be a
1224 new unique nonce then such an application could inadvertently encrypt
1225 messages with a reused nonce.
1227 Additionally the ignored bytes in a long nonce are not covered by the
1228 integrity guarantee of this cipher. Any application that relies on the
1229 integrity of these ignored leading bytes of a long nonce may be further
1230 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1231 is safe because no such use sets such a long nonce value. However user
1232 applications that use this cipher directly and set a non-default nonce
1233 length to be longer than 12 bytes may be vulnerable.
1235 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1241 * Add DEVRANDOM_WAIT feature for Linux systems
1243 On older Linux systems where the getrandom() system call is not available,
1244 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1245 Contrary to getrandom(), the /dev/urandom device will not block during
1246 early boot when the kernel CSPRNG has not been seeded yet.
1248 To mitigate this known weakness, use select() to wait for /dev/random to
1249 become readable before reading from /dev/urandom.
1251 * Ensure that SM2 only uses SM3 as digest algorithm
1255 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1257 * Change the info callback signals for the start and end of a post-handshake
1258 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1259 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1260 confused by this and assume that a TLSv1.2 renegotiation has started. This
1261 can break KeyUpdate handling. Instead we no longer signal the start and end
1262 of a post handshake message exchange (although the messages themselves are
1263 still signalled). This could break some applications that were expecting
1264 the old signals. However without this KeyUpdate is not usable for many
1269 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1271 * Timing vulnerability in DSA signature generation
1273 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1274 timing side channel attack. An attacker could use variations in the signing
1275 algorithm to recover the private key.
1277 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1282 * Timing vulnerability in ECDSA signature generation
1284 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1285 timing side channel attack. An attacker could use variations in the signing
1286 algorithm to recover the private key.
1288 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1293 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1294 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1295 of two gigabytes and the error handling improved.
1297 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1298 categorized as a normal bug, not a security issue, because the DRBG reseeds
1299 automatically and is fully functional even without additional randomness
1300 provided by the application.
1302 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1304 * Add a new ClientHello callback. Provides a callback interface that gives
1305 the application the ability to adjust the nascent SSL object at the
1306 earliest stage of ClientHello processing, immediately after extensions have
1307 been collected but before they have been processed. In particular, this
1308 callback can adjust the supported TLS versions in response to the contents
1313 * Add SM2 base algorithm support.
1317 * s390x assembly pack: add (improved) hardware-support for the following
1318 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1319 aes-cfb/cfb8, aes-ecb.
1323 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1324 parameter is no longer accepted, as it leads to a corrupt table. NULL
1325 pem_str is reserved for alias entries only.
1329 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1330 step for prime curves. The new implementation is based on formulae from
1331 differential addition-and-doubling in homogeneous projective coordinates
1332 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1333 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1334 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1335 to work in projective coordinates.
1337 *Billy Bob Brumley, Nicola Tuveri*
1339 * Change generating and checking of primes so that the error rate of not
1340 being prime depends on the intended use based on the size of the input.
1341 For larger primes this will result in more rounds of Miller-Rabin.
1342 The maximal error rate for primes with more than 1080 bits is lowered
1345 *Kurt Roeckx, Annie Yousar*
1347 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1351 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1352 moving between systems, and to avoid confusion when a Windows build is
1353 done with mingw vs with MSVC. For POSIX installs, there's still a
1354 symlink or copy named 'tsget' to avoid that confusion as well.
1358 * Revert blinding in ECDSA sign and instead make problematic addition
1359 length-invariant. Switch even to fixed-length Montgomery multiplication.
1363 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1364 step for binary curves. The new implementation is based on formulae from
1365 differential addition-and-doubling in mixed Lopez-Dahab projective
1366 coordinates, modified to independently blind the operands.
1368 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1370 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1371 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1372 EC_METHODs to implement their own specialized "ladder step", to take
1373 advantage of more favorable coordinate systems or more efficient
1374 differential addition-and-doubling algorithms.
1376 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1378 * Modified the random device based seed sources to keep the relevant
1379 file descriptors open rather than reopening them on each access.
1380 This allows such sources to operate in a chroot() jail without
1381 the associated device nodes being available. This behaviour can be
1382 controlled using RAND_keep_random_devices_open().
1386 * Numerous side-channel attack mitigations have been applied. This may have
1387 performance impacts for some algorithms for the benefit of improved
1388 security. Specific changes are noted in this change log by their respective
1393 * AIX shared library support overhaul. Switch to AIX "natural" way of
1394 handling shared libraries, which means collecting shared objects of
1395 different versions and bitnesses in one common archive. This allows to
1396 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1397 doesn't affect the way 3rd party applications are linked, only how
1398 multi-version installation is managed.
1402 * Make ec_group_do_inverse_ord() more robust and available to other
1403 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1404 mitigations are applied to the fallback BN_mod_inverse().
1405 When using this function rather than BN_mod_inverse() directly, new
1406 EC cryptosystem implementations are then safer-by-default.
1410 * Add coordinate blinding for EC_POINT and implement projective
1411 coordinate blinding for generic prime curves as a countermeasure to
1412 chosen point SCA attacks.
1414 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1416 * Add blinding to ECDSA and DSA signatures to protect against side channel
1417 attacks discovered by Keegan Ryan (NCC Group).
1421 * Enforce checking in the pkeyutl command line app to ensure that the input
1422 length does not exceed the maximum supported digest length when performing
1423 a sign, verify or verifyrecover operation.
1427 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1428 I/O in combination with something like select() or poll() will hang. This
1429 can be turned off again using SSL_CTX_clear_mode().
1430 Many applications do not properly handle non-application data records, and
1431 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1432 around the problems in those applications, but can also break some.
1433 It's recommended to read the manpages about SSL_read(), SSL_write(),
1434 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1435 SSL_CTX_set_read_ahead() again.
1439 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1440 now allow empty (zero character) pass phrases.
1444 * Apply blinding to binary field modular inversion and remove patent
1445 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1449 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1450 binary and prime elliptic curves.
1454 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1455 constant time fixed point multiplication.
1459 * Revise elliptic curve scalar multiplication with timing attack
1460 defenses: ec_wNAF_mul redirects to a constant time implementation
1461 when computing fixed point and variable point multiplication (which
1462 in OpenSSL are mostly used with secret scalars in keygen, sign,
1463 ECDH derive operations).
1464 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1467 * Updated CONTRIBUTING
1471 * Updated DRBG / RAND to request nonce and additional low entropy
1472 randomness from the system.
1474 *Matthias St. Pierre*
1476 * Updated 'openssl rehash' to use OpenSSL consistent default.
1480 * Moved the load of the ssl_conf module to libcrypto, which helps
1481 loading engines that libssl uses before libssl is initialised.
1485 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1489 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1491 *Ingo Schwarze, Rich Salz*
1493 * Added output of accepting IP address and port for 'openssl s_server'
1497 * Added a new API for TLSv1.3 ciphersuites:
1498 SSL_CTX_set_ciphersuites()
1499 SSL_set_ciphersuites()
1503 * Memory allocation failures consistently add an error to the error
1508 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1509 in libcrypto when run as setuid/setgid.
1513 * Load any config file by default when libssl is used.
1517 * Added new public header file <openssl/rand_drbg.h> and documentation
1518 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1520 *Matthias St. Pierre*
1522 * QNX support removed (cannot find contributors to get their approval
1523 for the license change).
1527 * TLSv1.3 replay protection for early data has been implemented. See the
1528 SSL_read_early_data() man page for further details.
1532 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1533 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1534 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1535 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1536 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1537 configuration has been separated out. See the ciphers man page or the
1538 SSL_CTX_set_ciphersuites() man page for more information.
1542 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1543 in responder mode now supports the new "-multi" option, which
1544 spawns the specified number of child processes to handle OCSP
1545 requests. The "-timeout" option now also limits the OCSP
1546 responder's patience to wait to receive the full client request
1547 on a newly accepted connection. Child processes are respawned
1548 as needed, and the CA index file is automatically reloaded
1549 when changed. This makes it possible to run the "ocsp" responder
1550 as a long-running service, making the OpenSSL CA somewhat more
1551 feature-complete. In this mode, most diagnostic messages logged
1552 after entering the event loop are logged via syslog(3) rather than
1557 * Added support for X448 and Ed448. Heavily based on original work by
1562 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1563 objects loaded. This adds the functions OSSL_STORE_expect() and
1564 OSSL_STORE_find() as well as needed tools to construct searches and
1565 get the search data out of them.
1569 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1570 version of OpenSSL should review their configuration settings to ensure
1571 that they are still appropriate for TLSv1.3. For further information see:
1572 <https://wiki.openssl.org/index.php/TLS1.3>
1576 * Grand redesign of the OpenSSL random generator
1578 The default RAND method now utilizes an AES-CTR DRBG according to
1579 NIST standard SP 800-90Ar1. The new random generator is essentially
1580 a port of the default random generator from the OpenSSL FIPS 2.0
1581 object module. It is a hybrid deterministic random bit generator
1582 using an AES-CTR bit stream and which seeds and reseeds itself
1583 automatically using trusted system entropy sources.
1585 Some of its new features are:
1586 - Support for multiple DRBG instances with seed chaining.
1587 - The default RAND method makes use of a DRBG.
1588 - There is a public and private DRBG instance.
1589 - The DRBG instances are fork-safe.
1590 - Keep all global DRBG instances on the secure heap if it is enabled.
1591 - The public and private DRBG instance are per thread for lock free
1594 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1596 * Changed Configure so it only says what it does and doesn't dump
1597 so much data. Instead, ./configdata.pm should be used as a script
1598 to display all sorts of configuration data.
1602 * Added processing of "make variables" to Configure.
1606 * Added SHA512/224 and SHA512/256 algorithm support.
1610 * The last traces of Netware support, first removed in 1.1.0, have
1615 * Get rid of Makefile.shared, and in the process, make the processing
1616 of certain files (rc.obj, or the .def/.map/.opt files produced from
1617 the ordinal files) more visible and hopefully easier to trace and
1618 debug (or make silent).
1622 * Make it possible to have environment variable assignments as
1623 arguments to config / Configure.
1627 * Add multi-prime RSA (RFC 8017) support.
1631 * Add SM3 implemented according to GB/T 32905-2016
1632 * Jack Lloyd <jack.lloyd@ribose.com>,
1633 Ronald Tse <ronald.tse@ribose.com>,
1634 Erick Borsboom <erick.borsboom@ribose.com> *
1636 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1637 as documented in RFC6066.
1638 Based on a patch from Tomasz Moń
1640 *Filipe Raimundo da Silva*
1642 * Add SM4 implemented according to GB/T 32907-2016.
1643 * Jack Lloyd <jack.lloyd@ribose.com>,
1644 Ronald Tse <ronald.tse@ribose.com>,
1645 Erick Borsboom <erick.borsboom@ribose.com> *
1647 * Reimplement -newreq-nodes and ERR_error_string_n; the
1648 original author does not agree with the license change.
1652 * Add ARIA AEAD TLS support.
1656 * Some macro definitions to support VS6 have been removed. Visual
1657 Studio 6 has not worked since 1.1.0
1661 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1662 without clearing the errors.
1666 * Add "atfork" functions. If building on a system that without
1667 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1668 requirements. The RAND facility now uses/requires this.
1676 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1677 not possible to disable entirely. However, it's still possible to
1678 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1681 To disable, configure with 'no-ui-console'. 'no-ui' is still
1682 possible to use as an alias. Check at compile time with the
1683 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1684 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1688 * Add a STORE module, which implements a uniform and URI based reader of
1689 stores that can contain keys, certificates, CRLs and numerous other
1690 objects. The main API is loosely based on a few stdio functions,
1691 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1692 OSSL_STORE_error and OSSL_STORE_close.
1693 The implementation uses backends called "loaders" to implement arbitrary
1694 URI schemes. There is one built in "loader" for the 'file' scheme.
1698 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1699 then adjusted to work on FreeBSD 8.4 as well.
1700 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1701 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1705 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1706 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1707 error code calls like this:
1709 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1711 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1712 that can be encoded in C. For the foreseeable future, this will only
1715 *Richard Levitte and Tim Hudson*
1717 * Removed BSD cryptodev engine.
1721 * Add a build target 'build_all_generated', to build all generated files
1722 and only that. This can be used to prepare everything that requires
1723 things like perl for a system that lacks perl and then move everything
1724 to that system and do the rest of the build there.
1728 * In the UI interface, make it possible to duplicate the user data. This
1729 can be used by engines that need to retain the data for a longer time
1730 than just the call where this user data is passed.
1734 * Ignore the '-named_curve auto' value for compatibility of applications
1737 *Tomas Mraz <tmraz@fedoraproject.org>*
1739 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1740 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1741 alerts across multiple records (some of which could be empty). In practice
1742 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1743 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1744 support this at all. Supporting it adds significant complexity to the
1745 record layer, and its removal is unlikely to cause interoperability
1750 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1751 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1752 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1757 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1758 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1760 *Richard Levitte, Andy Polyakov*
1762 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1767 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1768 platform rather than 'mingw'.
1772 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1773 success if they are asked to add an object which already exists
1774 in the store. This change cascades to other functions which load
1775 certificates and CRLs.
1779 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1780 facilitate stack unwinding even from assembly subroutines.
1784 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1785 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1789 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1790 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1791 which is the minimum version we support.
1795 * Certificate time validation (X509_cmp_time) enforces stricter
1796 compliance with RFC 5280. Fractional seconds and timezone offsets
1797 are no longer allowed.
1801 * Add support for ARIA
1805 * s_client will now send the Server Name Indication (SNI) extension by
1806 default unless the new "-noservername" option is used. The server name is
1807 based on the host provided to the "-connect" option unless overridden by
1808 using "-servername".
1812 * Add support for SipHash
1816 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1817 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1818 prevent issues where no progress is being made and the peer continually
1819 sends unrecognised record types, using up resources processing them.
1823 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1824 using the algorithm defined in
1825 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1829 * Heartbeat support has been removed; the ABI is changed for now.
1831 *Richard Levitte, Rich Salz*
1833 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1837 * The RSA "null" method, which was partially supported to avoid patent
1838 issues, has been replaced to always returns NULL.
1845 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1847 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1848 used even when parsing explicit parameters, when loading a serialized key
1849 or calling `EC_GROUP_new_from_ecpkparameters()`/
1850 `EC_GROUP_new_from_ecparameters()`.
1851 This prevents bypass of security hardening and performance gains,
1852 especially for curves with specialized EC_METHODs.
1853 By default, if a key encoded with explicit parameters is loaded and later
1854 serialized, the output is still encoded with explicit parameters, even if
1855 internally a "named" EC_GROUP is used for computation.
1859 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1860 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1861 NULL. After this change, only the cofactor parameter can be NULL. It also
1862 does some minimal sanity checks on the passed order.
1867 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1868 An attack is simple, if the first CMS_recipientInfo is valid but the
1869 second CMS_recipientInfo is chosen ciphertext. If the second
1870 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1871 encryption key will be replaced by garbage, and the message cannot be
1872 decoded, but if the RSA decryption fails, the correct encryption key is
1873 used and the recipient will not notice the attack.
1874 As a work around for this potential attack the length of the decrypted
1875 key must be equal to the cipher default key length, in case the
1876 certifiate is not given and all recipientInfo are tried out.
1877 The old behaviour can be re-enabled in the CMS code by setting the
1878 CMS_DEBUG_DECRYPT flag.
1883 * Use Windows installation paths in the mingw builds
1885 Mingw isn't a POSIX environment per se, which means that Windows
1886 paths should be used for installation.
1891 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
1893 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1894 This changes the size when using the genpkey app when no size is given. It
1895 fixes an omission in earlier changes that changed all RSA, DSA and DH
1896 generation apps to use 2048 bits by default.
1900 * Prevent over long nonces in ChaCha20-Poly1305.
1902 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1903 for every encryption operation. RFC 7539 specifies that the nonce value
1904 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1905 and front pads the nonce with 0 bytes if it is less than 12
1906 bytes. However it also incorrectly allows a nonce to be set of up to 16
1907 bytes. In this case only the last 12 bytes are significant and any
1908 additional leading bytes are ignored.
1910 It is a requirement of using this cipher that nonce values are
1911 unique. Messages encrypted using a reused nonce value are susceptible to
1912 serious confidentiality and integrity attacks. If an application changes
1913 the default nonce length to be longer than 12 bytes and then makes a
1914 change to the leading bytes of the nonce expecting the new value to be a
1915 new unique nonce then such an application could inadvertently encrypt
1916 messages with a reused nonce.
1918 Additionally the ignored bytes in a long nonce are not covered by the
1919 integrity guarantee of this cipher. Any application that relies on the
1920 integrity of these ignored leading bytes of a long nonce may be further
1921 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1922 is safe because no such use sets such a long nonce value. However user
1923 applications that use this cipher directly and set a non-default nonce
1924 length to be longer than 12 bytes may be vulnerable.
1926 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1932 * Added SCA hardening for modular field inversion in EC_GROUP through
1933 a new dedicated field_inv() pointer in EC_METHOD.
1934 This also addresses a leakage affecting conversions from projective
1935 to affine coordinates.
1937 *Billy Bob Brumley, Nicola Tuveri*
1939 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1940 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1944 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1948 * Remove the 'dist' target and add a tarball building script. The
1949 'dist' target has fallen out of use, and it shouldn't be
1950 necessary to configure just to create a source distribution.
1954 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
1956 * Timing vulnerability in DSA signature generation
1958 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1959 timing side channel attack. An attacker could use variations in the signing
1960 algorithm to recover the private key.
1962 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1967 * Timing vulnerability in ECDSA signature generation
1969 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1970 timing side channel attack. An attacker could use variations in the signing
1971 algorithm to recover the private key.
1973 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1978 * Add coordinate blinding for EC_POINT and implement projective
1979 coordinate blinding for generic prime curves as a countermeasure to
1980 chosen point SCA attacks.
1982 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1984 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
1986 * Client DoS due to large DH parameter
1988 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1989 malicious server can send a very large prime value to the client. This will
1990 cause the client to spend an unreasonably long period of time generating a
1991 key for this prime resulting in a hang until the client has finished. This
1992 could be exploited in a Denial Of Service attack.
1994 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1999 * Cache timing vulnerability in RSA Key Generation
2001 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2002 a cache timing side channel attack. An attacker with sufficient access to
2003 mount cache timing attacks during the RSA key generation process could
2004 recover the private key.
2006 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2007 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2012 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2013 parameter is no longer accepted, as it leads to a corrupt table. NULL
2014 pem_str is reserved for alias entries only.
2018 * Revert blinding in ECDSA sign and instead make problematic addition
2019 length-invariant. Switch even to fixed-length Montgomery multiplication.
2023 * Change generating and checking of primes so that the error rate of not
2024 being prime depends on the intended use based on the size of the input.
2025 For larger primes this will result in more rounds of Miller-Rabin.
2026 The maximal error rate for primes with more than 1080 bits is lowered
2029 *Kurt Roeckx, Annie Yousar*
2031 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2035 * Add blinding to ECDSA and DSA signatures to protect against side channel
2036 attacks discovered by Keegan Ryan (NCC Group).
2040 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2041 now allow empty (zero character) pass phrases.
2045 * Certificate time validation (X509_cmp_time) enforces stricter
2046 compliance with RFC 5280. Fractional seconds and timezone offsets
2047 are no longer allowed.
2051 * Fixed a text canonicalisation bug in CMS
2053 Where a CMS detached signature is used with text content the text goes
2054 through a canonicalisation process first prior to signing or verifying a
2055 signature. This process strips trailing space at the end of lines, converts
2056 line terminators to CRLF and removes additional trailing line terminators
2057 at the end of a file. A bug in the canonicalisation process meant that
2058 some characters, such as form-feed, were incorrectly treated as whitespace
2059 and removed. This is contrary to the specification (RFC5485). This fix
2060 could mean that detached text data signed with an earlier version of
2061 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2062 signed with a fixed OpenSSL may fail to verify with an earlier version of
2063 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2064 and use the "-binary" flag (for the "cms" command line application) or set
2065 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2069 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2071 * Constructed ASN.1 types with a recursive definition could exceed the stack
2073 Constructed ASN.1 types with a recursive definition (such as can be found
2074 in PKCS7) could eventually exceed the stack given malicious input with
2075 excessive recursion. This could result in a Denial Of Service attack. There
2076 are no such structures used within SSL/TLS that come from untrusted sources
2077 so this is considered safe.
2079 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2085 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2087 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2088 effectively reduced to only comparing the least significant bit of each
2089 byte. This allows an attacker to forge messages that would be considered as
2090 authenticated in an amount of tries lower than that guaranteed by the
2091 security claims of the scheme. The module can only be compiled by the
2092 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2094 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2100 * Add a build target 'build_all_generated', to build all generated files
2101 and only that. This can be used to prepare everything that requires
2102 things like perl for a system that lacks perl and then move everything
2103 to that system and do the rest of the build there.
2107 * Backport SSL_OP_NO_RENGOTIATION
2109 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2110 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2111 changes this is no longer possible in 1.1.0. Therefore the new
2112 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2113 1.1.0 to provide equivalent functionality.
2115 Note that if an application built against 1.1.0h headers (or above) is run
2116 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2117 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2121 * Removed the OS390-Unix config target. It relied on a script that doesn't
2126 * rsaz_1024_mul_avx2 overflow bug on x86_64
2128 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2129 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2130 Analysis suggests that attacks against RSA and DSA as a result of this
2131 defect would be very difficult to perform and are not believed likely.
2132 Attacks against DH1024 are considered just feasible, because most of the
2133 work necessary to deduce information about a private key may be performed
2134 offline. The amount of resources required for such an attack would be
2135 significant. However, for an attack on TLS to be meaningful, the server
2136 would have to share the DH1024 private key among multiple clients, which is
2137 no longer an option since CVE-2016-0701.
2139 This only affects processors that support the AVX2 but not ADX extensions
2140 like Intel Haswell (4th generation).
2142 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2143 was originally found via the OSS-Fuzz project.
2148 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2150 * bn_sqrx8x_internal carry bug on x86_64
2152 There is a carry propagating bug in the x86_64 Montgomery squaring
2153 procedure. No EC algorithms are affected. Analysis suggests that attacks
2154 against RSA and DSA as a result of this defect would be very difficult to
2155 perform and are not believed likely. Attacks against DH are considered just
2156 feasible (although very difficult) because most of the work necessary to
2157 deduce information about a private key may be performed offline. The amount
2158 of resources required for such an attack would be very significant and
2159 likely only accessible to a limited number of attackers. An attacker would
2160 additionally need online access to an unpatched system using the target
2161 private key in a scenario with persistent DH parameters and a private
2162 key that is shared between multiple clients.
2164 This only affects processors that support the BMI1, BMI2 and ADX extensions
2165 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2167 This issue was reported to OpenSSL by the OSS-Fuzz project.
2172 * Malformed X.509 IPAddressFamily could cause OOB read
2174 If an X.509 certificate has a malformed IPAddressFamily extension,
2175 OpenSSL could do a one-byte buffer overread. The most likely result
2176 would be an erroneous display of the certificate in text format.
2178 This issue was reported to OpenSSL by the OSS-Fuzz project.
2183 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2185 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2186 platform rather than 'mingw'.
2190 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2191 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2192 which is the minimum version we support.
2196 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2198 * Encrypt-Then-Mac renegotiation crash
2200 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2201 negotiated where it was not in the original handshake (or vice-versa) then
2202 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2203 and servers are affected.
2205 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2210 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2212 * Truncated packet could crash via OOB read
2214 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2215 cipher is being used, then a truncated packet can cause that host to
2216 perform an out-of-bounds read, usually resulting in a crash.
2218 This issue was reported to OpenSSL by Robert Święcki of Google.
2223 * Bad (EC)DHE parameters cause a client crash
2225 If a malicious server supplies bad parameters for a DHE or ECDHE key
2226 exchange then this can result in the client attempting to dereference a
2227 NULL pointer leading to a client crash. This could be exploited in a Denial
2230 This issue was reported to OpenSSL by Guido Vranken.
2235 * BN_mod_exp may produce incorrect results on x86_64
2237 There is a carry propagating bug in the x86_64 Montgomery squaring
2238 procedure. No EC algorithms are affected. Analysis suggests that attacks
2239 against RSA and DSA as a result of this defect would be very difficult to
2240 perform and are not believed likely. Attacks against DH are considered just
2241 feasible (although very difficult) because most of the work necessary to
2242 deduce information about a private key may be performed offline. The amount
2243 of resources required for such an attack would be very significant and
2244 likely only accessible to a limited number of attackers. An attacker would
2245 additionally need online access to an unpatched system using the target
2246 private key in a scenario with persistent DH parameters and a private
2247 key that is shared between multiple clients. For example this can occur by
2248 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2249 similar to CVE-2015-3193 but must be treated as a separate problem.
2251 This issue was reported to OpenSSL by the OSS-Fuzz project.
2256 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2258 * ChaCha20/Poly1305 heap-buffer-overflow
2260 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2261 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2262 crash. This issue is not considered to be exploitable beyond a DoS.
2264 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2269 * CMS Null dereference
2271 Applications parsing invalid CMS structures can crash with a NULL pointer
2272 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2273 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2274 structure callback if an attempt is made to free certain invalid encodings.
2275 Only CHOICE structures using a callback which do not handle NULL value are
2278 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2283 * Montgomery multiplication may produce incorrect results
2285 There is a carry propagating bug in the Broadwell-specific Montgomery
2286 multiplication procedure that handles input lengths divisible by, but
2287 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2288 and DH private keys are impossible. This is because the subroutine in
2289 question is not used in operations with the private key itself and an input
2290 of the attacker's direct choice. Otherwise the bug can manifest itself as
2291 transient authentication and key negotiation failures or reproducible
2292 erroneous outcome of public-key operations with specially crafted input.
2293 Among EC algorithms only Brainpool P-512 curves are affected and one
2294 presumably can attack ECDH key negotiation. Impact was not analyzed in
2295 detail, because pre-requisites for attack are considered unlikely. Namely
2296 multiple clients have to choose the curve in question and the server has to
2297 share the private key among them, neither of which is default behaviour.
2298 Even then only clients that chose the curve will be affected.
2300 This issue was publicly reported as transient failures and was not
2301 initially recognized as a security issue. Thanks to Richard Morgan for
2302 providing reproducible case.
2307 * Removed automatic addition of RPATH in shared libraries and executables,
2308 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2312 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2314 * Fix Use After Free for large message sizes
2316 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2317 message larger than approx 16k is received then the underlying buffer to
2318 store the incoming message is reallocated and moved. Unfortunately a
2319 dangling pointer to the old location is left which results in an attempt to
2320 write to the previously freed location. This is likely to result in a
2321 crash, however it could potentially lead to execution of arbitrary code.
2323 This issue only affects OpenSSL 1.1.0a.
2325 This issue was reported to OpenSSL by Robert Święcki.
2330 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2332 * OCSP Status Request extension unbounded memory growth
2334 A malicious client can send an excessively large OCSP Status Request
2335 extension. If that client continually requests renegotiation, sending a
2336 large OCSP Status Request extension each time, then there will be unbounded
2337 memory growth on the server. This will eventually lead to a Denial Of
2338 Service attack through memory exhaustion. Servers with a default
2339 configuration are vulnerable even if they do not support OCSP. Builds using
2340 the "no-ocsp" build time option are not affected.
2342 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2347 * SSL_peek() hang on empty record
2349 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2350 sends an empty record. This could be exploited by a malicious peer in a
2351 Denial Of Service attack.
2353 This issue was reported to OpenSSL by Alex Gaynor.
2358 * Excessive allocation of memory in tls_get_message_header() and
2359 dtls1_preprocess_fragment()
2361 A (D)TLS message includes 3 bytes for its length in the header for the
2362 message. This would allow for messages up to 16Mb in length. Messages of
2363 this length are excessive and OpenSSL includes a check to ensure that a
2364 peer is sending reasonably sized messages in order to avoid too much memory
2365 being consumed to service a connection. A flaw in the logic of version
2366 1.1.0 means that memory for the message is allocated too early, prior to
2367 the excessive message length check. Due to way memory is allocated in
2368 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2369 to service a connection. This could lead to a Denial of Service through
2370 memory exhaustion. However, the excessive message length check still takes
2371 place, and this would cause the connection to immediately fail. Assuming
2372 that the application calls SSL_free() on the failed connection in a timely
2373 manner then the 21Mb of allocated memory will then be immediately freed
2374 again. Therefore the excessive memory allocation will be transitory in
2375 nature. This then means that there is only a security impact if:
2377 1) The application does not call SSL_free() in a timely manner in the event
2378 that the connection fails
2380 2) The application is working in a constrained environment where there is
2381 very little free memory
2383 3) The attacker initiates multiple connection attempts such that there are
2384 multiple connections in a state where memory has been allocated for the
2385 connection; SSL_free() has not yet been called; and there is insufficient
2386 memory to service the multiple requests.
2388 Except in the instance of (1) above any Denial Of Service is likely to be
2389 transitory because as soon as the connection fails the memory is
2390 subsequently freed again in the SSL_free() call. However there is an
2391 increased risk during this period of application crashes due to the lack of
2392 memory - which would then mean a more serious Denial of Service.
2394 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2395 (CVE-2016-6307 and CVE-2016-6308)
2399 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2400 had to be removed. Primary reason is that vendor assembler can't
2401 assemble our modules with -KPIC flag. As result it, assembly
2402 support, was not even available as option. But its lack means
2403 lack of side-channel resistant code, which is incompatible with
2404 security by todays standards. Fortunately gcc is readily available
2405 prepackaged option, which we firmly point at...
2409 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2411 * Windows command-line tool supports UTF-8 opt-in option for arguments
2412 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2413 (to any value) allows Windows user to access PKCS#12 file generated
2414 with Windows CryptoAPI and protected with non-ASCII password, as well
2415 as files generated under UTF-8 locale on Linux also protected with
2420 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2421 have been disabled by default and removed from DEFAULT, just like RC4.
2422 See the RC4 item below to re-enable both.
2426 * The method for finding the storage location for the Windows RAND seed file
2427 has changed. First we check %RANDFILE%. If that is not set then we check
2428 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2429 all else fails we fall back to C:\.
2433 * The EVP_EncryptUpdate() function has had its return type changed from void
2434 to int. A return of 0 indicates and error while a return of 1 indicates
2439 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2440 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2441 off the constant time implementation for RSA, DSA and DH have been made
2442 no-ops and deprecated.
2446 * Windows RAND implementation was simplified to only get entropy by
2447 calling CryptGenRandom(). Various other RAND-related tickets
2450 *Joseph Wylie Yandle, Rich Salz*
2452 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2453 and `OPENSSL_LH_`, respectively. The old names are available
2454 with API compatibility. They new names are now completely documented.
2458 * Unify TYPE_up_ref(obj) methods signature.
2459 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2460 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2461 int (instead of void) like all others TYPE_up_ref() methods.
2462 So now these methods also check the return value of CRYPTO_atomic_add(),
2463 and the validity of object reference counter.
2465 *fdasilvayy@gmail.com*
2467 * With Windows Visual Studio builds, the .pdb files are installed
2468 alongside the installed libraries and executables. For a static
2469 library installation, ossl_static.pdb is the associate compiler
2470 generated .pdb file to be used when linking programs.
2474 * Remove openssl.spec. Packaging files belong with the packagers.
2478 * Automatic Darwin/OSX configuration has had a refresh, it will now
2479 recognise x86_64 architectures automatically. You can still decide
2480 to build for a different bitness with the environment variable
2481 KERNEL_BITS (can be 32 or 64), for example:
2483 KERNEL_BITS=32 ./config
2487 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2488 256 bit AES and HMAC with SHA256.
2492 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2496 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2500 * To enable users to have their own config files and build file templates,
2501 Configure looks in the directory indicated by the environment variable
2502 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2503 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2504 name and is used as is.
2508 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2509 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2510 X509_CERT_FILE_CTX was removed.
2514 * "shared" builds are now the default. To create only static libraries use
2515 the "no-shared" Configure option.
2519 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2520 All of these option have not worked for some while and are fundamental
2525 * Make various cleanup routines no-ops and mark them as deprecated. Most
2526 global cleanup functions are no longer required because they are handled
2527 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2528 Explicitly de-initing can cause problems (e.g. where a library that uses
2529 OpenSSL de-inits, but an application is still using it). The affected
2530 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2531 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2532 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2533 COMP_zlib_cleanup().
2537 * --strict-warnings no longer enables runtime debugging options
2538 such as REF_DEBUG. Instead, debug options are automatically
2539 enabled with '--debug' builds.
2541 *Andy Polyakov, Emilia Käsper*
2543 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2544 have been moved out of the public header files. New functions for managing
2545 these have been added.
2549 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2550 objects have been moved out of the public header files. New
2551 functions for managing these have been added.
2555 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2556 have been moved out of the public header files. New functions for managing
2557 these have been added.
2561 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2562 moved out of the public header files. New functions for managing these
2567 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2571 * Removed the mk1mf build scripts.
2575 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2576 it is always safe to #include a header now.
2580 * Removed the aged BC-32 config and all its supporting scripts
2584 * Removed support for Ultrix, Netware, and OS/2.
2588 * Add support for HKDF.
2590 *Alessandro Ghedini*
2592 * Add support for blake2b and blake2s
2596 * Added support for "pipelining". Ciphers that have the
2597 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2598 encryptions/decryptions simultaneously. There are currently no built-in
2599 ciphers with this property but the expectation is that engines will be able
2600 to offer it to significantly improve throughput. Support has been extended
2601 into libssl so that multiple records for a single connection can be
2602 processed in one go (for >=TLS 1.1).
2606 * Added the AFALG engine. This is an async capable engine which is able to
2607 offload work to the Linux kernel. In this initial version it only supports
2608 AES128-CBC. The kernel must be version 4.1.0 or greater.
2612 * OpenSSL now uses a new threading API. It is no longer necessary to
2613 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2614 are two supported threading models: pthreads and windows threads. It is
2615 also possible to configure OpenSSL at compile time for "no-threads". The
2616 old threading API should no longer be used. The functions have been
2617 replaced with "no-op" compatibility macros.
2619 *Alessandro Ghedini, Matt Caswell*
2621 * Modify behavior of ALPN to invoke callback after SNI/servername
2622 callback, such that updates to the SSL_CTX affect ALPN.
2626 * Add SSL_CIPHER queries for authentication and key-exchange.
2630 * Changes to the DEFAULT cipherlist:
2631 - Prefer (EC)DHE handshakes over plain RSA.
2632 - Prefer AEAD ciphers over legacy ciphers.
2633 - Prefer ECDSA over RSA when both certificates are available.
2634 - Prefer TLSv1.2 ciphers/PRF.
2635 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2640 * Change the ECC default curve list to be this, in order: x25519,
2641 secp256r1, secp521r1, secp384r1.
2645 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2646 disabled by default. They can be re-enabled using the
2647 enable-weak-ssl-ciphers option to Configure.
2651 * If the server has ALPN configured, but supports no protocols that the
2652 client advertises, send a fatal "no_application_protocol" alert.
2653 This behaviour is SHALL in RFC 7301, though it isn't universally
2654 implemented by other servers.
2658 * Add X25519 support.
2659 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2660 for public and private key encoding using the format documented in
2661 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2662 key generation and key derivation.
2664 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2669 * Deprecate SRP_VBASE_get_by_user.
2670 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2671 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2672 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2673 seed, even if the seed is configured.
2675 Users should use SRP_VBASE_get1_by_user instead. Note that in
2676 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2677 also that even though configuring the SRP seed attempts to hide
2678 invalid usernames by continuing the handshake with fake
2679 credentials, this behaviour is not constant time and no strong
2680 guarantees are made that the handshake is indistinguishable from
2681 that of a valid user.
2685 * Configuration change; it's now possible to build dynamic engines
2686 without having to build shared libraries and vice versa. This
2687 only applies to the engines in engines/, those in crypto/engine/
2688 will always be built into libcrypto (i.e. "static").
2690 Building dynamic engines is enabled by default; to disable, use
2691 the configuration option "disable-dynamic-engine".
2693 The only requirements for building dynamic engines are the
2694 presence of the DSO module and building with position independent
2695 code, so they will also automatically be disabled if configuring
2696 with "disable-dso" or "disable-pic".
2698 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2699 are also taken away from openssl/opensslconf.h, as they are
2704 * Configuration change; if there is a known flag to compile
2705 position independent code, it will always be applied on the
2706 libcrypto and libssl object files, and never on the application
2707 object files. This means other libraries that use routines from
2708 libcrypto / libssl can be made into shared libraries regardless
2709 of how OpenSSL was configured.
2711 If this isn't desirable, the configuration options "disable-pic"
2712 or "no-pic" can be used to disable the use of PIC. This will
2713 also disable building shared libraries and dynamic engines.
2717 * Removed JPAKE code. It was experimental and has no wide use.
2721 * The INSTALL_PREFIX Makefile variable has been renamed to
2722 DESTDIR. That makes for less confusion on what this variable
2723 is for. Also, the configuration option --install_prefix is
2728 * Heartbeat for TLS has been removed and is disabled by default
2729 for DTLS; configure with enable-heartbeats. Code that uses the
2730 old #define's might need to be updated.
2732 *Emilia Käsper, Rich Salz*
2734 * Rename REF_CHECK to REF_DEBUG.
2738 * New "unified" build system
2740 The "unified" build system is aimed to be a common system for all
2741 platforms we support. With it comes new support for VMS.
2743 This system builds supports building in a different directory tree
2744 than the source tree. It produces one Makefile (for unix family
2745 or lookalikes), or one descrip.mms (for VMS).
2747 The source of information to make the Makefile / descrip.mms is
2748 small files called 'build.info', holding the necessary
2749 information for each directory with source to compile, and a
2750 template in Configurations, like unix-Makefile.tmpl or
2753 With this change, the library names were also renamed on Windows
2754 and on VMS. They now have names that are closer to the standard
2755 on Unix, and include the major version number, and in certain
2756 cases, the architecture they are built for. See "Notes on shared
2757 libraries" in INSTALL.
2759 We rely heavily on the perl module Text::Template.
2763 * Added support for auto-initialisation and de-initialisation of the library.
2764 OpenSSL no longer requires explicit init or deinit routines to be called,
2765 except in certain circumstances. See the OPENSSL_init_crypto() and
2766 OPENSSL_init_ssl() man pages for further information.
2770 * The arguments to the DTLSv1_listen function have changed. Specifically the
2771 "peer" argument is now expected to be a BIO_ADDR object.
2773 * Rewrite of BIO networking library. The BIO library lacked consistent
2774 support of IPv6, and adding it required some more extensive
2775 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2776 which hold all types of addresses and chains of address information.
2777 It also introduces a new API, with functions like BIO_socket,
2778 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2779 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2780 have been adapted accordingly.
2784 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2789 * CRIME protection: disable compression by default, even if OpenSSL is
2790 compiled with zlib enabled. Applications can still enable compression
2791 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2792 using the SSL_CONF library to configure compression.
2796 * The signature of the session callback configured with
2797 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2798 was explicitly marked as `const unsigned char*` instead of
2803 * Always DPURIFY. Remove the use of uninitialized memory in the
2804 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2808 * Removed many obsolete configuration items, including
2809 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2810 MD2_CHAR, MD2_INT, MD2_LONG
2812 IDEA_SHORT, IDEA_LONG
2813 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2815 *Rich Salz, with advice from Andy Polyakov*
2817 * Many BN internals have been moved to an internal header file.
2819 *Rich Salz with help from Andy Polyakov*
2821 * Configuration and writing out the results from it has changed.
2822 Files such as Makefile include/openssl/opensslconf.h and are now
2823 produced through general templates, such as Makefile.in and
2824 crypto/opensslconf.h.in and some help from the perl module
2827 Also, the center of configuration information is no longer
2828 Makefile. Instead, Configure produces a perl module in
2829 configdata.pm which holds most of the config data (in the hash
2830 table %config), the target data that comes from the target
2831 configuration in one of the `Configurations/*.conf~ files (in
2836 * To clarify their intended purposes, the Configure options
2837 --prefix and --openssldir change their semantics, and become more
2838 straightforward and less interdependent.
2840 --prefix shall be used exclusively to give the location INSTALLTOP
2841 where programs, scripts, libraries, include files and manuals are
2842 going to be installed. The default is now /usr/local.
2844 --openssldir shall be used exclusively to give the default
2845 location OPENSSLDIR where certificates, private keys, CRLs are
2846 managed. This is also where the default openssl.cnf gets
2848 If the directory given with this option is a relative path, the
2849 values of both the --prefix value and the --openssldir value will
2850 be combined to become OPENSSLDIR.
2851 The default for --openssldir is INSTALLTOP/ssl.
2853 Anyone who uses --openssldir to specify where OpenSSL is to be
2854 installed MUST change to use --prefix instead.
2858 * The GOST engine was out of date and therefore it has been removed. An up
2859 to date GOST engine is now being maintained in an external repository.
2860 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2861 support for GOST ciphersuites (these are only activated if a GOST engine
2866 * EGD is no longer supported by default; use enable-egd when
2869 *Ben Kaduk and Rich Salz*
2871 * The distribution now has Makefile.in files, which are used to
2872 create Makefile's when Configure is run. *Configure must be run
2873 before trying to build now.*
2877 * The return value for SSL_CIPHER_description() for error conditions
2882 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2884 Obtaining and performing DNSSEC validation of TLSA records is
2885 the application's responsibility. The application provides
2886 the TLSA records of its choice to OpenSSL, and these are then
2887 used to authenticate the peer.
2889 The TLSA records need not even come from DNS. They can, for
2890 example, be used to implement local end-entity certificate or
2891 trust-anchor "pinning", where the "pin" data takes the form
2892 of TLSA records, which can augment or replace verification
2893 based on the usual WebPKI public certification authorities.
2897 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2898 continues to support deprecated interfaces in default builds.
2899 However, applications are strongly advised to compile their
2900 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2901 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2902 or the 1.1.0 releases.
2904 In environments in which all applications have been ported to
2905 not use any deprecated interfaces OpenSSL's Configure script
2906 should be used with the --api=1.1.0 option to entirely remove
2907 support for the deprecated features from the library and
2908 unconditionally disable them in the installed headers.
2909 Essentially the same effect can be achieved with the "no-deprecated"
2910 argument to Configure, except that this will always restrict
2911 the build to just the latest API, rather than a fixed API
2914 As applications are ported to future revisions of the API,
2915 they should update their compile-time OPENSSL_API_COMPAT define
2916 accordingly, but in most cases should be able to continue to
2917 compile with later releases.
2919 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2920 0x10000000L and 0x00908000L, respectively. However those
2921 versions did not support the OPENSSL_API_COMPAT feature, and
2922 so applications are not typically tested for explicit support
2923 of just the undeprecated features of either release.
2927 * Add support for setting the minimum and maximum supported protocol.
2928 It can bet set via the SSL_set_min_proto_version() and
2929 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2930 MaxProtocol. It's recommended to use the new APIs to disable
2931 protocols instead of disabling individual protocols using
2932 SSL_set_options() or SSL_CONF's Protocol. This change also
2933 removes support for disabling TLS 1.2 in the OpenSSL TLS
2934 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2938 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2942 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2943 and integrates ECDSA and ECDH functionality into EC. Implementations can
2944 now redirect key generation and no longer need to convert to or from
2947 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2948 include the ec.h header file instead.
2952 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2953 ciphers who are no longer supported and drops support the ephemeral RSA key
2954 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2958 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2959 opaque. For HMAC_CTX, the following constructors and destructors
2962 HMAC_CTX *HMAC_CTX_new(void);
2963 void HMAC_CTX_free(HMAC_CTX *ctx);
2965 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2966 destroy such methods has been added. See EVP_MD_meth_new(3) and
2967 EVP_CIPHER_meth_new(3) for documentation.
2970 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2971 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2972 EVP_MD_CTX_reset() should be called instead to reinitialise
2973 an already created structure.
2974 2) For consistency with the majority of our object creators and
2975 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2976 EVP_MD_CTX_(new|free). The old names are retained as macros
2977 for deprecated builds.
2981 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2982 cryptographic operations to be performed asynchronously as long as an
2983 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2984 further details. Libssl has also had this capability integrated with the
2985 introduction of the new mode SSL_MODE_ASYNC and associated error
2986 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2987 pages. This work was developed in partnership with Intel Corp.
2991 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2992 always enabled now. If you want to disable the support you should
2993 exclude it using the list of supported ciphers. This also means that the
2994 "-no_ecdhe" option has been removed from s_server.
2998 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2999 SSL_{CTX_}set1_curves() which can set a list.
3003 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3004 curve you want to support using SSL_{CTX_}set1_curves().
3008 * State machine rewrite. The state machine code has been significantly
3009 refactored in order to remove much duplication of code and solve issues
3010 with the old code (see ssl/statem/README for further details). This change
3011 does have some associated API changes. Notably the SSL_state() function
3012 has been removed and replaced by SSL_get_state which now returns an
3013 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
3014 altogether. The previous handshake states defined in ssl.h and ssl3.h have
3019 * All instances of the string "ssleay" in the public API were replaced
3020 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3021 Some error codes related to internal RSA_eay API's were renamed.
3025 * The demo files in crypto/threads were moved to demo/threads.
3029 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3032 *Matt Caswell, Rich Salz*
3034 * New ASN.1 embed macro.
3036 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3037 structure is not allocated: it is part of the parent. That is instead of
3045 This reduces memory fragmentation and make it impossible to accidentally
3046 set a mandatory field to NULL.
3048 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3049 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3050 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3055 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3059 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3060 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3061 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3062 DES and RC4 ciphersuites.
3066 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3067 This changes the decoding behaviour for some invalid messages,
3068 though the change is mostly in the more lenient direction, and
3069 legacy behaviour is preserved as much as possible.
3073 * Fix no-stdio build.
3074 * David Woodhouse <David.Woodhouse@intel.com> and also
3075 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3077 * New testing framework
3078 The testing framework has been largely rewritten and is now using
3079 perl and the perl modules Test::Harness and an extended variant of
3080 Test::More called OpenSSL::Test to do its work. All test scripts in
3081 test/ have been rewritten into test recipes, and all direct calls to
3082 executables in test/Makefile have become individual recipes using the
3083 simplified testing OpenSSL::Test::Simple.
3085 For documentation on our testing modules, do:
3087 perldoc test/testlib/OpenSSL/Test/Simple.pm
3088 perldoc test/testlib/OpenSSL/Test.pm
3092 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3093 are used; the latter aborts on memory leaks (usually checked on exit).
3094 Some undocumented "set malloc, etc., hooks" functions were removed
3095 and others were changed. All are now documented.
3099 * In DSA_generate_parameters_ex, if the provided seed is too short,
3102 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3104 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3105 from RFC4279, RFC4785, RFC5487, RFC5489.
3107 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3108 original RSA_PSK patch.
3112 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3113 era flag was never set throughout the codebase (only read). Also removed
3114 SSL3_FLAGS_POP_BUFFER which was only used if
3115 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3119 * Changed the default name options in the "ca", "crl", "req" and "x509"
3120 to be "oneline" instead of "compat".
3124 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3125 not aware of clients that still exhibit this bug, and the workaround
3126 hasn't been working properly for a while.
3130 * The return type of BIO_number_read() and BIO_number_written() as well as
3131 the corresponding num_read and num_write members in the BIO structure has
3132 changed from unsigned long to uint64_t. On platforms where an unsigned
3133 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3138 * Given the pervasive nature of TLS extensions it is inadvisable to run
3139 OpenSSL without support for them. It also means that maintaining
3140 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3141 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3145 * Removed support for the two export grade static DH ciphersuites
3146 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3147 were newly added (along with a number of other static DH ciphersuites) to
3148 1.0.2. However the two export ones have *never* worked since they were
3149 introduced. It seems strange in any case to be adding new export
3150 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3154 * Version negotiation has been rewritten. In particular SSLv23_method(),
3155 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3156 and turned into macros which simply call the new preferred function names
3157 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3158 should use the new names instead. Also as part of this change the ssl23.h
3159 header file has been removed.
3163 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3164 code and the associated standard is no longer considered fit-for-purpose.
3168 * RT2547 was closed. When generating a private key, try to make the
3169 output file readable only by the owner. This behavior change might
3170 be noticeable when interacting with other software.
3172 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3177 * Added HTTP GET support to the ocsp command.
3181 * Changed default digest for the dgst and enc commands from MD5 to
3186 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3190 * Added support for TLS extended master secret from
3191 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3192 initial patch which was a great help during development.
3196 * All libssl internal structures have been removed from the public header
3197 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3198 now redundant). Users should not attempt to access internal structures
3199 directly. Instead they should use the provided API functions.
3203 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3204 Access to deprecated functions can be re-enabled by running config with
3205 "enable-deprecated". In addition applications wishing to use deprecated
3206 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3207 will, by default, disable some transitive includes that previously existed
3208 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3212 * Added support for OCB mode. OpenSSL has been granted a patent license
3213 compatible with the OpenSSL license for use of OCB. Details are available
3214 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3215 for OCB can be removed by calling config with no-ocb.
3219 * SSLv2 support has been removed. It still supports receiving a SSLv2
3220 compatible client hello.
3224 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3225 done while fixing the error code for the key-too-small case.
3227 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3229 * CA.sh has been removed; use CA.pl instead.
3233 * Removed old DES API.
3237 * Remove various unsupported platforms:
3243 Sinix/ReliantUNIX RM400
3248 16-bit platforms such as WIN16
3252 * Clean up OPENSSL_NO_xxx #define's
3253 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3254 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3255 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3256 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3257 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3258 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3259 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3260 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3261 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3262 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3266 * Cleaned up dead code
3267 Remove all but one '#ifdef undef' which is to be looked at.
3271 * Clean up calling of xxx_free routines.
3272 Just like free(), fix most of the xxx_free routines to accept
3273 NULL. Remove the non-null checks from callers. Save much code.
3277 * Add secure heap for storage of private keys (when possible).
3278 Add BIO_s_secmem(), CBIGNUM, etc.
3279 Contributed by Akamai Technologies under our Corporate CLA.
3283 * Experimental support for a new, fast, unbiased prime candidate generator,
3284 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3286 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3288 * New output format NSS in the sess_id command line tool. This allows
3289 exporting the session id and the master key in NSS keylog format.
3291 *Martin Kaiser <martin@kaiser.cx>*
3293 * Harmonize version and its documentation. -f flag is used to display
3296 *mancha <mancha1@zoho.com>*
3298 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3299 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3301 *mancha <mancha1@zoho.com>*
3303 * Fix some double frees. These are not thought to be exploitable.
3305 *mancha <mancha1@zoho.com>*
3307 * A missing bounds check in the handling of the TLS heartbeat extension
3308 can be used to reveal up to 64k of memory to a connected client or
3311 Thanks for Neel Mehta of Google Security for discovering this bug and to
3312 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3313 preparing the fix [CVE-2014-0160][]
3315 *Adam Langley, Bodo Moeller*
3317 * Fix for the attack described in the paper "Recovering OpenSSL
3318 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3319 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3320 <http://eprint.iacr.org/2014/140>
3322 Thanks to Yuval Yarom and Naomi Benger for discovering this
3323 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3325 *Yuval Yarom and Naomi Benger*
3327 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3328 this fixes a limitation in previous versions of OpenSSL.
3332 * Experimental encrypt-then-mac support.
3334 Experimental support for encrypt then mac from
3335 draft-gutmann-tls-encrypt-then-mac-02.txt
3337 To enable it set the appropriate extension number (0x42 for the test
3338 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3340 For non-compliant peers (i.e. just about everything) this should have no
3343 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3347 * Add EVP support for key wrapping algorithms, to avoid problems with
3348 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3349 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3350 algorithms and include tests cases.
3354 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3359 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3360 MGF1 digest and OAEP label.
3364 * Make openssl verify return errors.
3366 *Chris Palmer <palmer@google.com> and Ben Laurie*
3368 * New function ASN1_TIME_diff to calculate the difference between two
3369 ASN1_TIME structures or one structure and the current time.
3373 * Update fips_test_suite to support multiple command line options. New
3374 test to induce all self test errors in sequence and check expected
3379 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3380 sign or verify all in one operation.
3384 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3385 test programs and fips_test_suite. Includes functionality to parse
3386 the minimal script output of fipsalgest.pl directly.
3390 * Add authorisation parameter to FIPS_module_mode_set().
3394 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3398 * Use separate DRBG fields for internal and external flags. New function
3399 FIPS_drbg_health_check() to perform on demand health checking. Add
3400 generation tests to fips_test_suite with reduced health check interval to
3401 demonstrate periodic health checking. Add "nodh" option to
3402 fips_test_suite to skip very slow DH test.
3406 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3411 * More extensive health check for DRBG checking many more failure modes.
3412 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3413 combination: call this in fips_test_suite.
3417 * Add support for canonical generation of DSA parameter 'g'. See
3420 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3421 POST to handle HMAC cases.
3425 * Add functions FIPS_module_version() and FIPS_module_version_text()
3426 to return numerical and string versions of the FIPS module number.
3430 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3431 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3432 outside the validated module in the FIPS capable OpenSSL.
3436 * Minor change to DRBG entropy callback semantics. In some cases
3437 there is no multiple of the block length between min_len and
3438 max_len. Allow the callback to return more than max_len bytes
3439 of entropy but discard any extra: it is the callback's responsibility
3440 to ensure that the extra data discarded does not impact the
3441 requested amount of entropy.
3445 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3446 information in FIPS186-3, SP800-57 and SP800-131A.
3450 * CCM support via EVP. Interface is very similar to GCM case except we
3451 must supply all data in one chunk (i.e. no update, final) and the
3452 message length must be supplied if AAD is used. Add algorithm test
3457 * Initial version of POST overhaul. Add POST callback to allow the status
3458 of POST to be monitored and/or failures induced. Modify fips_test_suite
3459 to use callback. Always run all selftests even if one fails.
3463 * XTS support including algorithm test driver in the fips_gcmtest program.
3464 Note: this does increase the maximum key length from 32 to 64 bytes but
3465 there should be no binary compatibility issues as existing applications
3466 will never use XTS mode.
3470 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3471 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3472 performs algorithm blocking for unapproved PRNG types. Also do not
3473 set PRNG type in FIPS_mode_set(): leave this to the application.
3474 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3475 the standard OpenSSL PRNG: set additional data to a date time vector.
3479 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3480 This shouldn't present any incompatibility problems because applications
3481 shouldn't be using these directly and any that are will need to rethink
3482 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3486 * Extensive self tests and health checking required by SP800-90 DRBG.
3487 Remove strength parameter from FIPS_drbg_instantiate and always
3488 instantiate at maximum supported strength.
3492 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3496 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3500 * New function DH_compute_key_padded() to compute a DH key and pad with
3501 leading zeroes if needed: this complies with SP800-56A et al.
3505 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3506 anything, incomplete, subject to change and largely untested at present.
3510 * Modify fipscanisteronly build option to only build the necessary object
3511 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3515 * Add experimental option FIPSSYMS to give all symbols in
3516 fipscanister.o and FIPS or fips prefix. This will avoid
3517 conflicts with future versions of OpenSSL. Add perl script
3518 util/fipsas.pl to preprocess assembly language source files
3519 and rename any affected symbols.
3523 * Add selftest checks and algorithm block of non-fips algorithms in
3524 FIPS mode. Remove DES2 from selftests.
3528 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3529 return internal method without any ENGINE dependencies. Add new
3530 tiny fips sign and verify functions.
3534 * New build option no-ec2m to disable characteristic 2 code.
3538 * New build option "fipscanisteronly". This only builds fipscanister.o
3539 and (currently) associated fips utilities. Uses the file Makefile.fips
3540 instead of Makefile.org as the prototype.
3544 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3545 Update fips_gcmtest to use IV generator.
3549 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3550 setting output buffer to NULL. The `*Final` function must be
3551 called although it will not retrieve any additional data. The tag
3552 can be set or retrieved with a ctrl. The IV length is by default 12
3553 bytes (96 bits) but can be set to an alternative value. If the IV
3554 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3559 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3560 underlying do_cipher function handles all cipher semantics itself
3561 including padding and finalisation. This is useful if (for example)
3562 an ENGINE cipher handles block padding itself. The behaviour of
3563 do_cipher is subtly changed if this flag is set: the return value
3564 is the number of characters written to the output buffer (zero is
3565 no longer an error code) or a negative error code. Also if the
3566 input buffer is NULL and length 0 finalisation should be performed.
3570 * If a candidate issuer certificate is already part of the constructed
3571 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3575 * Improve forward-security support: add functions
3577 void SSL_CTX_set_not_resumable_session_callback(
3578 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3579 void SSL_set_not_resumable_session_callback(
3580 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3582 for use by SSL/TLS servers; the callback function will be called whenever a
3583 new session is created, and gets to decide whether the session may be
3584 cached to make it resumable (return 0) or not (return 1). (As by the
3585 SSL/TLS protocol specifications, the session_id sent by the server will be
3586 empty to indicate that the session is not resumable; also, the server will
3587 not generate RFC 4507 (RFC 5077) session tickets.)
3589 A simple reasonable callback implementation is to return is_forward_secure.
3590 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3591 by the SSL/TLS server library, indicating whether it can provide forward
3594 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3596 * New -verify_name option in command line utilities to set verification
3601 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3602 Add CMAC pkey methods.
3606 * Experimental renegotiation in s_server -www mode. If the client
3607 browses /reneg connection is renegotiated. If /renegcert it is
3608 renegotiated requesting a certificate.
3612 * Add an "external" session cache for debugging purposes to s_server. This
3613 should help trace issues which normally are only apparent in deployed
3614 multi-process servers.
3618 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3619 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3620 BIO_set_cipher() and some obscure PEM functions were changed so they
3621 can now return an error. The RAND changes required a change to the
3622 RAND_METHOD structure.
3626 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3627 a gcc attribute to warn if the result of a function is ignored. This
3628 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3629 whose return value is often ignored.
3633 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3634 These allow SCTs (signed certificate timestamps) to be requested and
3635 validated when establishing a connection.
3637 *Rob Percival <robpercival@google.com>*
3642 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3644 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3645 used even when parsing explicit parameters, when loading a serialized key
3646 or calling `EC_GROUP_new_from_ecpkparameters()`/
3647 `EC_GROUP_new_from_ecparameters()`.
3648 This prevents bypass of security hardening and performance gains,
3649 especially for curves with specialized EC_METHODs.
3650 By default, if a key encoded with explicit parameters is loaded and later
3651 serialized, the output is still encoded with explicit parameters, even if
3652 internally a "named" EC_GROUP is used for computation.
3656 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3657 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3658 NULL. After this change, only the cofactor parameter can be NULL. It also
3659 does some minimal sanity checks on the passed order.
3664 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3665 An attack is simple, if the first CMS_recipientInfo is valid but the
3666 second CMS_recipientInfo is chosen ciphertext. If the second
3667 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3668 encryption key will be replaced by garbage, and the message cannot be
3669 decoded, but if the RSA decryption fails, the correct encryption key is
3670 used and the recipient will not notice the attack.
3671 As a work around for this potential attack the length of the decrypted
3672 key must be equal to the cipher default key length, in case the
3673 certifiate is not given and all recipientInfo are tried out.
3674 The old behaviour can be re-enabled in the CMS code by setting the
3675 CMS_DEBUG_DECRYPT flag.
3680 * Document issue with installation paths in diverse Windows builds
3682 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3683 binaries and run-time config file.
3688 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3690 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3691 This changes the size when using the genpkey app when no size is given. It
3692 fixes an omission in earlier changes that changed all RSA, DSA and DH
3693 generation apps to use 2048 bits by default.
3697 * Add FIPS support for Android Arm 64-bit
3699 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3700 Module in Version 2.0.10. For some reason, the corresponding target
3701 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3702 built with FIPS support on Android Arm 64-bit. This omission has been
3705 *Matthias St. Pierre*
3707 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3709 * 0-byte record padding oracle
3711 If an application encounters a fatal protocol error and then calls
3712 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3713 then OpenSSL can respond differently to the calling application if a 0 byte
3714 record is received with invalid padding compared to if a 0 byte record is
3715 received with an invalid MAC. If the application then behaves differently
3716 based on that in a way that is detectable to the remote peer, then this
3717 amounts to a padding oracle that could be used to decrypt data.
3719 In order for this to be exploitable "non-stitched" ciphersuites must be in
3720 use. Stitched ciphersuites are optimised implementations of certain
3721 commonly used ciphersuites. Also the application must call SSL_shutdown()
3722 twice even if a protocol error has occurred (applications should not do
3723 this but some do anyway).
3725 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3726 Aviram, with additional investigation by Steven Collison and Andrew
3727 Hourselt. It was reported to OpenSSL on 10th December 2018.
3732 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3736 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3738 * Microarchitecture timing vulnerability in ECC scalar multiplication
3740 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3741 shown to be vulnerable to a microarchitecture timing side channel attack.
3742 An attacker with sufficient access to mount local timing attacks during
3743 ECDSA signature generation could recover the private key.
3745 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3746 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3752 * Timing vulnerability in DSA signature generation
3754 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3755 timing side channel attack. An attacker could use variations in the signing
3756 algorithm to recover the private key.
3758 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3763 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3764 Module, accidentally introduced while backporting security fixes from the
3765 development branch and hindering the use of ECC in FIPS mode.
3769 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3771 * Client DoS due to large DH parameter
3773 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3774 malicious server can send a very large prime value to the client. This will
3775 cause the client to spend an unreasonably long period of time generating a
3776 key for this prime resulting in a hang until the client has finished. This
3777 could be exploited in a Denial Of Service attack.
3779 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3784 * Cache timing vulnerability in RSA Key Generation
3786 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3787 a cache timing side channel attack. An attacker with sufficient access to
3788 mount cache timing attacks during the RSA key generation process could
3789 recover the private key.
3791 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3792 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3797 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3798 parameter is no longer accepted, as it leads to a corrupt table. NULL
3799 pem_str is reserved for alias entries only.
3803 * Revert blinding in ECDSA sign and instead make problematic addition
3804 length-invariant. Switch even to fixed-length Montgomery multiplication.
3808 * Change generating and checking of primes so that the error rate of not
3809 being prime depends on the intended use based on the size of the input.
3810 For larger primes this will result in more rounds of Miller-Rabin.
3811 The maximal error rate for primes with more than 1080 bits is lowered
3814 *Kurt Roeckx, Annie Yousar*
3816 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3820 * Add blinding to ECDSA and DSA signatures to protect against side channel
3821 attacks discovered by Keegan Ryan (NCC Group).
3825 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3826 now allow empty (zero character) pass phrases.
3830 * Certificate time validation (X509_cmp_time) enforces stricter
3831 compliance with RFC 5280. Fractional seconds and timezone offsets
3832 are no longer allowed.
3836 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3838 * Constructed ASN.1 types with a recursive definition could exceed the stack
3840 Constructed ASN.1 types with a recursive definition (such as can be found
3841 in PKCS7) could eventually exceed the stack given malicious input with
3842 excessive recursion. This could result in a Denial Of Service attack. There
3843 are no such structures used within SSL/TLS that come from untrusted sources
3844 so this is considered safe.
3846 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3852 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3854 * Read/write after SSL object in error state
3856 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3857 mechanism. The intent was that if a fatal error occurred during a handshake
3858 then OpenSSL would move into the error state and would immediately fail if
3859 you attempted to continue the handshake. This works as designed for the
3860 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3861 SSL_connect()), however due to a bug it does not work correctly if
3862 SSL_read() or SSL_write() is called directly. In that scenario, if the
3863 handshake fails then a fatal error will be returned in the initial function
3864 call. If SSL_read()/SSL_write() is subsequently called by the application
3865 for the same SSL object then it will succeed and the data is passed without
3866 being decrypted/encrypted directly from the SSL/TLS record layer.
3868 In order to exploit this issue an application bug would have to be present
3869 that resulted in a call to SSL_read()/SSL_write() being issued after having
3870 already received a fatal error.
3872 This issue was reported to OpenSSL by David Benjamin (Google).
3877 * rsaz_1024_mul_avx2 overflow bug on x86_64
3879 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3880 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3881 Analysis suggests that attacks against RSA and DSA as a result of this
3882 defect would be very difficult to perform and are not believed likely.
3883 Attacks against DH1024 are considered just feasible, because most of the
3884 work necessary to deduce information about a private key may be performed
3885 offline. The amount of resources required for such an attack would be
3886 significant. However, for an attack on TLS to be meaningful, the server
3887 would have to share the DH1024 private key among multiple clients, which is
3888 no longer an option since CVE-2016-0701.
3890 This only affects processors that support the AVX2 but not ADX extensions
3891 like Intel Haswell (4th generation).
3893 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3894 was originally found via the OSS-Fuzz project.
3899 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
3901 * bn_sqrx8x_internal carry bug on x86_64
3903 There is a carry propagating bug in the x86_64 Montgomery squaring
3904 procedure. No EC algorithms are affected. Analysis suggests that attacks
3905 against RSA and DSA as a result of this defect would be very difficult to
3906 perform and are not believed likely. Attacks against DH are considered just
3907 feasible (although very difficult) because most of the work necessary to
3908 deduce information about a private key may be performed offline. The amount
3909 of resources required for such an attack would be very significant and
3910 likely only accessible to a limited number of attackers. An attacker would
3911 additionally need online access to an unpatched system using the target
3912 private key in a scenario with persistent DH parameters and a private
3913 key that is shared between multiple clients.
3915 This only affects processors that support the BMI1, BMI2 and ADX extensions
3916 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3918 This issue was reported to OpenSSL by the OSS-Fuzz project.
3923 * Malformed X.509 IPAddressFamily could cause OOB read
3925 If an X.509 certificate has a malformed IPAddressFamily extension,
3926 OpenSSL could do a one-byte buffer overread. The most likely result
3927 would be an erroneous display of the certificate in text format.
3929 This issue was reported to OpenSSL by the OSS-Fuzz project.
3934 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
3936 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3937 platform rather than 'mingw'.
3941 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
3943 * Truncated packet could crash via OOB read
3945 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3946 cipher is being used, then a truncated packet can cause that host to
3947 perform an out-of-bounds read, usually resulting in a crash.
3949 This issue was reported to OpenSSL by Robert Święcki of Google.
3954 * BN_mod_exp may produce incorrect results on x86_64
3956 There is a carry propagating bug in the x86_64 Montgomery squaring
3957 procedure. No EC algorithms are affected. Analysis suggests that attacks
3958 against RSA and DSA as a result of this defect would be very difficult to
3959 perform and are not believed likely. Attacks against DH are considered just
3960 feasible (although very difficult) because most of the work necessary to
3961 deduce information about a private key may be performed offline. The amount
3962 of resources required for such an attack would be very significant and
3963 likely only accessible to a limited number of attackers. An attacker would
3964 additionally need online access to an unpatched system using the target
3965 private key in a scenario with persistent DH parameters and a private
3966 key that is shared between multiple clients. For example this can occur by
3967 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3968 similar to CVE-2015-3193 but must be treated as a separate problem.
3970 This issue was reported to OpenSSL by the OSS-Fuzz project.
3975 * Montgomery multiplication may produce incorrect results
3977 There is a carry propagating bug in the Broadwell-specific Montgomery
3978 multiplication procedure that handles input lengths divisible by, but
3979 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3980 and DH private keys are impossible. This is because the subroutine in
3981 question is not used in operations with the private key itself and an input
3982 of the attacker's direct choice. Otherwise the bug can manifest itself as
3983 transient authentication and key negotiation failures or reproducible
3984 erroneous outcome of public-key operations with specially crafted input.
3985 Among EC algorithms only Brainpool P-512 curves are affected and one
3986 presumably can attack ECDH key negotiation. Impact was not analyzed in
3987 detail, because pre-requisites for attack are considered unlikely. Namely
3988 multiple clients have to choose the curve in question and the server has to
3989 share the private key among them, neither of which is default behaviour.
3990 Even then only clients that chose the curve will be affected.
3992 This issue was publicly reported as transient failures and was not
3993 initially recognized as a security issue. Thanks to Richard Morgan for
3994 providing reproducible case.
3999 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4000 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4001 prevent issues where no progress is being made and the peer continually
4002 sends unrecognised record types, using up resources processing them.
4006 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4008 * Missing CRL sanity check
4010 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4011 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4012 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4014 This issue only affects the OpenSSL 1.0.2i
4019 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4021 * OCSP Status Request extension unbounded memory growth
4023 A malicious client can send an excessively large OCSP Status Request
4024 extension. If that client continually requests renegotiation, sending a
4025 large OCSP Status Request extension each time, then there will be unbounded
4026 memory growth on the server. This will eventually lead to a Denial Of
4027 Service attack through memory exhaustion. Servers with a default
4028 configuration are vulnerable even if they do not support OCSP. Builds using
4029 the "no-ocsp" build time option are not affected.
4031 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4036 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4039 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4045 * OOB write in MDC2_Update()
4047 An overflow can occur in MDC2_Update() either if called directly or
4048 through the EVP_DigestUpdate() function using MDC2. If an attacker
4049 is able to supply very large amounts of input data after a previous
4050 call to EVP_EncryptUpdate() with a partial block then a length check
4051 can overflow resulting in a heap corruption.
4053 The amount of data needed is comparable to SIZE_MAX which is impractical
4056 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4061 * Malformed SHA512 ticket DoS
4063 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4064 DoS attack where a malformed ticket will result in an OOB read which will
4067 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4068 a custom server callback and ticket lookup mechanism.
4070 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4075 * OOB write in BN_bn2dec()
4077 The function BN_bn2dec() does not check the return value of BN_div_word().
4078 This can cause an OOB write if an application uses this function with an
4079 overly large BIGNUM. This could be a problem if an overly large certificate
4080 or CRL is printed out from an untrusted source. TLS is not affected because
4081 record limits will reject an oversized certificate before it is parsed.
4083 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4088 * OOB read in TS_OBJ_print_bio()
4090 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4091 the total length the OID text representation would use and not the amount
4092 of data written. This will result in OOB reads when large OIDs are
4095 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4100 * Pointer arithmetic undefined behaviour
4102 Avoid some undefined pointer arithmetic
4104 A common idiom in the codebase is to check limits in the following manner:
4107 Where "p" points to some malloc'd data of SIZE bytes and
4110 "len" here could be from some externally supplied data (e.g. from a TLS
4113 The rules of C pointer arithmetic are such that "p + len" is only well
4114 defined where len <= SIZE. Therefore the above idiom is actually
4115 undefined behaviour.
4117 For example this could cause problems if some malloc implementation
4118 provides an address for "p" such that "p + len" actually overflows for
4119 values of len that are too big and therefore p + len < limit.
4121 This issue was reported to OpenSSL by Guido Vranken
4126 * Constant time flag not preserved in DSA signing
4128 Operations in the DSA signing algorithm should run in constant time in
4129 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4130 implementation means that a non-constant time codepath is followed for
4131 certain operations. This has been demonstrated through a cache-timing
4132 attack to be sufficient for an attacker to recover the private DSA key.
4134 This issue was reported by César Pereida (Aalto University), Billy Brumley
4135 (Tampere University of Technology), and Yuval Yarom (The University of
4136 Adelaide and NICTA).
4141 * DTLS buffered message DoS
4143 In a DTLS connection where handshake messages are delivered out-of-order
4144 those messages that OpenSSL is not yet ready to process will be buffered
4145 for later use. Under certain circumstances, a flaw in the logic means that
4146 those messages do not get removed from the buffer even though the handshake
4147 has been completed. An attacker could force up to approx. 15 messages to
4148 remain in the buffer when they are no longer required. These messages will
4149 be cleared when the DTLS connection is closed. The default maximum size for
4150 a message is 100k. Therefore the attacker could force an additional 1500k
4151 to be consumed per connection. By opening many simulataneous connections an
4152 attacker could cause a DoS attack through memory exhaustion.
4154 This issue was reported to OpenSSL by Quan Luo.
4159 * DTLS replay protection DoS
4161 A flaw in the DTLS replay attack protection mechanism means that records
4162 that arrive for future epochs update the replay protection "window" before
4163 the MAC for the record has been validated. This could be exploited by an
4164 attacker by sending a record for the next epoch (which does not have to
4165 decrypt or have a valid MAC), with a very large sequence number. This means
4166 that all subsequent legitimate packets are dropped causing a denial of
4167 service for a specific DTLS connection.
4169 This issue was reported to OpenSSL by the OCAP audit team.
4174 * Certificate message OOB reads
4176 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4177 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4178 theoretical DoS risk but this has not been observed in practice on common
4181 The messages affected are client certificate, client certificate request
4182 and server certificate. As a result the attack can only be performed
4183 against a client or a server which enables client authentication.
4185 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4190 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4192 * Prevent padding oracle in AES-NI CBC MAC check
4194 A MITM attacker can use a padding oracle attack to decrypt traffic
4195 when the connection uses an AES CBC cipher and the server support
4198 This issue was introduced as part of the fix for Lucky 13 padding
4199 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4200 constant time by making sure that always the same bytes are read and
4201 compared against either the MAC or padding bytes. But it no longer
4202 checked that there was enough data to have both the MAC and padding
4205 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4210 * Fix EVP_EncodeUpdate overflow
4212 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4213 Base64 encoding of binary data. If an attacker is able to supply very large
4214 amounts of input data then a length check can overflow resulting in a heap
4217 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4218 the `PEM_write_bio*` family of functions. These are mainly used within the
4219 OpenSSL command line applications, so any application which processes data
4220 from an untrusted source and outputs it as a PEM file should be considered
4221 vulnerable to this issue. User applications that call these APIs directly
4222 with large amounts of untrusted data may also be vulnerable.
4224 This issue was reported by Guido Vranken.
4229 * Fix EVP_EncryptUpdate overflow
4231 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4232 is able to supply very large amounts of input data after a previous call to
4233 EVP_EncryptUpdate() with a partial block then a length check can overflow
4234 resulting in a heap corruption. Following an analysis of all OpenSSL
4235 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4236 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4237 the first called function after an EVP_EncryptInit(), and therefore that
4238 specific call must be safe. The second form is where the length passed to
4239 EVP_EncryptUpdate() can be seen from the code to be some small value and
4240 therefore there is no possibility of an overflow. Since all instances are
4241 one of these two forms, it is believed that there can be no overflows in
4242 internal code due to this problem. It should be noted that
4243 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4244 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4245 of these calls have also been analysed too and it is believed there are no
4246 instances in internal usage where an overflow could occur.
4248 This issue was reported by Guido Vranken.
4253 * Prevent ASN.1 BIO excessive memory allocation
4255 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4256 a short invalid encoding can cause allocation of large amounts of memory
4257 potentially consuming excessive resources or exhausting memory.
4259 Any application parsing untrusted data through d2i BIO functions is
4260 affected. The memory based functions such as d2i_X509() are *not* affected.
4261 Since the memory based functions are used by the TLS library, TLS
4262 applications are not affected.
4264 This issue was reported by Brian Carpenter.
4271 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4272 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4273 in arbitrary stack data being returned in the buffer.
4275 This issue was reported by Guido Vranken.
4280 * Modify behavior of ALPN to invoke callback after SNI/servername
4281 callback, such that updates to the SSL_CTX affect ALPN.
4285 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4290 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4291 methods are enabled and ssl2 is disabled the methods return NULL.
4295 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4297 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4298 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4299 provide any "EXPORT" or "LOW" strength ciphers.
4303 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4304 is by default disabled at build-time. Builds that are not configured with
4305 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4306 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4307 will need to explicitly call either of:
4309 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4311 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4313 as appropriate. Even if either of those is used, or the application
4314 explicitly uses the version-specific SSLv2_method() or its client and
4315 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4316 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4317 ciphers, and SSLv2 56-bit DES are no longer available.
4322 * Fix a double-free in DSA code
4324 A double free bug was discovered when OpenSSL parses malformed DSA private
4325 keys and could lead to a DoS attack or memory corruption for applications
4326 that receive DSA private keys from untrusted sources. This scenario is
4329 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4335 * Disable SRP fake user seed to address a server memory leak.
4337 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4339 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4340 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4341 was changed to ignore the "fake user" SRP seed, even if the seed
4344 Users should use SRP_VBASE_get1_by_user instead. Note that in
4345 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4346 also that even though configuring the SRP seed attempts to hide
4347 invalid usernames by continuing the handshake with fake
4348 credentials, this behaviour is not constant time and no strong
4349 guarantees are made that the handshake is indistinguishable from
4350 that of a valid user.
4355 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4357 In the BN_hex2bn function the number of hex digits is calculated using an
4358 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4359 large values of |i| this can result in |bn_expand| not allocating any
4360 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4361 field as NULL leading to a subsequent NULL ptr deref. For very large values
4362 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4363 In this case memory is allocated to the internal BIGNUM data field, but it
4364 is insufficiently sized leading to heap corruption. A similar issue exists
4365 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4366 is ever called by user applications with very large untrusted hex/dec data.
4367 This is anticipated to be a rare occurrence.
4369 All OpenSSL internal usage of these functions use data that is not expected
4370 to be untrusted, e.g. config file data or application command line
4371 arguments. If user developed applications generate config file data based
4372 on untrusted data then it is possible that this could also lead to security
4373 consequences. This is also anticipated to be rare.
4375 This issue was reported to OpenSSL by Guido Vranken.
4380 * Fix memory issues in `BIO_*printf` functions
4382 The internal |fmtstr| function used in processing a "%s" format string in
4383 the `BIO_*printf` functions could overflow while calculating the length of a
4384 string and cause an OOB read when printing very long strings.
4386 Additionally the internal |doapr_outch| function can attempt to write to an
4387 OOB memory location (at an offset from the NULL pointer) in the event of a
4388 memory allocation failure. In 1.0.2 and below this could be caused where
4389 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4390 could be in processing a very long "%s" format string. Memory leaks can
4393 The first issue may mask the second issue dependent on compiler behaviour.
4394 These problems could enable attacks where large amounts of untrusted data
4395 is passed to the `BIO_*printf` functions. If applications use these functions
4396 in this way then they could be vulnerable. OpenSSL itself uses these
4397 functions when printing out human-readable dumps of ASN.1 data. Therefore
4398 applications that print this data could be vulnerable if the data is from
4399 untrusted sources. OpenSSL command line applications could also be
4400 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4401 as command line arguments.
4403 Libssl is not considered directly vulnerable. Additionally certificates etc
4404 received via remote connections via libssl are also unlikely to be able to
4405 trigger these issues because of message size limits enforced within libssl.
4407 This issue was reported to OpenSSL Guido Vranken.
4412 * Side channel attack on modular exponentiation
4414 A side-channel attack was found which makes use of cache-bank conflicts on
4415 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4416 of RSA keys. The ability to exploit this issue is limited as it relies on
4417 an attacker who has control of code in a thread running on the same
4418 hyper-threaded core as the victim thread which is performing decryptions.
4420 This issue was reported to OpenSSL by Yuval Yarom, The University of
4421 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4422 Nadia Heninger, University of Pennsylvania with more information at
4423 <http://cachebleed.info>.
4428 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4429 if no keysize is specified with default_bits. This fixes an
4430 omission in an earlier change that changed all RSA/DSA key generation
4431 apps to use 2048 bits by default.
4435 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4437 * DH small subgroups
4439 Historically OpenSSL only ever generated DH parameters based on "safe"
4440 primes. More recently (in version 1.0.2) support was provided for
4441 generating X9.42 style parameter files such as those required for RFC 5114
4442 support. The primes used in such files may not be "safe". Where an
4443 application is using DH configured with parameters based on primes that are
4444 not "safe" then an attacker could use this fact to find a peer's private
4445 DH exponent. This attack requires that the attacker complete multiple
4446 handshakes in which the peer uses the same private DH exponent. For example
4447 this could be used to discover a TLS server's private DH exponent if it's
4448 reusing the private DH exponent or it's using a static DH ciphersuite.
4450 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4451 TLS. It is not on by default. If the option is not set then the server
4452 reuses the same private DH exponent for the life of the server process and
4453 would be vulnerable to this attack. It is believed that many popular
4454 applications do set this option and would therefore not be at risk.
4456 The fix for this issue adds an additional check where a "q" parameter is
4457 available (as is the case in X9.42 based parameters). This detects the
4458 only known attack, and is the only possible defense for static DH
4459 ciphersuites. This could have some performance impact.
4461 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4462 default and cannot be disabled. This could have some performance impact.
4464 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4469 * SSLv2 doesn't block disabled ciphers
4471 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4472 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4473 been disabled, provided that the SSLv2 protocol was not also disabled via
4476 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4477 and Sebastian Schinzel.
4482 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4484 * BN_mod_exp may produce incorrect results on x86_64
4486 There is a carry propagating bug in the x86_64 Montgomery squaring
4487 procedure. No EC algorithms are affected. Analysis suggests that attacks
4488 against RSA and DSA as a result of this defect would be very difficult to
4489 perform and are not believed likely. Attacks against DH are considered just
4490 feasible (although very difficult) because most of the work necessary to
4491 deduce information about a private key may be performed offline. The amount
4492 of resources required for such an attack would be very significant and
4493 likely only accessible to a limited number of attackers. An attacker would
4494 additionally need online access to an unpatched system using the target
4495 private key in a scenario with persistent DH parameters and a private
4496 key that is shared between multiple clients. For example this can occur by
4497 default in OpenSSL DHE based SSL/TLS ciphersuites.
4499 This issue was reported to OpenSSL by Hanno Böck.
4504 * Certificate verify crash with missing PSS parameter
4506 The signature verification routines will crash with a NULL pointer
4507 dereference if presented with an ASN.1 signature using the RSA PSS
4508 algorithm and absent mask generation function parameter. Since these
4509 routines are used to verify certificate signature algorithms this can be
4510 used to crash any certificate verification operation and exploited in a
4511 DoS attack. Any application which performs certificate verification is
4512 vulnerable including OpenSSL clients and servers which enable client
4515 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4520 * X509_ATTRIBUTE memory leak
4522 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4523 memory. This structure is used by the PKCS#7 and CMS routines so any
4524 application which reads PKCS#7 or CMS data from untrusted sources is
4525 affected. SSL/TLS is not affected.
4527 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4533 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4534 This changes the decoding behaviour for some invalid messages,
4535 though the change is mostly in the more lenient direction, and
4536 legacy behaviour is preserved as much as possible.
4540 * In DSA_generate_parameters_ex, if the provided seed is too short,
4543 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4545 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4547 * Alternate chains certificate forgery
4549 During certificate verification, OpenSSL will attempt to find an
4550 alternative certificate chain if the first attempt to build such a chain
4551 fails. An error in the implementation of this logic can mean that an
4552 attacker could cause certain checks on untrusted certificates to be
4553 bypassed, such as the CA flag, enabling them to use a valid leaf
4554 certificate to act as a CA and "issue" an invalid certificate.
4556 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4561 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4563 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4564 incompatibility in the handling of HMAC. The previous ABI has now been
4569 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4571 * Malformed ECParameters causes infinite loop
4573 When processing an ECParameters structure OpenSSL enters an infinite loop
4574 if the curve specified is over a specially malformed binary polynomial
4577 This can be used to perform denial of service against any
4578 system which processes public keys, certificate requests or
4579 certificates. This includes TLS clients and TLS servers with
4580 client authentication enabled.
4582 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4587 * Exploitable out-of-bounds read in X509_cmp_time
4589 X509_cmp_time does not properly check the length of the ASN1_TIME
4590 string and can read a few bytes out of bounds. In addition,
4591 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4594 An attacker can use this to craft malformed certificates and CRLs of
4595 various sizes and potentially cause a segmentation fault, resulting in
4596 a DoS on applications that verify certificates or CRLs. TLS clients
4597 that verify CRLs are affected. TLS clients and servers with client
4598 authentication enabled may be affected if they use custom verification
4601 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4602 independently by Hanno Böck.
4607 * PKCS7 crash with missing EnvelopedContent
4609 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4610 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4611 with missing content and trigger a NULL pointer dereference on parsing.
4613 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4614 structures from untrusted sources are affected. OpenSSL clients and
4615 servers are not affected.
4617 This issue was reported to OpenSSL by Michal Zalewski (Google).
4622 * CMS verify infinite loop with unknown hash function
4624 When verifying a signedData message the CMS code can enter an infinite loop
4625 if presented with an unknown hash function OID. This can be used to perform
4626 denial of service against any system which verifies signedData messages using
4628 This issue was reported to OpenSSL by Johannes Bauer.
4633 * Race condition handling NewSessionTicket
4635 If a NewSessionTicket is received by a multi-threaded client when attempting to
4636 reuse a previous ticket then a race condition can occur potentially leading to
4637 a double free of the ticket data.
4642 * Only support 256-bit or stronger elliptic curves with the
4643 'ecdh_auto' setting (server) or by default (client). Of supported
4644 curves, prefer P-256 (both).
4648 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4650 * ClientHello sigalgs DoS fix
4652 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4653 invalid signature algorithms extension a NULL pointer dereference will
4654 occur. This can be exploited in a DoS attack against the server.
4656 This issue was was reported to OpenSSL by David Ramos of Stanford
4660 *Stephen Henson and Matt Caswell*
4662 * Multiblock corrupted pointer fix
4664 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4665 feature only applies on 64 bit x86 architecture platforms that support AES
4666 NI instructions. A defect in the implementation of "multiblock" can cause
4667 OpenSSL's internal write buffer to become incorrectly set to NULL when
4668 using non-blocking IO. Typically, when the user application is using a
4669 socket BIO for writing, this will only result in a failed connection.
4670 However if some other BIO is used then it is likely that a segmentation
4671 fault will be triggered, thus enabling a potential DoS attack.
4673 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4678 * Segmentation fault in DTLSv1_listen fix
4680 The DTLSv1_listen function is intended to be stateless and processes the
4681 initial ClientHello from many peers. It is common for user code to loop
4682 over the call to DTLSv1_listen until a valid ClientHello is received with
4683 an associated cookie. A defect in the implementation of DTLSv1_listen means
4684 that state is preserved in the SSL object from one invocation to the next
4685 that can lead to a segmentation fault. Errors processing the initial
4686 ClientHello can trigger this scenario. An example of such an error could be
4687 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4690 This issue was reported to OpenSSL by Per Allansson.
4695 * Segmentation fault in ASN1_TYPE_cmp fix
4697 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4698 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4699 certificate signature algorithm consistency this can be used to crash any
4700 certificate verification operation and exploited in a DoS attack. Any
4701 application which performs certificate verification is vulnerable including
4702 OpenSSL clients and servers which enable client authentication.
4707 * Segmentation fault for invalid PSS parameters fix
4709 The signature verification routines will crash with a NULL pointer
4710 dereference if presented with an ASN.1 signature using the RSA PSS
4711 algorithm and invalid parameters. Since these routines are used to verify
4712 certificate signature algorithms this can be used to crash any
4713 certificate verification operation and exploited in a DoS attack. Any
4714 application which performs certificate verification is vulnerable including
4715 OpenSSL clients and servers which enable client authentication.
4717 This issue was was reported to OpenSSL by Brian Carpenter.
4722 * ASN.1 structure reuse memory corruption fix
4724 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4725 memory corruption via an invalid write. Such reuse is and has been
4726 strongly discouraged and is believed to be rare.
4728 Applications that parse structures containing CHOICE or ANY DEFINED BY
4729 components may be affected. Certificate parsing (d2i_X509 and related
4730 functions) are however not affected. OpenSSL clients and servers are
4736 * PKCS7 NULL pointer dereferences fix
4738 The PKCS#7 parsing code does not handle missing outer ContentInfo
4739 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4740 missing content and trigger a NULL pointer dereference on parsing.
4742 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4743 otherwise parse PKCS#7 structures from untrusted sources are
4744 affected. OpenSSL clients and servers are not affected.
4746 This issue was reported to OpenSSL by Michal Zalewski (Google).
4751 * DoS via reachable assert in SSLv2 servers fix
4753 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4754 servers that both support SSLv2 and enable export cipher suites by sending
4755 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4757 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4758 (OpenSSL development team).
4763 * Empty CKE with client auth and DHE fix
4765 If client auth is used then a server can seg fault in the event of a DHE
4766 ciphersuite being selected and a zero length ClientKeyExchange message
4767 being sent by the client. This could be exploited in a DoS attack.
4772 * Handshake with unseeded PRNG fix
4774 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4775 with an unseeded PRNG. The conditions are:
4776 - The client is on a platform where the PRNG has not been seeded
4777 automatically, and the user has not seeded manually
4778 - A protocol specific client method version has been used (i.e. not
4779 SSL_client_methodv23)
4780 - A ciphersuite is used that does not require additional random data from
4781 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4783 If the handshake succeeds then the client random that has been used will
4784 have been generated from a PRNG with insufficient entropy and therefore the
4785 output may be predictable.
4787 For example using the following command with an unseeded openssl will
4788 succeed on an unpatched platform:
4790 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4795 * Use After Free following d2i_ECPrivatekey error fix
4797 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4798 could cause a use after free condition. This, in turn, could cause a double
4799 free in several private key parsing functions (such as d2i_PrivateKey
4800 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4801 for applications that receive EC private keys from untrusted
4802 sources. This scenario is considered rare.
4804 This issue was discovered by the BoringSSL project and fixed in their
4810 * X509_to_X509_REQ NULL pointer deref fix
4812 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4813 the certificate key is invalid. This function is rarely used in practice.
4815 This issue was discovered by Brian Carpenter.
4820 * Removed the export ciphers from the DEFAULT ciphers
4824 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4826 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4827 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4828 So far those who have to target multiple platforms would compromise
4829 and argue that binary targeting say ARMv5 would still execute on
4830 ARMv8. "Universal" build resolves this compromise by providing
4831 near-optimal performance even on newer platforms.
4835 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4836 (other platforms pending).
4838 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4840 * Add support for the SignedCertificateTimestampList certificate and
4841 OCSP response extensions from RFC6962.
4845 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4846 for corner cases. (Certain input points at infinity could lead to
4847 bogus results, with non-infinity inputs mapped to infinity too.)
4851 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4852 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4853 common cases are optimized and there still is room for further
4854 improvements. Vector Permutation AES for Altivec is also added.
4858 * Add support for little-endian ppc64 Linux target.
4860 *Marcelo Cerri (IBM)*
4862 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4863 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4864 are optimized and there still is room for further improvements.
4865 Both 32- and 64-bit modes are supported.
4867 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4869 * Improved ARMv7 NEON support.
4873 * Support for SPARC Architecture 2011 crypto extensions, first
4874 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4875 SHA256/512, MD5, GHASH and modular exponentiation.
4877 *Andy Polyakov, David Miller*
4879 * Accelerated modular exponentiation for Intel processors, a.k.a.
4882 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4884 * Support for new and upcoming Intel processors, including AVX2,
4885 BMI and SHA ISA extensions. This includes additional "stitched"
4886 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4889 This work was sponsored by Intel Corp.
4893 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4894 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4895 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4899 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4900 this fixes a limitation in previous versions of OpenSSL.
4904 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4905 MGF1 digest and OAEP label.
4909 * Add EVP support for key wrapping algorithms, to avoid problems with
4910 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4911 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4912 algorithms and include tests cases.
4916 * Add functions to allocate and set the fields of an ECDSA_METHOD
4919 *Douglas E. Engert, Steve Henson*
4921 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4922 difference in days and seconds between two tm or ASN1_TIME structures.
4926 * Add -rev test option to s_server to just reverse order of characters
4927 received by client and send back to server. Also prints an abbreviated
4928 summary of the connection parameters.
4932 * New option -brief for s_client and s_server to print out a brief summary
4933 of connection parameters.
4937 * Add callbacks for arbitrary TLS extensions.
4939 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4941 * New option -crl_download in several openssl utilities to download CRLs
4942 from CRLDP extension in certificates.
4946 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4950 * New function X509_CRL_diff to generate a delta CRL from the difference
4951 of two full CRLs. Add support to "crl" utility.
4955 * New functions to set lookup_crls function and to retrieve
4956 X509_STORE from X509_STORE_CTX.
4960 * Print out deprecated issuer and subject unique ID fields in
4965 * Extend OCSP I/O functions so they can be used for simple general purpose
4966 HTTP as well as OCSP. New wrapper function which can be used to download
4967 CRLs using the OCSP API.
4971 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4975 * `SSL_CONF*` functions. These provide a common framework for application
4976 configuration using configuration files or command lines.
4980 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4981 message callback and prints the results. Needs compile time option
4982 "enable-ssl-trace". New options to s_client and s_server to enable
4987 * New ctrl and macro to retrieve supported points extensions.
4988 Print out extension in s_server and s_client.
4992 * New functions to retrieve certificate signature and signature
4997 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5002 * New Suite B modes for TLS code. These use and enforce the requirements
5003 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5004 only use Suite B curves. The Suite B modes can be set by using the
5005 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5009 * New chain verification flags for Suite B levels of security. Check
5010 algorithms are acceptable when flags are set in X509_verify_cert.
5014 * Make tls1_check_chain return a set of flags indicating checks passed
5015 by a certificate chain. Add additional tests to handle client
5016 certificates: checks for matching certificate type and issuer name
5021 * If an attempt is made to use a signature algorithm not in the peer
5022 preference list abort the handshake. If client has no suitable
5023 signature algorithms in response to a certificate request do not
5024 use the certificate.
5028 * If server EC tmp key is not in client preference list abort handshake.
5032 * Add support for certificate stores in CERT structure. This makes it
5033 possible to have different stores per SSL structure or one store in
5034 the parent SSL_CTX. Include distinct stores for certificate chain
5035 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5036 to build and store a certificate chain in CERT structure: returning
5037 an error if the chain cannot be built: this will allow applications
5038 to test if a chain is correctly configured.
5040 Note: if the CERT based stores are not set then the parent SSL_CTX
5041 store is used to retain compatibility with existing behaviour.
5045 * New function ssl_set_client_disabled to set a ciphersuite disabled
5046 mask based on the current session, check mask when sending client
5047 hello and checking the requested ciphersuite.
5051 * New ctrls to retrieve and set certificate types in a certificate
5052 request message. Print out received values in s_client. If certificate
5053 types is not set with custom values set sensible values based on
5054 supported signature algorithms.
5058 * Support for distinct client and server supported signature algorithms.
5062 * Add certificate callback. If set this is called whenever a certificate
5063 is required by client or server. An application can decide which
5064 certificate chain to present based on arbitrary criteria: for example
5065 supported signature algorithms. Add very simple example to s_server.
5066 This fixes many of the problems and restrictions of the existing client
5067 certificate callback: for example you can now clear an existing
5068 certificate and specify the whole chain.
5072 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5073 the certificate can be used for (if anything). Set valid_flags field
5074 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5075 to have similar checks in it.
5077 Add new "cert_flags" field to CERT structure and include a "strict mode".
5078 This enforces some TLS certificate requirements (such as only permitting
5079 certificate signature algorithms contained in the supported algorithms
5080 extension) which some implementations ignore: this option should be used
5081 with caution as it could cause interoperability issues.
5085 * Update and tidy signature algorithm extension processing. Work out
5086 shared signature algorithms based on preferences and peer algorithms
5087 and print them out in s_client and s_server. Abort handshake if no
5088 shared signature algorithms.
5092 * Add new functions to allow customised supported signature algorithms
5093 for SSL and SSL_CTX structures. Add options to s_client and s_server
5098 * New function SSL_certs_clear() to delete all references to certificates
5099 from an SSL structure. Before this once a certificate had been added
5100 it couldn't be removed.
5104 * Integrate hostname, email address and IP address checking with certificate
5105 verification. New verify options supporting checking in openssl utility.
5109 * Fixes and wildcard matching support to hostname and email checking
5110 functions. Add manual page.
5112 *Florian Weimer (Red Hat Product Security Team)*
5114 * New functions to check a hostname email or IP address against a
5115 certificate. Add options x509 utility to print results of checks against
5120 * Fix OCSP checking.
5122 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5124 * Initial experimental support for explicitly trusted non-root CAs.
5125 OpenSSL still tries to build a complete chain to a root but if an
5126 intermediate CA has a trust setting included that is used. The first
5127 setting is used: whether to trust (e.g., -addtrust option to the x509
5132 * Add -trusted_first option which attempts to find certificates in the
5133 trusted store even if an untrusted chain is also supplied.
5137 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5138 platform support for Linux and Android.
5142 * Support for linux-x32, ILP32 environment in x86_64 framework.
5146 * Experimental multi-implementation support for FIPS capable OpenSSL.
5147 When in FIPS mode the approved implementations are used as normal,
5148 when not in FIPS mode the internal unapproved versions are used instead.
5149 This means that the FIPS capable OpenSSL isn't forced to use the
5150 (often lower performance) FIPS implementations outside FIPS mode.
5154 * Transparently support X9.42 DH parameters when calling
5155 PEM_read_bio_DHparameters. This means existing applications can handle
5156 the new parameter format automatically.
5160 * Initial experimental support for X9.42 DH parameter format: mainly
5161 to support use of 'q' parameter for RFC5114 parameters.
5165 * Add DH parameters from RFC5114 including test data to dhtest.
5169 * Support for automatic EC temporary key parameter selection. If enabled
5170 the most preferred EC parameters are automatically used instead of
5171 hardcoded fixed parameters. Now a server just has to call:
5172 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5173 support ECDH and use the most appropriate parameters.
5177 * Enhance and tidy EC curve and point format TLS extension code. Use
5178 static structures instead of allocation if default values are used.
5179 New ctrls to set curves we wish to support and to retrieve shared curves.
5180 Print out shared curves in s_server. New options to s_server and s_client
5181 to set list of supported curves.
5185 * New ctrls to retrieve supported signature algorithms and
5186 supported curve values as an array of NIDs. Extend openssl utility
5187 to print out received values.
5191 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5192 between NIDs and the more common NIST names such as "P-256". Enhance
5193 ecparam utility and ECC method to recognise the NIST names for curves.
5197 * Enhance SSL/TLS certificate chain handling to support different
5198 chains for each certificate instead of one chain in the parent SSL_CTX.
5202 * Support for fixed DH ciphersuite client authentication: where both
5203 server and client use DH certificates with common parameters.
5207 * Support for fixed DH ciphersuites: those requiring DH server
5212 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5214 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5215 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5216 X509_CINF_get_signature were reverted post internal team review.
5221 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5223 * OCSP Status Request extension unbounded memory growth
5225 A malicious client can send an excessively large OCSP Status Request
5226 extension. If that client continually requests renegotiation, sending a
5227 large OCSP Status Request extension each time, then there will be unbounded
5228 memory growth on the server. This will eventually lead to a Denial Of
5229 Service attack through memory exhaustion. Servers with a default
5230 configuration are vulnerable even if they do not support OCSP. Builds using
5231 the "no-ocsp" build time option are not affected.
5233 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5238 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5241 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5247 * OOB write in MDC2_Update()
5249 An overflow can occur in MDC2_Update() either if called directly or
5250 through the EVP_DigestUpdate() function using MDC2. If an attacker
5251 is able to supply very large amounts of input data after a previous
5252 call to EVP_EncryptUpdate() with a partial block then a length check
5253 can overflow resulting in a heap corruption.
5255 The amount of data needed is comparable to SIZE_MAX which is impractical
5258 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5263 * Malformed SHA512 ticket DoS
5265 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5266 DoS attack where a malformed ticket will result in an OOB read which will
5269 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5270 a custom server callback and ticket lookup mechanism.
5272 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5277 * OOB write in BN_bn2dec()
5279 The function BN_bn2dec() does not check the return value of BN_div_word().
5280 This can cause an OOB write if an application uses this function with an
5281 overly large BIGNUM. This could be a problem if an overly large certificate
5282 or CRL is printed out from an untrusted source. TLS is not affected because
5283 record limits will reject an oversized certificate before it is parsed.
5285 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5290 * OOB read in TS_OBJ_print_bio()
5292 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5293 the total length the OID text representation would use and not the amount
5294 of data written. This will result in OOB reads when large OIDs are
5297 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5302 * Pointer arithmetic undefined behaviour
5304 Avoid some undefined pointer arithmetic
5306 A common idiom in the codebase is to check limits in the following manner:
5309 Where "p" points to some malloc'd data of SIZE bytes and
5312 "len" here could be from some externally supplied data (e.g. from a TLS
5315 The rules of C pointer arithmetic are such that "p + len" is only well
5316 defined where len <= SIZE. Therefore the above idiom is actually
5317 undefined behaviour.
5319 For example this could cause problems if some malloc implementation
5320 provides an address for "p" such that "p + len" actually overflows for
5321 values of len that are too big and therefore p + len < limit.
5323 This issue was reported to OpenSSL by Guido Vranken
5328 * Constant time flag not preserved in DSA signing
5330 Operations in the DSA signing algorithm should run in constant time in
5331 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5332 implementation means that a non-constant time codepath is followed for
5333 certain operations. This has been demonstrated through a cache-timing
5334 attack to be sufficient for an attacker to recover the private DSA key.
5336 This issue was reported by César Pereida (Aalto University), Billy Brumley
5337 (Tampere University of Technology), and Yuval Yarom (The University of
5338 Adelaide and NICTA).
5343 * DTLS buffered message DoS
5345 In a DTLS connection where handshake messages are delivered out-of-order
5346 those messages that OpenSSL is not yet ready to process will be buffered
5347 for later use. Under certain circumstances, a flaw in the logic means that
5348 those messages do not get removed from the buffer even though the handshake
5349 has been completed. An attacker could force up to approx. 15 messages to
5350 remain in the buffer when they are no longer required. These messages will
5351 be cleared when the DTLS connection is closed. The default maximum size for
5352 a message is 100k. Therefore the attacker could force an additional 1500k
5353 to be consumed per connection. By opening many simulataneous connections an
5354 attacker could cause a DoS attack through memory exhaustion.
5356 This issue was reported to OpenSSL by Quan Luo.
5361 * DTLS replay protection DoS
5363 A flaw in the DTLS replay attack protection mechanism means that records
5364 that arrive for future epochs update the replay protection "window" before
5365 the MAC for the record has been validated. This could be exploited by an
5366 attacker by sending a record for the next epoch (which does not have to
5367 decrypt or have a valid MAC), with a very large sequence number. This means
5368 that all subsequent legitimate packets are dropped causing a denial of
5369 service for a specific DTLS connection.
5371 This issue was reported to OpenSSL by the OCAP audit team.
5376 * Certificate message OOB reads
5378 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5379 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5380 theoretical DoS risk but this has not been observed in practice on common
5383 The messages affected are client certificate, client certificate request
5384 and server certificate. As a result the attack can only be performed
5385 against a client or a server which enables client authentication.
5387 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5392 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5394 * Prevent padding oracle in AES-NI CBC MAC check
5396 A MITM attacker can use a padding oracle attack to decrypt traffic
5397 when the connection uses an AES CBC cipher and the server support
5400 This issue was introduced as part of the fix for Lucky 13 padding
5401 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5402 constant time by making sure that always the same bytes are read and
5403 compared against either the MAC or padding bytes. But it no longer
5404 checked that there was enough data to have both the MAC and padding
5407 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5412 * Fix EVP_EncodeUpdate overflow
5414 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5415 Base64 encoding of binary data. If an attacker is able to supply very large
5416 amounts of input data then a length check can overflow resulting in a heap
5419 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5420 the `PEM_write_bio*` family of functions. These are mainly used within the
5421 OpenSSL command line applications, so any application which processes data
5422 from an untrusted source and outputs it as a PEM file should be considered
5423 vulnerable to this issue. User applications that call these APIs directly
5424 with large amounts of untrusted data may also be vulnerable.
5426 This issue was reported by Guido Vranken.
5431 * Fix EVP_EncryptUpdate overflow
5433 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5434 is able to supply very large amounts of input data after a previous call to
5435 EVP_EncryptUpdate() with a partial block then a length check can overflow
5436 resulting in a heap corruption. Following an analysis of all OpenSSL
5437 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5438 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5439 the first called function after an EVP_EncryptInit(), and therefore that
5440 specific call must be safe. The second form is where the length passed to
5441 EVP_EncryptUpdate() can be seen from the code to be some small value and
5442 therefore there is no possibility of an overflow. Since all instances are
5443 one of these two forms, it is believed that there can be no overflows in
5444 internal code due to this problem. It should be noted that
5445 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5446 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5447 of these calls have also been analysed too and it is believed there are no
5448 instances in internal usage where an overflow could occur.
5450 This issue was reported by Guido Vranken.
5455 * Prevent ASN.1 BIO excessive memory allocation
5457 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5458 a short invalid encoding can casuse allocation of large amounts of memory
5459 potentially consuming excessive resources or exhausting memory.
5461 Any application parsing untrusted data through d2i BIO functions is
5462 affected. The memory based functions such as d2i_X509() are *not* affected.
5463 Since the memory based functions are used by the TLS library, TLS
5464 applications are not affected.
5466 This issue was reported by Brian Carpenter.
5473 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5474 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5475 in arbitrary stack data being returned in the buffer.
5477 This issue was reported by Guido Vranken.
5482 * Modify behavior of ALPN to invoke callback after SNI/servername
5483 callback, such that updates to the SSL_CTX affect ALPN.
5487 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5492 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5493 methods are enabled and ssl2 is disabled the methods return NULL.
5497 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5499 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5500 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5501 provide any "EXPORT" or "LOW" strength ciphers.
5505 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5506 is by default disabled at build-time. Builds that are not configured with
5507 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5508 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5509 will need to explicitly call either of:
5511 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5513 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5515 as appropriate. Even if either of those is used, or the application
5516 explicitly uses the version-specific SSLv2_method() or its client and
5517 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5518 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5519 ciphers, and SSLv2 56-bit DES are no longer available.
5524 * Fix a double-free in DSA code
5526 A double free bug was discovered when OpenSSL parses malformed DSA private
5527 keys and could lead to a DoS attack or memory corruption for applications
5528 that receive DSA private keys from untrusted sources. This scenario is
5531 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5537 * Disable SRP fake user seed to address a server memory leak.
5539 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5541 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5542 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5543 was changed to ignore the "fake user" SRP seed, even if the seed
5546 Users should use SRP_VBASE_get1_by_user instead. Note that in
5547 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5548 also that even though configuring the SRP seed attempts to hide
5549 invalid usernames by continuing the handshake with fake
5550 credentials, this behaviour is not constant time and no strong
5551 guarantees are made that the handshake is indistinguishable from
5552 that of a valid user.
5557 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5559 In the BN_hex2bn function the number of hex digits is calculated using an
5560 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5561 large values of |i| this can result in |bn_expand| not allocating any
5562 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5563 field as NULL leading to a subsequent NULL ptr deref. For very large values
5564 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5565 In this case memory is allocated to the internal BIGNUM data field, but it
5566 is insufficiently sized leading to heap corruption. A similar issue exists
5567 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5568 is ever called by user applications with very large untrusted hex/dec data.
5569 This is anticipated to be a rare occurrence.
5571 All OpenSSL internal usage of these functions use data that is not expected
5572 to be untrusted, e.g. config file data or application command line
5573 arguments. If user developed applications generate config file data based
5574 on untrusted data then it is possible that this could also lead to security
5575 consequences. This is also anticipated to be rare.
5577 This issue was reported to OpenSSL by Guido Vranken.
5582 * Fix memory issues in `BIO_*printf` functions
5584 The internal |fmtstr| function used in processing a "%s" format string in
5585 the `BIO_*printf` functions could overflow while calculating the length of a
5586 string and cause an OOB read when printing very long strings.
5588 Additionally the internal |doapr_outch| function can attempt to write to an
5589 OOB memory location (at an offset from the NULL pointer) in the event of a
5590 memory allocation failure. In 1.0.2 and below this could be caused where
5591 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5592 could be in processing a very long "%s" format string. Memory leaks can
5595 The first issue may mask the second issue dependent on compiler behaviour.
5596 These problems could enable attacks where large amounts of untrusted data
5597 is passed to the `BIO_*printf` functions. If applications use these functions
5598 in this way then they could be vulnerable. OpenSSL itself uses these
5599 functions when printing out human-readable dumps of ASN.1 data. Therefore
5600 applications that print this data could be vulnerable if the data is from
5601 untrusted sources. OpenSSL command line applications could also be
5602 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5603 as command line arguments.
5605 Libssl is not considered directly vulnerable. Additionally certificates etc
5606 received via remote connections via libssl are also unlikely to be able to
5607 trigger these issues because of message size limits enforced within libssl.
5609 This issue was reported to OpenSSL Guido Vranken.
5614 * Side channel attack on modular exponentiation
5616 A side-channel attack was found which makes use of cache-bank conflicts on
5617 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5618 of RSA keys. The ability to exploit this issue is limited as it relies on
5619 an attacker who has control of code in a thread running on the same
5620 hyper-threaded core as the victim thread which is performing decryptions.
5622 This issue was reported to OpenSSL by Yuval Yarom, The University of
5623 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5624 Nadia Heninger, University of Pennsylvania with more information at
5625 <http://cachebleed.info>.
5630 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5631 if no keysize is specified with default_bits. This fixes an
5632 omission in an earlier change that changed all RSA/DSA key generation
5633 apps to use 2048 bits by default.
5637 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5639 * Protection for DH small subgroup attacks
5641 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5642 switched on by default and cannot be disabled. This could have some
5647 * SSLv2 doesn't block disabled ciphers
5649 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5650 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5651 been disabled, provided that the SSLv2 protocol was not also disabled via
5654 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5655 and Sebastian Schinzel.
5660 * Reject DH handshakes with parameters shorter than 1024 bits.
5664 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5666 * Certificate verify crash with missing PSS parameter
5668 The signature verification routines will crash with a NULL pointer
5669 dereference if presented with an ASN.1 signature using the RSA PSS
5670 algorithm and absent mask generation function parameter. Since these
5671 routines are used to verify certificate signature algorithms this can be
5672 used to crash any certificate verification operation and exploited in a
5673 DoS attack. Any application which performs certificate verification is
5674 vulnerable including OpenSSL clients and servers which enable client
5677 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5682 * X509_ATTRIBUTE memory leak
5684 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5685 memory. This structure is used by the PKCS#7 and CMS routines so any
5686 application which reads PKCS#7 or CMS data from untrusted sources is
5687 affected. SSL/TLS is not affected.
5689 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5695 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5696 This changes the decoding behaviour for some invalid messages,
5697 though the change is mostly in the more lenient direction, and
5698 legacy behaviour is preserved as much as possible.
5702 * In DSA_generate_parameters_ex, if the provided seed is too short,
5703 use a random seed, as already documented.
5705 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5707 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5709 * Alternate chains certificate forgery
5711 During certificate verfification, OpenSSL will attempt to find an
5712 alternative certificate chain if the first attempt to build such a chain
5713 fails. An error in the implementation of this logic can mean that an
5714 attacker could cause certain checks on untrusted certificates to be
5715 bypassed, such as the CA flag, enabling them to use a valid leaf
5716 certificate to act as a CA and "issue" an invalid certificate.
5718 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5724 * Race condition handling PSK identify hint
5726 If PSK identity hints are received by a multi-threaded client then
5727 the values are wrongly updated in the parent SSL_CTX structure. This can
5728 result in a race condition potentially leading to a double free of the
5734 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5736 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5737 incompatibility in the handling of HMAC. The previous ABI has now been
5740 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5742 * Malformed ECParameters causes infinite loop
5744 When processing an ECParameters structure OpenSSL enters an infinite loop
5745 if the curve specified is over a specially malformed binary polynomial
5748 This can be used to perform denial of service against any
5749 system which processes public keys, certificate requests or
5750 certificates. This includes TLS clients and TLS servers with
5751 client authentication enabled.
5753 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5758 * Exploitable out-of-bounds read in X509_cmp_time
5760 X509_cmp_time does not properly check the length of the ASN1_TIME
5761 string and can read a few bytes out of bounds. In addition,
5762 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5765 An attacker can use this to craft malformed certificates and CRLs of
5766 various sizes and potentially cause a segmentation fault, resulting in
5767 a DoS on applications that verify certificates or CRLs. TLS clients
5768 that verify CRLs are affected. TLS clients and servers with client
5769 authentication enabled may be affected if they use custom verification
5772 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5773 independently by Hanno Böck.
5778 * PKCS7 crash with missing EnvelopedContent
5780 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5781 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5782 with missing content and trigger a NULL pointer dereference on parsing.
5784 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5785 structures from untrusted sources are affected. OpenSSL clients and
5786 servers are not affected.
5788 This issue was reported to OpenSSL by Michal Zalewski (Google).
5793 * CMS verify infinite loop with unknown hash function
5795 When verifying a signedData message the CMS code can enter an infinite loop
5796 if presented with an unknown hash function OID. This can be used to perform
5797 denial of service against any system which verifies signedData messages using
5799 This issue was reported to OpenSSL by Johannes Bauer.
5804 * Race condition handling NewSessionTicket
5806 If a NewSessionTicket is received by a multi-threaded client when attempting to
5807 reuse a previous ticket then a race condition can occur potentially leading to
5808 a double free of the ticket data.
5813 * Reject DH handshakes with parameters shorter than 768 bits.
5815 *Kurt Roeckx and Emilia Kasper*
5817 * dhparam: generate 2048-bit parameters by default.
5819 *Kurt Roeckx and Emilia Kasper*
5821 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5823 * Segmentation fault in ASN1_TYPE_cmp fix
5825 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5826 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5827 certificate signature algorithm consistency this can be used to crash any
5828 certificate verification operation and exploited in a DoS attack. Any
5829 application which performs certificate verification is vulnerable including
5830 OpenSSL clients and servers which enable client authentication.
5835 * ASN.1 structure reuse memory corruption fix
5837 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5838 memory corruption via an invalid write. Such reuse is and has been
5839 strongly discouraged and is believed to be rare.
5841 Applications that parse structures containing CHOICE or ANY DEFINED BY
5842 components may be affected. Certificate parsing (d2i_X509 and related
5843 functions) are however not affected. OpenSSL clients and servers are
5849 * PKCS7 NULL pointer dereferences fix
5851 The PKCS#7 parsing code does not handle missing outer ContentInfo
5852 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5853 missing content and trigger a NULL pointer dereference on parsing.
5855 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5856 otherwise parse PKCS#7 structures from untrusted sources are
5857 affected. OpenSSL clients and servers are not affected.
5859 This issue was reported to OpenSSL by Michal Zalewski (Google).
5864 * DoS via reachable assert in SSLv2 servers fix
5866 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5867 servers that both support SSLv2 and enable export cipher suites by sending
5868 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5870 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5871 (OpenSSL development team).
5876 * Use After Free following d2i_ECPrivatekey error fix
5878 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5879 could cause a use after free condition. This, in turn, could cause a double
5880 free in several private key parsing functions (such as d2i_PrivateKey
5881 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5882 for applications that receive EC private keys from untrusted
5883 sources. This scenario is considered rare.
5885 This issue was discovered by the BoringSSL project and fixed in their
5891 * X509_to_X509_REQ NULL pointer deref fix
5893 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5894 the certificate key is invalid. This function is rarely used in practice.
5896 This issue was discovered by Brian Carpenter.
5901 * Removed the export ciphers from the DEFAULT ciphers
5905 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
5907 * Build fixes for the Windows and OpenVMS platforms
5909 *Matt Caswell and Richard Levitte*
5911 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
5913 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5914 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5915 dereference. This could lead to a Denial Of Service attack. Thanks to
5916 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5921 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5922 dtls1_buffer_record function under certain conditions. In particular this
5923 could occur if an attacker sent repeated DTLS records with the same
5924 sequence number but for the next epoch. The memory leak could be exploited
5925 by an attacker in a Denial of Service attack through memory exhaustion.
5926 Thanks to Chris Mueller for reporting this issue.
5931 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5932 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5933 method would be set to NULL which could later result in a NULL pointer
5934 dereference. Thanks to Frank Schmirler for reporting this issue.
5939 * Abort handshake if server key exchange message is omitted for ephemeral
5942 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5943 reporting this issue.
5948 * Remove non-export ephemeral RSA code on client and server. This code
5949 violated the TLS standard by allowing the use of temporary RSA keys in
5950 non-export ciphersuites and could be used by a server to effectively
5951 downgrade the RSA key length used to a value smaller than the server
5952 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5953 INRIA or reporting this issue.
5958 * Fixed issue where DH client certificates are accepted without verification.
5959 An OpenSSL server will accept a DH certificate for client authentication
5960 without the certificate verify message. This effectively allows a client to
5961 authenticate without the use of a private key. This only affects servers
5962 which trust a client certificate authority which issues certificates
5963 containing DH keys: these are extremely rare and hardly ever encountered.
5964 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5970 * Ensure that the session ID context of an SSL is updated when its
5971 SSL_CTX is updated via SSL_set_SSL_CTX.
5973 The session ID context is typically set from the parent SSL_CTX,
5974 and can vary with the CTX.
5978 * Fix various certificate fingerprint issues.
5980 By using non-DER or invalid encodings outside the signed portion of a
5981 certificate the fingerprint can be changed without breaking the signature.
5982 Although no details of the signed portion of the certificate can be changed
5983 this can cause problems with some applications: e.g. those using the
5984 certificate fingerprint for blacklists.
5986 1. Reject signatures with non zero unused bits.
5988 If the BIT STRING containing the signature has non zero unused bits reject
5989 the signature. All current signature algorithms require zero unused bits.
5991 2. Check certificate algorithm consistency.
5993 Check the AlgorithmIdentifier inside TBS matches the one in the
5994 certificate signature. NB: this will result in signature failure
5995 errors for some broken certificates.
5997 Thanks to Konrad Kraszewski from Google for reporting this issue.
5999 3. Check DSA/ECDSA signatures use DER.
6001 Re-encode DSA/ECDSA signatures and compare with the original received
6002 signature. Return an error if there is a mismatch.
6004 This will reject various cases including garbage after signature
6005 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6006 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6007 (negative or with leading zeroes).
6009 Further analysis was conducted and fixes were developed by Stephen Henson
6010 of the OpenSSL core team.
6016 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6017 results on some platforms, including x86_64. This bug occurs at random
6018 with a very low probability, and is not known to be exploitable in any
6019 way, though its exact impact is difficult to determine. Thanks to Pieter
6020 Wuille (Blockstream) who reported this issue and also suggested an initial
6021 fix. Further analysis was conducted by the OpenSSL development team and
6022 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6023 the OpenSSL core team.
6028 * Do not resume sessions on the server if the negotiated protocol
6029 version does not match the session's version. Resuming with a different
6030 version, while not strictly forbidden by the RFC, is of questionable
6031 sanity and breaks all known clients.
6033 *David Benjamin, Emilia Käsper*
6035 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6036 early CCS messages during renegotiation. (Note that because
6037 renegotiation is encrypted, this early CCS was not exploitable.)
6041 * Tighten client-side session ticket handling during renegotiation:
6042 ensure that the client only accepts a session ticket if the server sends
6043 the extension anew in the ServerHello. Previously, a TLS client would
6044 reuse the old extension state and thus accept a session ticket if one was
6045 announced in the initial ServerHello.
6047 Similarly, ensure that the client requires a session ticket if one
6048 was advertised in the ServerHello. Previously, a TLS client would
6049 ignore a missing NewSessionTicket message.
6053 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6057 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6058 sends a carefully crafted handshake message, to cause OpenSSL to fail
6059 to free up to 64k of memory causing a memory leak. This could be
6060 exploited in a Denial Of Service attack. This issue affects OpenSSL
6061 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6062 whether SRTP is used or configured. Implementations of OpenSSL that
6063 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6065 The fix was developed by the OpenSSL team.
6070 * Session Ticket Memory Leak.
6072 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6073 integrity of that ticket is first verified. In the event of a session
6074 ticket integrity check failing, OpenSSL will fail to free memory
6075 causing a memory leak. By sending a large number of invalid session
6076 tickets an attacker could exploit this issue in a Denial Of Service
6082 * Build option no-ssl3 is incomplete.
6084 When OpenSSL is configured with "no-ssl3" as a build option, servers
6085 could accept and complete a SSL 3.0 handshake, and clients could be
6086 configured to send them.
6089 *Akamai and the OpenSSL team*
6091 * Add support for TLS_FALLBACK_SCSV.
6092 Client applications doing fallback retries should call
6093 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6096 *Adam Langley, Bodo Moeller*
6098 * Add additional DigestInfo checks.
6100 Re-encode DigestInto in DER and check against the original when
6101 verifying RSA signature: this will reject any improperly encoded
6102 DigestInfo structures.
6104 Note: this is a precautionary measure and no attacks are currently known.
6108 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6110 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6111 SRP code can be overrun an internal buffer. Add sanity check that
6112 g, A, B < N to SRP code.
6114 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6115 Group for discovering this issue.
6120 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6121 TLS 1.0 instead of higher protocol versions when the ClientHello message
6122 is badly fragmented. This allows a man-in-the-middle attacker to force a
6123 downgrade to TLS 1.0 even if both the server and the client support a
6124 higher protocol version, by modifying the client's TLS records.
6126 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6127 researching this issue.
6132 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6133 to a denial of service attack. A malicious server can crash the client
6134 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6135 ciphersuite and sending carefully crafted handshake messages.
6137 Thanks to Felix Gröbert (Google) for discovering and researching this
6143 * By sending carefully crafted DTLS packets an attacker could cause openssl
6144 to leak memory. This can be exploited through a Denial of Service attack.
6145 Thanks to Adam Langley for discovering and researching this issue.
6150 * An attacker can force openssl to consume large amounts of memory whilst
6151 processing DTLS handshake messages. This can be exploited through a
6152 Denial of Service attack.
6153 Thanks to Adam Langley for discovering and researching this issue.
6158 * An attacker can force an error condition which causes openssl to crash
6159 whilst processing DTLS packets due to memory being freed twice. This
6160 can be exploited through a Denial of Service attack.
6161 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6167 * If a multithreaded client connects to a malicious server using a resumed
6168 session and the server sends an ec point format extension it could write
6169 up to 255 bytes to freed memory.
6171 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6177 * A malicious server can crash an OpenSSL client with a null pointer
6178 dereference (read) by specifying an SRP ciphersuite even though it was not
6179 properly negotiated with the client. This can be exploited through a
6180 Denial of Service attack.
6182 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6183 discovering and researching this issue.
6188 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6189 X509_name_oneline, X509_name_print_ex et al. to leak some information
6190 from the stack. Applications may be affected if they echo pretty printing
6191 output to the attacker.
6193 Thanks to Ivan Fratric (Google) for discovering this issue.
6196 *Emilia Käsper, and Steve Henson*
6198 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6199 for corner cases. (Certain input points at infinity could lead to
6200 bogus results, with non-infinity inputs mapped to infinity too.)
6204 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6206 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6207 handshake can force the use of weak keying material in OpenSSL
6208 SSL/TLS clients and servers.
6210 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6211 researching this issue. [CVE-2014-0224][]
6213 *KIKUCHI Masashi, Steve Henson*
6215 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6216 OpenSSL DTLS client the code can be made to recurse eventually crashing
6219 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6222 *Imre Rad, Steve Henson*
6224 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6225 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6226 client or server. This is potentially exploitable to run arbitrary
6227 code on a vulnerable client or server.
6229 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6231 *Jüri Aedla, Steve Henson*
6233 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6234 are subject to a denial of service attack.
6236 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6237 this issue. [CVE-2014-3470][]
6239 *Felix Gröbert, Ivan Fratric, Steve Henson*
6241 * Harmonize version and its documentation. -f flag is used to display
6244 *mancha <mancha1@zoho.com>*
6246 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6247 in i2d_ECPrivateKey.
6249 *mancha <mancha1@zoho.com>*
6251 * Fix some double frees. These are not thought to be exploitable.
6253 *mancha <mancha1@zoho.com>*
6255 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6257 * A missing bounds check in the handling of the TLS heartbeat extension
6258 can be used to reveal up to 64k of memory to a connected client or
6261 Thanks for Neel Mehta of Google Security for discovering this bug and to
6262 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6263 preparing the fix [CVE-2014-0160][]
6265 *Adam Langley, Bodo Moeller*
6267 * Fix for the attack described in the paper "Recovering OpenSSL
6268 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6269 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6270 <http://eprint.iacr.org/2014/140>
6272 Thanks to Yuval Yarom and Naomi Benger for discovering this
6273 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6275 *Yuval Yarom and Naomi Benger*
6277 * TLS pad extension: draft-agl-tls-padding-03
6279 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6280 TLS client Hello record length value would otherwise be > 255 and
6281 less that 512 pad with a dummy extension containing zeroes so it
6282 is at least 512 bytes long.
6284 *Adam Langley, Steve Henson*
6286 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6288 * Fix for TLS record tampering bug. A carefully crafted invalid
6289 handshake could crash OpenSSL with a NULL pointer exception.
6290 Thanks to Anton Johansson for reporting this issues.
6293 * Keep original DTLS digest and encryption contexts in retransmission
6294 structures so we can use the previous session parameters if they need
6295 to be resent. [CVE-2013-6450][]
6299 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6300 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6301 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6302 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6303 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6304 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6306 *Rob Stradling, Adam Langley*
6308 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6310 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6311 supporting platforms or when small records were transferred.
6313 *Andy Polyakov, Steve Henson*
6315 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6317 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6319 This addresses the flaw in CBC record processing discovered by
6320 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6321 at: <http://www.isg.rhul.ac.uk/tls/>
6323 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6324 Security Group at Royal Holloway, University of London
6325 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6326 Emilia Käsper for the initial patch.
6329 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6331 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6332 ciphersuites which can be exploited in a denial of service attack.
6333 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6334 and detecting this bug and to Wolfgang Ettlinger
6335 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6340 * Return an error when checking OCSP signatures when key is NULL.
6341 This fixes a DoS attack. [CVE-2013-0166][]
6345 * Make openssl verify return errors.
6347 *Chris Palmer <palmer@google.com> and Ben Laurie*
6349 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6350 the right response is stapled. Also change SSL_get_certificate()
6351 so it returns the certificate actually sent.
6352 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6354 *Rob Stradling <rob.stradling@comodo.com>*
6356 * Fix possible deadlock when decoding public keys.
6360 * Don't use TLS 1.0 record version number in initial client hello
6365 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6367 * Sanity check record length before skipping explicit IV in TLS
6368 1.2, 1.1 and DTLS to fix DoS attack.
6370 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6371 fuzzing as a service testing platform.
6376 * Initialise tkeylen properly when encrypting CMS messages.
6377 Thanks to Solar Designer of Openwall for reporting this issue.
6381 * In FIPS mode don't try to use composite ciphers as they are not
6386 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6388 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6389 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6390 mean any application compiled against OpenSSL 1.0.0 headers setting
6391 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6392 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6393 0x10000000L Any application which was previously compiled against
6394 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6395 will need to be recompiled as a result. Letting be results in
6396 inability to disable specifically TLS 1.1 and in client context,
6397 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6401 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6402 disable just protocol X, but all protocols above X *if* there are
6403 protocols *below* X still enabled. In more practical terms it means
6404 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6405 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6406 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6411 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6413 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6414 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6415 in CRYPTO_realloc_clean.
6417 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6418 issue and to Adam Langley <agl@chromium.org> for fixing it.
6421 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6423 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6427 * Workarounds for some broken servers that "hang" if a client hello
6428 record length exceeds 255 bytes.
6430 1. Do not use record version number > TLS 1.0 in initial client
6431 hello: some (but not all) hanging servers will now work.
6432 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6433 the number of ciphers sent in the client hello. This should be
6434 set to an even number, such as 50, for example by passing:
6435 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6436 Most broken servers should now work.
6437 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6438 TLS 1.2 client support entirely.
6442 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6446 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6448 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6449 STRING form instead of a DigestInfo.
6453 * The format used for MDC2 RSA signatures is inconsistent between EVP
6454 and the RSA_sign/RSA_verify functions. This was made more apparent when
6455 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6456 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6457 the correct format in RSA_verify so both forms transparently work.
6461 * Some servers which support TLS 1.0 can choke if we initially indicate
6462 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6463 encrypted premaster secret. As a workaround use the maximum permitted
6464 client version in client hello, this should keep such servers happy
6465 and still work with previous versions of OpenSSL.
6469 * Add support for TLS/DTLS heartbeats.
6471 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6473 * Add support for SCTP.
6475 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6477 * Improved PRNG seeding for VOS.
6479 *Paul Green <Paul.Green@stratus.com>*
6481 * Extensive assembler packs updates, most notably:
6483 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6484 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6485 - x86_64: bit-sliced AES implementation;
6486 - ARM: NEON support, contemporary platforms optimizations;
6487 - s390x: z196 support;
6488 - `*`: GHASH and GF(2^m) multiplication implementations;
6492 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6493 (removal of unnecessary code)
6495 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6497 * Add TLS key material exporter from RFC 5705.
6501 * Add DTLS-SRTP negotiation from RFC 5764.
6505 * Add Next Protocol Negotiation,
6506 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6507 disabled with a no-npn flag to config or Configure. Code donated
6510 *Adam Langley <agl@google.com> and Ben Laurie*
6512 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6513 NIST-P256, NIST-P521, with constant-time single point multiplication on
6514 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6515 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6516 Code made available under Apache License version 2.0.
6518 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6519 line to include this in your build of OpenSSL, and run "make depend" (or
6520 "make update"). This enables the following EC_METHODs:
6522 EC_GFp_nistp224_method()
6523 EC_GFp_nistp256_method()
6524 EC_GFp_nistp521_method()
6526 EC_GROUP_new_by_curve_name() will automatically use these (while
6527 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6530 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6532 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6533 all platforms. Move ssize_t definition from e_os.h to the public
6534 header file e_os2.h as it now appears in public header file cms.h
6538 * New -sigopt option to the ca, req and x509 utilities. Additional
6539 signature parameters can be passed using this option and in
6544 * Add RSA PSS signing function. This will generate and set the
6545 appropriate AlgorithmIdentifiers for PSS based on those in the
6546 corresponding EVP_MD_CTX structure. No application support yet.
6550 * Support for companion algorithm specific ASN1 signing routines.
6551 New function ASN1_item_sign_ctx() signs a pre-initialised
6552 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6553 the appropriate parameters.
6557 * Add new algorithm specific ASN1 verification initialisation function
6558 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6559 handling will be the same no matter what EVP_PKEY_METHOD is used.
6560 Add a PSS handler to support verification of PSS signatures: checked
6561 against a number of sample certificates.
6565 * Add signature printing for PSS. Add PSS OIDs.
6567 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6569 * Add algorithm specific signature printing. An individual ASN1 method
6570 can now print out signatures instead of the standard hex dump.
6572 More complex signatures (e.g. PSS) can print out more meaningful
6573 information. Include DSA version that prints out the signature
6578 * Password based recipient info support for CMS library: implementing
6583 * Split password based encryption into PBES2 and PBKDF2 functions. This
6584 neatly separates the code into cipher and PBE sections and is required
6585 for some algorithms that split PBES2 into separate pieces (such as
6586 password based CMS).
6590 * Session-handling fixes:
6591 - Fix handling of connections that are resuming with a session ID,
6592 but also support Session Tickets.
6593 - Fix a bug that suppressed issuing of a new ticket if the client
6594 presented a ticket with an expired session.
6595 - Try to set the ticket lifetime hint to something reasonable.
6596 - Make tickets shorter by excluding irrelevant information.
6597 - On the client side, don't ignore renewed tickets.
6599 *Adam Langley, Bodo Moeller (Google)*
6601 * Fix PSK session representation.
6605 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6607 This work was sponsored by Intel.
6611 * Add GCM support to TLS library. Some custom code is needed to split
6612 the IV between the fixed (from PRF) and explicit (from TLS record)
6613 portions. This adds all GCM ciphersuites supported by RFC5288 and
6614 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6615 add a special AESGCM string for GCM only.
6619 * Expand range of ctrls for AES GCM. Permit setting invocation
6620 field on decrypt and retrieval of invocation field only on encrypt.
6624 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6625 As required by RFC5289 these ciphersuites cannot be used if for
6626 versions of TLS earlier than 1.2.
6630 * For FIPS capable OpenSSL interpret a NULL default public key method
6631 as unset and return the appropriate default but do *not* set the default.
6632 This means we can return the appropriate method in applications that
6633 switch between FIPS and non-FIPS modes.
6637 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6638 ENGINE is used then we cannot handle that in the FIPS module so we
6639 keep original code iff non-FIPS operations are allowed.
6643 * Add -attime option to openssl utilities.
6645 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6647 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6651 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6652 FIPS EC methods unconditionally for now.
6656 * New build option no-ec2m to disable characteristic 2 code.
6660 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6661 all cases can be covered as some introduce binary incompatibilities.
6665 * Redirect RSA operations to FIPS module including keygen,
6666 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6670 * Add similar low level API blocking to ciphers.
6674 * Low level digest APIs are not approved in FIPS mode: any attempt
6675 to use these will cause a fatal error. Applications that *really* want
6676 to use them can use the `private_*` version instead.
6680 * Redirect cipher operations to FIPS module for FIPS builds.
6684 * Redirect digest operations to FIPS module for FIPS builds.
6688 * Update build system to add "fips" flag which will link in fipscanister.o
6689 for static and shared library builds embedding a signature if needed.
6693 * Output TLS supported curves in preference order instead of numerical
6694 order. This is currently hardcoded for the highest order curves first.
6695 This should be configurable so applications can judge speed vs strength.
6699 * Add TLS v1.2 server support for client authentication.
6703 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6708 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6709 FIPS modules versions.
6713 * Add TLS v1.2 client side support for client authentication. Keep cache
6714 of handshake records longer as we don't know the hash algorithm to use
6715 until after the certificate request message is received.
6719 * Initial TLS v1.2 client support. Add a default signature algorithms
6720 extension including all the algorithms we support. Parse new signature
6721 format in client key exchange. Relax some ECC signing restrictions for
6722 TLS v1.2 as indicated in RFC5246.
6726 * Add server support for TLS v1.2 signature algorithms extension. Switch
6727 to new signature format when needed using client digest preference.
6728 All server ciphersuites should now work correctly in TLS v1.2. No client
6729 support yet and no support for client certificates.
6733 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6734 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6735 ciphersuites. At present only RSA key exchange ciphersuites work with
6736 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6737 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6738 and version checking.
6742 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6743 with this defined it will not be affected by any changes to ssl internal
6744 structures. Add several utility functions to allow openssl application
6745 to work with OPENSSL_NO_SSL_INTERN defined.
6749 * A long standing patch to add support for SRP from EdelWeb (Peter
6750 Sylvester and Christophe Renou) was integrated.
6751 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6752 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6755 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6759 * Permit abbreviated handshakes when renegotiating using the function
6760 SSL_renegotiate_abbreviated().
6762 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6764 * Add call to ENGINE_register_all_complete() to
6765 ENGINE_load_builtin_engines(), so some implementations get used
6766 automatically instead of needing explicit application support.
6770 * Add support for TLS key exporter as described in RFC5705.
6772 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6774 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6775 a few changes are required:
6777 Add SSL_OP_NO_TLSv1_1 flag.
6778 Add TLSv1_1 methods.
6779 Update version checking logic to handle version 1.1.
6780 Add explicit IV handling (ported from DTLS code).
6781 Add command line options to s_client/s_server.
6788 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6790 * X509_ATTRIBUTE memory leak
6792 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6793 memory. This structure is used by the PKCS#7 and CMS routines so any
6794 application which reads PKCS#7 or CMS data from untrusted sources is
6795 affected. SSL/TLS is not affected.
6797 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6803 * Race condition handling PSK identify hint
6805 If PSK identity hints are received by a multi-threaded client then
6806 the values are wrongly updated in the parent SSL_CTX structure. This can
6807 result in a race condition potentially leading to a double free of the
6813 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6815 * Malformed ECParameters causes infinite loop
6817 When processing an ECParameters structure OpenSSL enters an infinite loop
6818 if the curve specified is over a specially malformed binary polynomial
6821 This can be used to perform denial of service against any
6822 system which processes public keys, certificate requests or
6823 certificates. This includes TLS clients and TLS servers with
6824 client authentication enabled.
6826 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6831 * Exploitable out-of-bounds read in X509_cmp_time
6833 X509_cmp_time does not properly check the length of the ASN1_TIME
6834 string and can read a few bytes out of bounds. In addition,
6835 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6838 An attacker can use this to craft malformed certificates and CRLs of
6839 various sizes and potentially cause a segmentation fault, resulting in
6840 a DoS on applications that verify certificates or CRLs. TLS clients
6841 that verify CRLs are affected. TLS clients and servers with client
6842 authentication enabled may be affected if they use custom verification
6845 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6846 independently by Hanno Böck.
6851 * PKCS7 crash with missing EnvelopedContent
6853 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6854 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6855 with missing content and trigger a NULL pointer dereference on parsing.
6857 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6858 structures from untrusted sources are affected. OpenSSL clients and
6859 servers are not affected.
6861 This issue was reported to OpenSSL by Michal Zalewski (Google).
6866 * CMS verify infinite loop with unknown hash function
6868 When verifying a signedData message the CMS code can enter an infinite loop
6869 if presented with an unknown hash function OID. This can be used to perform
6870 denial of service against any system which verifies signedData messages using
6872 This issue was reported to OpenSSL by Johannes Bauer.
6877 * Race condition handling NewSessionTicket
6879 If a NewSessionTicket is received by a multi-threaded client when attempting to
6880 reuse a previous ticket then a race condition can occur potentially leading to
6881 a double free of the ticket data.
6886 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
6888 * Segmentation fault in ASN1_TYPE_cmp fix
6890 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6891 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6892 certificate signature algorithm consistency this can be used to crash any
6893 certificate verification operation and exploited in a DoS attack. Any
6894 application which performs certificate verification is vulnerable including
6895 OpenSSL clients and servers which enable client authentication.
6900 * ASN.1 structure reuse memory corruption fix
6902 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6903 memory corruption via an invalid write. Such reuse is and has been
6904 strongly discouraged and is believed to be rare.
6906 Applications that parse structures containing CHOICE or ANY DEFINED BY
6907 components may be affected. Certificate parsing (d2i_X509 and related
6908 functions) are however not affected. OpenSSL clients and servers are
6914 * PKCS7 NULL pointer dereferences fix
6916 The PKCS#7 parsing code does not handle missing outer ContentInfo
6917 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6918 missing content and trigger a NULL pointer dereference on parsing.
6920 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6921 otherwise parse PKCS#7 structures from untrusted sources are
6922 affected. OpenSSL clients and servers are not affected.
6924 This issue was reported to OpenSSL by Michal Zalewski (Google).
6929 * DoS via reachable assert in SSLv2 servers fix
6931 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6932 servers that both support SSLv2 and enable export cipher suites by sending
6933 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6935 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6936 (OpenSSL development team).
6941 * Use After Free following d2i_ECPrivatekey error fix
6943 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6944 could cause a use after free condition. This, in turn, could cause a double
6945 free in several private key parsing functions (such as d2i_PrivateKey
6946 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6947 for applications that receive EC private keys from untrusted
6948 sources. This scenario is considered rare.
6950 This issue was discovered by the BoringSSL project and fixed in their
6956 * X509_to_X509_REQ NULL pointer deref fix
6958 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6959 the certificate key is invalid. This function is rarely used in practice.
6961 This issue was discovered by Brian Carpenter.
6966 * Removed the export ciphers from the DEFAULT ciphers
6970 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
6972 * Build fixes for the Windows and OpenVMS platforms
6974 *Matt Caswell and Richard Levitte*
6976 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
6978 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6979 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6980 dereference. This could lead to a Denial Of Service attack. Thanks to
6981 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6986 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6987 dtls1_buffer_record function under certain conditions. In particular this
6988 could occur if an attacker sent repeated DTLS records with the same
6989 sequence number but for the next epoch. The memory leak could be exploited
6990 by an attacker in a Denial of Service attack through memory exhaustion.
6991 Thanks to Chris Mueller for reporting this issue.
6996 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6997 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6998 method would be set to NULL which could later result in a NULL pointer
6999 dereference. Thanks to Frank Schmirler for reporting this issue.
7004 * Abort handshake if server key exchange message is omitted for ephemeral
7007 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7008 reporting this issue.
7013 * Remove non-export ephemeral RSA code on client and server. This code
7014 violated the TLS standard by allowing the use of temporary RSA keys in
7015 non-export ciphersuites and could be used by a server to effectively
7016 downgrade the RSA key length used to a value smaller than the server
7017 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7018 INRIA or reporting this issue.
7023 * Fixed issue where DH client certificates are accepted without verification.
7024 An OpenSSL server will accept a DH certificate for client authentication
7025 without the certificate verify message. This effectively allows a client to
7026 authenticate without the use of a private key. This only affects servers
7027 which trust a client certificate authority which issues certificates
7028 containing DH keys: these are extremely rare and hardly ever encountered.
7029 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7035 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7036 results on some platforms, including x86_64. This bug occurs at random
7037 with a very low probability, and is not known to be exploitable in any
7038 way, though its exact impact is difficult to determine. Thanks to Pieter
7039 Wuille (Blockstream) who reported this issue and also suggested an initial
7040 fix. Further analysis was conducted by the OpenSSL development team and
7041 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7042 the OpenSSL core team.
7047 * Fix various certificate fingerprint issues.
7049 By using non-DER or invalid encodings outside the signed portion of a
7050 certificate the fingerprint can be changed without breaking the signature.
7051 Although no details of the signed portion of the certificate can be changed
7052 this can cause problems with some applications: e.g. those using the
7053 certificate fingerprint for blacklists.
7055 1. Reject signatures with non zero unused bits.
7057 If the BIT STRING containing the signature has non zero unused bits reject
7058 the signature. All current signature algorithms require zero unused bits.
7060 2. Check certificate algorithm consistency.
7062 Check the AlgorithmIdentifier inside TBS matches the one in the
7063 certificate signature. NB: this will result in signature failure
7064 errors for some broken certificates.
7066 Thanks to Konrad Kraszewski from Google for reporting this issue.
7068 3. Check DSA/ECDSA signatures use DER.
7070 Reencode DSA/ECDSA signatures and compare with the original received
7071 signature. Return an error if there is a mismatch.
7073 This will reject various cases including garbage after signature
7074 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7075 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7076 (negative or with leading zeroes).
7078 Further analysis was conducted and fixes were developed by Stephen Henson
7079 of the OpenSSL core team.
7085 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7087 * Session Ticket Memory Leak.
7089 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7090 integrity of that ticket is first verified. In the event of a session
7091 ticket integrity check failing, OpenSSL will fail to free memory
7092 causing a memory leak. By sending a large number of invalid session
7093 tickets an attacker could exploit this issue in a Denial Of Service
7099 * Build option no-ssl3 is incomplete.
7101 When OpenSSL is configured with "no-ssl3" as a build option, servers
7102 could accept and complete a SSL 3.0 handshake, and clients could be
7103 configured to send them.
7106 *Akamai and the OpenSSL team*
7108 * Add support for TLS_FALLBACK_SCSV.
7109 Client applications doing fallback retries should call
7110 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7113 *Adam Langley, Bodo Moeller*
7115 * Add additional DigestInfo checks.
7117 Reencode DigestInto in DER and check against the original when
7118 verifying RSA signature: this will reject any improperly encoded
7119 DigestInfo structures.
7121 Note: this is a precautionary measure and no attacks are currently known.
7125 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7127 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7128 to a denial of service attack. A malicious server can crash the client
7129 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7130 ciphersuite and sending carefully crafted handshake messages.
7132 Thanks to Felix Gröbert (Google) for discovering and researching this
7138 * By sending carefully crafted DTLS packets an attacker could cause openssl
7139 to leak memory. This can be exploited through a Denial of Service attack.
7140 Thanks to Adam Langley for discovering and researching this issue.
7145 * An attacker can force openssl to consume large amounts of memory whilst
7146 processing DTLS handshake messages. This can be exploited through a
7147 Denial of Service attack.
7148 Thanks to Adam Langley for discovering and researching this issue.
7153 * An attacker can force an error condition which causes openssl to crash
7154 whilst processing DTLS packets due to memory being freed twice. This
7155 can be exploited through a Denial of Service attack.
7156 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7162 * If a multithreaded client connects to a malicious server using a resumed
7163 session and the server sends an ec point format extension it could write
7164 up to 255 bytes to freed memory.
7166 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7172 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7173 X509_name_oneline, X509_name_print_ex et al. to leak some information
7174 from the stack. Applications may be affected if they echo pretty printing
7175 output to the attacker.
7177 Thanks to Ivan Fratric (Google) for discovering this issue.
7180 *Emilia Käsper, and Steve Henson*
7182 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7183 for corner cases. (Certain input points at infinity could lead to
7184 bogus results, with non-infinity inputs mapped to infinity too.)
7188 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7190 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7191 handshake can force the use of weak keying material in OpenSSL
7192 SSL/TLS clients and servers.
7194 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7195 researching this issue. [CVE-2014-0224][]
7197 *KIKUCHI Masashi, Steve Henson*
7199 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7200 OpenSSL DTLS client the code can be made to recurse eventually crashing
7203 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7206 *Imre Rad, Steve Henson*
7208 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7209 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7210 client or server. This is potentially exploitable to run arbitrary
7211 code on a vulnerable client or server.
7213 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7215 *Jüri Aedla, Steve Henson*
7217 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7218 are subject to a denial of service attack.
7220 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7221 this issue. [CVE-2014-3470][]
7223 *Felix Gröbert, Ivan Fratric, Steve Henson*
7225 * Harmonize version and its documentation. -f flag is used to display
7228 *mancha <mancha1@zoho.com>*
7230 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7231 in i2d_ECPrivateKey.
7233 *mancha <mancha1@zoho.com>*
7235 * Fix some double frees. These are not thought to be exploitable.
7237 *mancha <mancha1@zoho.com>*
7239 * Fix for the attack described in the paper "Recovering OpenSSL
7240 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7241 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7242 <http://eprint.iacr.org/2014/140>
7244 Thanks to Yuval Yarom and Naomi Benger for discovering this
7245 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7247 *Yuval Yarom and Naomi Benger*
7249 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7251 * Keep original DTLS digest and encryption contexts in retransmission
7252 structures so we can use the previous session parameters if they need
7253 to be resent. [CVE-2013-6450][]
7257 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7258 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7259 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7260 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7261 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7262 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7264 *Rob Stradling, Adam Langley*
7266 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7268 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7270 This addresses the flaw in CBC record processing discovered by
7271 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7272 at: <http://www.isg.rhul.ac.uk/tls/>
7274 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7275 Security Group at Royal Holloway, University of London
7276 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7277 Emilia Käsper for the initial patch.
7280 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7282 * Return an error when checking OCSP signatures when key is NULL.
7283 This fixes a DoS attack. [CVE-2013-0166][]
7287 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7288 the right response is stapled. Also change SSL_get_certificate()
7289 so it returns the certificate actually sent.
7290 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7291 (This is a backport)
7293 *Rob Stradling <rob.stradling@comodo.com>*
7295 * Fix possible deadlock when decoding public keys.
7299 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7301 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7304 * Sanity check record length before skipping explicit IV in DTLS
7307 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7308 fuzzing as a service testing platform.
7313 * Initialise tkeylen properly when encrypting CMS messages.
7314 Thanks to Solar Designer of Openwall for reporting this issue.
7318 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7320 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7321 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7322 in CRYPTO_realloc_clean.
7324 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7325 issue and to Adam Langley <agl@chromium.org> for fixing it.
7328 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7330 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7332 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7333 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7334 content decryption and always return the same error. Note: this attack
7335 needs on average 2^20 messages so it only affects automated senders. The
7336 old behaviour can be re-enabled in the CMS code by setting the
7337 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7338 an MMA defence is not necessary.
7339 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7340 this issue. [CVE-2012-0884][]
7344 * Fix CVE-2011-4619: make sure we really are receiving a
7345 client hello before rejecting multiple SGC restarts. Thanks to
7346 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7350 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7352 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7353 Thanks to Antonio Martin, Enterprise Secure Access Research and
7354 Development, Cisco Systems, Inc. for discovering this bug and
7355 preparing a fix. [CVE-2012-0050][]
7359 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7361 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7362 of the Vaudenay padding oracle attack on CBC mode encryption
7363 which enables an efficient plaintext recovery attack against
7364 the OpenSSL implementation of DTLS. Their attack exploits timing
7365 differences arising during decryption processing. A research
7366 paper describing this attack can be found at:
7367 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7368 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7369 Security Group at Royal Holloway, University of London
7370 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7371 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7372 for preparing the fix. [CVE-2011-4108][]
7374 *Robin Seggelmann, Michael Tuexen*
7376 * Clear bytes used for block padding of SSL 3.0 records.
7379 *Adam Langley (Google)*
7381 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7382 Kadianakis <desnacked@gmail.com> for discovering this issue and
7383 Adam Langley for preparing the fix. [CVE-2011-4619][]
7385 *Adam Langley (Google)*
7387 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7389 *Andrey Kulikov <amdeich@gmail.com>*
7391 * Prevent malformed RFC3779 data triggering an assertion failure.
7392 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7393 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7395 *Rob Austein <sra@hactrn.net>*
7397 * Improved PRNG seeding for VOS.
7399 *Paul Green <Paul.Green@stratus.com>*
7401 * Fix ssl_ciph.c set-up race.
7403 *Adam Langley (Google)*
7405 * Fix spurious failures in ecdsatest.c.
7407 *Emilia Käsper (Google)*
7409 * Fix the BIO_f_buffer() implementation (which was mixing different
7410 interpretations of the `..._len` fields).
7412 *Adam Langley (Google)*
7414 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7415 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7416 threads won't reuse the same blinding coefficients.
7418 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7419 lock to call BN_BLINDING_invert_ex, and avoids one use of
7420 BN_BLINDING_update for each BN_BLINDING structure (previously,
7421 the last update always remained unused).
7423 *Emilia Käsper (Google)*
7425 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7427 *Bob Buckholz (Google)*
7429 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7431 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7432 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7434 *Kaspar Brand <ossl@velox.ch>*
7436 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7437 for multi-threaded use of ECDH. [CVE-2011-3210][]
7439 *Adam Langley (Google)*
7441 * Fix x509_name_ex_d2i memory leak on bad inputs.
7445 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7446 signature public key algorithm by using OID xref utilities instead.
7447 Before this you could only use some ECC ciphersuites with SHA1 only.
7451 * Add protection against ECDSA timing attacks as mentioned in the paper
7452 by Billy Bob Brumley and Nicola Tuveri, see:
7453 <http://eprint.iacr.org/2011/232.pdf>
7455 *Billy Bob Brumley and Nicola Tuveri*
7457 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7459 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7461 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7463 * Fix bug in string printing code: if *any* escaping is enabled we must
7464 escape the escape character (backslash) or the resulting string is
7469 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7471 * Disable code workaround for ancient and obsolete Netscape browsers
7472 and servers: an attacker can use it in a ciphersuite downgrade attack.
7473 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7477 * Fixed J-PAKE implementation error, originally discovered by
7478 Sebastien Martini, further info and confirmation from Stefan
7479 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7483 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7485 * Fix extension code to avoid race conditions which can result in a buffer
7486 overrun vulnerability: resumed sessions must not be modified as they can
7487 be shared by multiple threads. CVE-2010-3864
7491 * Fix WIN32 build system to correctly link an ENGINE directory into
7496 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7498 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7501 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7503 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7505 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7506 context. The operation can be customised via the ctrl mechanism in
7507 case ENGINEs want to include additional functionality.
7511 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7515 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7516 output hashes compatible with older versions of OpenSSL.
7518 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7520 * Fix compression algorithm handling: if resuming a session use the
7521 compression algorithm of the resumed session instead of determining
7522 it from client hello again. Don't allow server to change algorithm.
7526 * Add load_crls() function to apps tidying load_certs() too. Add option
7527 to verify utility to allow additional CRLs to be included.
7531 * Update OCSP request code to permit adding custom headers to the request:
7532 some responders need this.
7536 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7539 *Julia Lawall <julia@diku.dk>*
7541 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7542 needlessly dereferenced structures, used obsolete functions and
7543 didn't handle all updated verify codes correctly.
7547 * Disable MD2 in the default configuration.
7551 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7552 indicate the initial BIO being pushed or popped. This makes it possible
7553 to determine whether the BIO is the one explicitly called or as a result
7554 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7555 it handles reference counts correctly and doesn't zero out the I/O bio
7556 when it is not being explicitly popped. WARNING: applications which
7557 included workarounds for the old buggy behaviour will need to be modified
7558 or they could free up already freed BIOs.
7562 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7563 renaming to all platforms (within the 0.9.8 branch, this was
7564 done conditionally on Netware platforms to avoid a name clash).
7566 *Guenter <lists@gknw.net>*
7568 * Add ECDHE and PSK support to DTLS.
7570 *Michael Tuexen <tuexen@fh-muenster.de>*
7572 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7577 * Add "missing" function EVP_MD_flags() (without this the only way to
7578 retrieve a digest flags is by accessing the structure directly. Update
7579 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7580 or cipher is registered as in the "from" argument. Print out all
7581 registered digests in the dgst usage message instead of manually
7582 attempting to work them out.
7586 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7587 this allows the use of compression and extensions. Change default cipher
7588 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7589 by default unless an application cipher string requests it.
7593 * Alter match criteria in PKCS12_parse(). It used to try to use local
7594 key ids to find matching certificates and keys but some PKCS#12 files
7595 don't follow the (somewhat unwritten) rules and this strategy fails.
7596 Now just gather all certificates together and the first private key
7597 then look for the first certificate that matches the key.
7601 * Support use of registered digest and cipher names for dgst and cipher
7602 commands instead of having to add each one as a special case. So now
7609 openssl dgst -sha256 foo
7611 and this works for ENGINE based algorithms too.
7615 * Update Gost ENGINE to support parameter files.
7617 *Victor B. Wagner <vitus@cryptocom.ru>*
7619 * Support GeneralizedTime in ca utility.
7621 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7623 * Enhance the hash format used for certificate directory links. The new
7624 form uses the canonical encoding (meaning equivalent names will work
7625 even if they aren't identical) and uses SHA1 instead of MD5. This form
7626 is incompatible with the older format and as a result c_rehash should
7627 be used to rebuild symbolic links.
7631 * Make PKCS#8 the default write format for private keys, replacing the
7632 traditional format. This form is standardised, more secure and doesn't
7633 include an implicit MD5 dependency.
7637 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7638 committed to OpenSSL should pass this lot as a minimum.
7642 * Add session ticket override functionality for use by EAP-FAST.
7644 *Jouni Malinen <j@w1.fi>*
7646 * Modify HMAC functions to return a value. Since these can be implemented
7647 in an ENGINE errors can occur.
7651 * Type-checked OBJ_bsearch_ex.
7655 * Type-checked OBJ_bsearch. Also some constification necessitated
7656 by type-checking. Still to come: TXT_DB, bsearch(?),
7657 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7662 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7663 seconds to a tm structure directly, instead of going through OS
7664 specific date routines. This avoids any issues with OS routines such
7665 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7666 and X509_time_adj_ex() to cover the extended range. The existing
7667 X509_time_adj() is still usable and will no longer have any date issues.
7671 * Delta CRL support. New use deltas option which will attempt to locate
7672 and search any appropriate delta CRLs available.
7674 This work was sponsored by Google.
7678 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7679 code and add additional score elements. Validate alternate CRL paths
7680 as part of the CRL checking and indicate a new error "CRL path validation
7681 error" in this case. Applications wanting additional details can use
7682 the verify callback and check the new "parent" field. If this is not
7683 NULL CRL path validation is taking place. Existing applications won't
7684 see this because it requires extended CRL support which is off by
7687 This work was sponsored by Google.
7691 * Support for freshest CRL extension.
7693 This work was sponsored by Google.
7697 * Initial indirect CRL support. Currently only supported in the CRLs
7698 passed directly and not via lookup. Process certificate issuer
7699 CRL entry extension and lookup CRL entries by bother issuer name
7700 and serial number. Check and process CRL issuer entry in IDP extension.
7702 This work was sponsored by Google.
7706 * Add support for distinct certificate and CRL paths. The CRL issuer
7707 certificate is validated separately in this case. Only enabled if
7708 an extended CRL support flag is set: this flag will enable additional
7709 CRL functionality in future.
7711 This work was sponsored by Google.
7715 * Add support for policy mappings extension.
7717 This work was sponsored by Google.
7721 * Fixes to pathlength constraint, self issued certificate handling,
7722 policy processing to align with RFC3280 and PKITS tests.
7724 This work was sponsored by Google.
7728 * Support for name constraints certificate extension. DN, email, DNS
7729 and URI types are currently supported.
7731 This work was sponsored by Google.
7735 * To cater for systems that provide a pointer-based thread ID rather
7736 than numeric, deprecate the current numeric thread ID mechanism and
7737 replace it with a structure and associated callback type. This
7738 mechanism allows a numeric "hash" to be extracted from a thread ID in
7739 either case, and on platforms where pointers are larger than 'long',
7740 mixing is done to help ensure the numeric 'hash' is usable even if it
7741 can't be guaranteed unique. The default mechanism is to use "&errno"
7742 as a pointer-based thread ID to distinguish between threads.
7744 Applications that want to provide their own thread IDs should now use
7745 CRYPTO_THREADID_set_callback() to register a callback that will call
7746 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7748 Note that ERR_remove_state() is now deprecated, because it is tied
7749 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7750 to free the current thread's error state should be replaced by
7751 ERR_remove_thread_state(NULL).
7753 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7754 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7755 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7756 application was previously providing a numeric thread callback that
7757 was inappropriate for distinguishing threads, then uniqueness might
7758 have been obtained with &errno that happened immediately in the
7759 intermediate development versions of OpenSSL; this is no longer the
7760 case, the numeric thread callback will now override the automatic use
7763 *Geoff Thorpe, with help from Bodo Moeller*
7765 * Initial support for different CRL issuing certificates. This covers a
7766 simple case where the self issued certificates in the chain exist and
7767 the real CRL issuer is higher in the existing chain.
7769 This work was sponsored by Google.
7773 * Removed effectively defunct crypto/store from the build.
7777 * Revamp of STACK to provide stronger type-checking. Still to come:
7778 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7779 ASN1_STRING, CONF_VALUE.
7783 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7784 RAM on SSL connections. This option can save about 34k per idle SSL.
7788 * Revamp of LHASH to provide stronger type-checking. Still to come:
7789 STACK, TXT_DB, bsearch, qsort.
7793 * Initial support for Cryptographic Message Syntax (aka CMS) based
7794 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7795 support for data, signedData, compressedData, digestedData and
7796 encryptedData, envelopedData types included. Scripts to check against
7797 RFC4134 examples draft and interop and consistency checks of many
7798 content types and variants.
7802 * Add options to enc utility to support use of zlib compression BIO.
7806 * Extend mk1mf to support importing of options and assembly language
7807 files from Configure script, currently only included in VC-WIN32.
7808 The assembly language rules can now optionally generate the source
7809 files from the associated perl scripts.
7813 * Implement remaining functionality needed to support GOST ciphersuites.
7814 Interop testing has been performed using CryptoPro implementations.
7816 *Victor B. Wagner <vitus@cryptocom.ru>*
7818 * s390x assembler pack.
7822 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7827 * Implement Opaque PRF Input TLS extension as specified in
7828 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7829 official specification yet and no extension type assignment by
7830 IANA exists, this extension (for now) will have to be explicitly
7831 enabled when building OpenSSL by providing the extension number
7832 to use. For example, specify an option
7834 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7836 to the "config" or "Configure" script to enable the extension,
7837 assuming extension number 0x9527 (which is a completely arbitrary
7838 and unofficial assignment based on the MD5 hash of the Internet
7839 Draft). Note that by doing so, you potentially lose
7840 interoperability with other TLS implementations since these might
7841 be using the same extension number for other purposes.
7843 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7844 opaque PRF input value to use in the handshake. This will create
7845 an internal copy of the length-'len' string at 'src', and will
7846 return non-zero for success.
7848 To get more control and flexibility, provide a callback function
7851 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7852 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7856 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7859 Callback function 'cb' will be called in handshakes, and is
7860 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7861 Argument 'arg' is for application purposes (the value as given to
7862 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7863 be provided to the callback function). The callback function
7864 has to return non-zero to report success: usually 1 to use opaque
7865 PRF input just if possible, or 2 to enforce use of the opaque PRF
7866 input. In the latter case, the library will abort the handshake
7867 if opaque PRF input is not successfully negotiated.
7869 Arguments 'peerinput' and 'len' given to the callback function
7870 will always be NULL and 0 in the case of a client. A server will
7871 see the client's opaque PRF input through these variables if
7872 available (NULL and 0 otherwise). Note that if the server
7873 provides an opaque PRF input, the length must be the same as the
7874 length of the client's opaque PRF input.
7876 Note that the callback function will only be called when creating
7877 a new session (session resumption can resume whatever was
7878 previously negotiated), and will not be called in SSL 2.0
7879 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7880 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7881 for applications that need to enforce opaque PRF input.
7885 * Update ssl code to support digests other than SHA1+MD5 for handshake
7888 *Victor B. Wagner <vitus@cryptocom.ru>*
7890 * Add RFC4507 support to OpenSSL. This includes the corrections in
7891 RFC4507bis. The encrypted ticket format is an encrypted encoded
7892 SSL_SESSION structure, that way new session features are automatically
7895 If a client application caches session in an SSL_SESSION structure
7896 support is transparent because tickets are now stored in the encoded
7899 The SSL_CTX structure automatically generates keys for ticket
7900 protection in servers so again support should be possible
7901 with no application modification.
7903 If a client or server wishes to disable RFC4507 support then the option
7904 SSL_OP_NO_TICKET can be set.
7906 Add a TLS extension debugging callback to allow the contents of any client
7907 or server extensions to be examined.
7909 This work was sponsored by Google.
7913 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7914 OpenSSL should now compile cleanly on gcc 4.2
7916 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7918 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7919 support including streaming MAC support: this is required for GOST
7920 ciphersuite support.
7922 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7924 * Add option -stream to use PKCS#7 streaming in smime utility. New
7925 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7926 to output in BER and PEM format.
7930 * Experimental support for use of HMAC via EVP_PKEY interface. This
7931 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
7932 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7933 ENGINE support for HMAC keys which are unextractable. New -mac and
7934 -macopt options to dgst utility.
7938 * New option -sigopt to dgst utility. Update dgst to use
7939 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
7940 alternative signing parameters such as X9.31 or PSS in the dgst
7945 * Change ssl_cipher_apply_rule(), the internal function that does
7946 the work each time a ciphersuite string requests enabling
7947 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7948 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7949 the order of disabled ciphersuites such that those ciphersuites
7950 that most recently went from enabled to disabled not only stay
7951 in order with respect to each other, but also have higher priority
7952 than other disabled ciphersuites the next time ciphersuites are
7955 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7956 the same ciphersuites as with "HIGH" alone, but in a specific
7957 order where the PSK ciphersuites come first (since they are the
7958 most recently disabled ciphersuites when "HIGH" is parsed).
7960 Also, change ssl_create_cipher_list() (using this new
7961 functionality) such that between otherwise identical
7962 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7967 * Change ssl_create_cipher_list() so that it automatically
7968 arranges the ciphersuites in reasonable order before starting
7969 to process the rule string. Thus, the definition for "DEFAULT"
7970 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7971 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7972 This makes it much easier to arrive at a reasonable default order
7973 in applications for which anonymous ciphers are OK (meaning
7974 that you can't actually use DEFAULT).
7976 *Bodo Moeller; suggested by Victor Duchovni*
7978 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7979 processing) into multiple integers instead of setting
7980 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7981 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7982 (These masks as well as the individual bit definitions are hidden
7983 away into the non-exported interface ssl/ssl_locl.h, so this
7984 change to the definition of the SSL_CIPHER structure shouldn't
7985 affect applications.) This give us more bits for each of these
7986 categories, so there is no longer a need to coagulate AES128 and
7987 AES256 into a single algorithm bit, and to coagulate Camellia128
7988 and Camellia256 into a single algorithm bit, which has led to all
7991 Thus, among other things, the kludge introduced in 0.9.7m and
7992 0.9.8e for masking out AES256 independently of AES128 or masking
7993 out Camellia256 independently of AES256 is not needed here in 0.9.9.
7995 With the change, we also introduce new ciphersuite aliases that
7996 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8001 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8002 Use the leftmost N bytes of the signature input if the input is
8003 larger than the prime q (with N being the size in bytes of q).
8007 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8008 it yet and it is largely untested.
8012 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8016 * Initial incomplete changes to avoid need for function casts in OpenSSL
8017 some compilers (gcc 4.2 and later) reject their use. Safestack is
8018 reimplemented. Update ASN1 to avoid use of legacy functions.
8022 * Win32/64 targets are linked with Winsock2.
8026 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8027 to external functions. This can be used to increase CRL handling
8028 efficiency especially when CRLs are very large by (for example) storing
8029 the CRL revoked certificates in a database.
8033 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8034 new CRLs added to a directory can be used. New command line option
8035 -verify_return_error to s_client and s_server. This causes real errors
8036 to be returned by the verify callback instead of carrying on no matter
8037 what. This reflects the way a "real world" verify callback would behave.
8041 * GOST engine, supporting several GOST algorithms and public key formats.
8042 Kindly donated by Cryptocom.
8046 * Partial support for Issuing Distribution Point CRL extension. CRLs
8047 partitioned by DP are handled but no indirect CRL or reason partitioning
8048 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8049 selected via a scoring technique which handles IDP and AKID in CRLs.
8053 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8054 will ultimately be used for all verify operations: this will remove the
8055 X509_STORE dependency on certificate verification and allow alternative
8056 lookup methods. X509_STORE based implementations of these two callbacks.
8060 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8061 Modify get_crl() to find a valid (unexpired) CRL if possible.
8065 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8066 this would be called X509_CRL_cmp() but that name is already used by
8067 a function that just compares CRL issuer names. Cache several CRL
8068 extensions in X509_CRL structure and cache CRLDP in X509.
8072 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8073 this maps equivalent X509_NAME structures into a consistent structure.
8074 Name comparison can then be performed rapidly using memcmp().
8078 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8083 * Allow digests to supply their own micalg string for S/MIME type using
8084 the ctrl EVP_MD_CTRL_MICALG.
8088 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8089 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8090 ctrl. It can then customise the structure before and/or after signing
8095 * New function OBJ_add_sigid() to allow application defined signature OIDs
8096 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8097 to free up any added signature OIDs.
8101 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8102 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8103 digest and cipher tables. New options added to openssl utility:
8104 list-message-digest-algorithms and list-cipher-algorithms.
8108 * Change the array representation of binary polynomials: the list
8109 of degrees of non-zero coefficients is now terminated with -1.
8110 Previously it was terminated with 0, which was also part of the
8111 value; thus, the array representation was not applicable to
8112 polynomials where t^0 has coefficient zero. This change makes
8113 the array representation useful in a more general context.
8117 * Various modifications and fixes to SSL/TLS cipher string
8118 handling. For ECC, the code now distinguishes between fixed ECDH
8119 with RSA certificates on the one hand and with ECDSA certificates
8120 on the other hand, since these are separate ciphersuites. The
8121 unused code for Fortezza ciphersuites has been removed.
8123 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8124 (not "ECDHE"). For consistency with the code for DH
8125 certificates, use of ECDH certificates is now considered ECDH
8126 authentication, not RSA or ECDSA authentication (the latter is
8127 merely the CA's signing algorithm and not actively used in the
8130 The temporary ciphersuite alias "ECCdraft" is no longer
8131 available, and ECC ciphersuites are no longer excluded from "ALL"
8132 and "DEFAULT". The following aliases now exist for RFC 4492
8133 ciphersuites, most of these by analogy with the DH case:
8135 kECDHr - ECDH cert, signed with RSA
8136 kECDHe - ECDH cert, signed with ECDSA
8137 kECDH - ECDH cert (signed with either RSA or ECDSA)
8138 kEECDH - ephemeral ECDH
8139 ECDH - ECDH cert or ephemeral ECDH
8145 AECDH - anonymous ECDH
8146 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8150 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8151 Use correct micalg parameters depending on digest(s) in signed message.
8155 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8156 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8160 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8161 an engine to register a method. Add ENGINE lookups for methods and
8162 functional reference processing.
8166 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8167 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8172 * New -resign option to smime utility. This adds one or more signers
8173 to an existing PKCS#7 signedData structure. Also -md option to use an
8174 alternative message digest algorithm for signing.
8178 * Tidy up PKCS#7 routines and add new functions to make it easier to
8179 create PKCS7 structures containing multiple signers. Update smime
8180 application to support multiple signers.
8184 * New -macalg option to pkcs12 utility to allow setting of an alternative
8189 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8190 Reorganize PBE internals to lookup from a static table using NIDs,
8191 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8192 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8193 PRF which will be automatically used with PBES2.
8197 * Replace the algorithm specific calls to generate keys in "req" with the
8202 * Update PKCS#7 enveloped data routines to use new API. This is now
8203 supported by any public key method supporting the encrypt operation. A
8204 ctrl is added to allow the public key algorithm to examine or modify
8205 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8210 * Add a ctrl to asn1 method to allow a public key algorithm to express
8211 a default digest type to use. In most cases this will be SHA1 but some
8212 algorithms (such as GOST) need to specify an alternative digest. The
8213 return value indicates how strong the preference is 1 means optional and
8214 2 is mandatory (that is it is the only supported type). Modify
8215 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8216 use the default md. Update openssl utilities to use the default digest
8217 type for signing if it is not explicitly indicated.
8221 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8222 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8223 signing method from the key type. This effectively removes the link
8224 between digests and public key types.
8228 * Add an OID cross reference table and utility functions. Its purpose is to
8229 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8230 rsaEncryption. This will allow some of the algorithm specific hackery
8231 needed to use the correct OID to be removed.
8235 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8236 structures for PKCS7_sign(). They are now set up by the relevant public
8241 * Add provisional EC pkey method with support for ECDSA and ECDH.
8245 * Add support for key derivation (agreement) in the API, DH method and
8250 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8251 public and private key formats. As a side effect these add additional
8252 command line functionality not previously available: DSA signatures can be
8253 generated and verified using pkeyutl and DH key support and generation in
8260 *Oliver Tappe <zooey@hirschkaefer.de>*
8262 * New make target "install_html_docs" installs HTML renditions of the
8265 *Oliver Tappe <zooey@hirschkaefer.de>*
8267 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8268 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8269 support key and parameter generation and add initial key generation
8270 functionality for RSA.
8274 * Add functions for main EVP_PKEY_method operations. The undocumented
8275 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8276 `EVP_PKEY_{encrypt,decrypt}_old`.
8280 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8281 key API, doesn't do much yet.
8285 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8286 public key algorithms. New option to openssl utility:
8287 "list-public-key-algorithms" to print out info.
8291 * Implement the Supported Elliptic Curves Extension for
8292 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8296 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8297 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8301 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8302 utilities such as rsa, dsa, dsaparam etc except they process any key
8307 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8308 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8309 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8314 * Initial support for pluggable public key ASN1.
8315 De-spaghettify the public key ASN1 handling. Move public and private
8316 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8317 algorithm specific handling to a single module within the relevant
8318 algorithm directory. Add functions to allow (near) opaque processing
8319 of public and private key structures.
8323 * Implement the Supported Point Formats Extension for
8324 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8328 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8329 for the psk identity [hint] and the psk callback functions to the
8330 SSL_SESSION, SSL and SSL_CTX structure.
8333 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8337 SSL_CTX_use_psk_identity_hint
8338 SSL_get_psk_identity_hint
8339 SSL_get_psk_identity
8340 SSL_use_psk_identity_hint
8342 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8344 * Add RFC 3161 compliant time stamp request creation, response generation
8345 and response verification functionality.
8347 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8349 * Add initial support for TLS extensions, specifically for the server_name
8350 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8351 have new members for a host name. The SSL data structure has an
8352 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8353 stored in that context to allow for session resumption, even after the
8354 SSL has been switched to a new SSL_CTX in reaction to a client's
8355 server_name extension.
8357 New functions (subject to change):
8359 SSL_get_servername()
8360 SSL_get_servername_type()
8363 New CTRL codes and macros (subject to change):
8365 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8366 - SSL_CTX_set_tlsext_servername_callback()
8367 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8368 - SSL_CTX_set_tlsext_servername_arg()
8369 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8371 openssl s_client has a new '-servername ...' option.
8373 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8374 '-key2 ...', '-servername_fatal' (subject to change). This allows
8375 testing the HostName extension for a specific single host name ('-cert'
8376 and '-key' remain fallbacks for handshakes without HostName
8377 negotiation). If the unrecognized_name alert has to be sent, this by
8378 default is a warning; it becomes fatal with the '-servername_fatal'
8381 *Peter Sylvester, Remy Allais, Christophe Renou*
8383 * Whirlpool hash implementation is added.
8387 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8388 bn(64,32). Because of instruction set limitations it doesn't have
8389 any negative impact on performance. This was done mostly in order
8390 to make it possible to share assembler modules, such as bn_mul_mont
8391 implementations, between 32- and 64-bit builds without hassle.
8395 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8396 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8401 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8402 dedicated Montgomery multiplication procedure, is introduced.
8403 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8404 "64-bit" performance on certain 32-bit targets.
8408 * New option SSL_OP_NO_COMP to disable use of compression selectively
8409 in SSL structures. New SSL ctrl to set maximum send fragment size.
8410 Save memory by setting the I/O buffer sizes dynamically instead of
8411 using the maximum available value.
8415 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8416 in addition to the text details.
8420 * Very, very preliminary EXPERIMENTAL support for printing of general
8421 ASN1 structures. This currently produces rather ugly output and doesn't
8422 handle several customised structures at all.
8426 * Integrated support for PVK file format and some related formats such
8427 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8428 these in the 'rsa' and 'dsa' utilities.
8432 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8436 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8437 place for the (very old) "NETSCAPE" format certificates which are now
8438 handled using new ASN1 code equivalents.
8442 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8443 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8444 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8448 * Modify CRL distribution points extension code to print out previously
8449 unsupported fields. Enhance extension setting code to allow setting of
8454 * Add print and set support for Issuing Distribution Point CRL extension.
8458 * Change 'Configure' script to enable Camellia by default.
8465 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8467 * When rejecting SSL/TLS records due to an incorrect version number, never
8468 update s->server with a new major version number. As of
8469 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8470 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8471 the previous behavior could result in a read attempt at NULL when
8472 receiving specific incorrect SSL/TLS records once record payload
8473 protection is active. [CVE-2010-0740][]
8475 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8477 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8478 could be crashed if the relevant tables were not present (e.g. chrooted).
8480 *Tomas Hoger <thoger@redhat.com>*
8482 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8484 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8486 *Martin Olsson, Neel Mehta*
8488 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8489 accommodate for stack sorting, always a write lock!).
8493 * On some versions of WIN32 Heap32Next is very slow. This can cause
8494 excessive delays in the RAND_poll(): over a minute. As a workaround
8495 include a time check in the inner Heap32Next loop too.
8499 * The code that handled flushing of data in SSL/TLS originally used the
8500 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8501 the problem outlined in PR#1949. The fix suggested there however can
8502 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8503 of Apache). So instead simplify the code to flush unconditionally.
8504 This should be fine since flushing with no data to flush is a no op.
8508 * Handle TLS versions 2.0 and later properly and correctly use the
8509 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8510 off ancient servers have a habit of sticking around for a while...
8514 * Modify compression code so it frees up structures without using the
8515 ex_data callbacks. This works around a problem where some applications
8516 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8517 restarting) then use compression (e.g. SSL with compression) later.
8518 This results in significant per-connection memory leaks and
8519 has caused some security issues including CVE-2008-1678 and
8524 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8525 change when encrypting or decrypting.
8529 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8530 connect and renegotiate with servers which do not support RI.
8531 Until RI is more widely deployed this option is enabled by default.
8535 * Add "missing" ssl ctrls to clear options and mode.
8539 * If client attempts to renegotiate and doesn't support RI respond with
8540 a no_renegotiation alert as required by RFC5746. Some renegotiating
8541 TLS clients will continue a connection gracefully when they receive
8542 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8543 waiting for a server hello which it will never receive. Now we treat a
8544 received no_renegotiation alert as a fatal error. This is because
8545 applications requesting a renegotiation might well expect it to succeed
8546 and would have no code in place to handle the server denying it so the
8547 only safe thing to do is to terminate the connection.
8551 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8552 peer supports secure renegotiation and 0 otherwise. Print out peer
8553 renegotiation support in s_client/s_server.
8557 * Replace the highly broken and deprecated SPKAC certification method with
8558 the updated NID creation version. This should correctly handle UTF8.
8562 * Implement RFC5746. Re-enable renegotiation but require the extension
8563 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8564 turns out to be a bad idea. It has been replaced by
8565 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8566 SSL_CTX_set_options(). This is really not recommended unless you
8567 know what you are doing.
8569 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8571 * Fixes to stateless session resumption handling. Use initial_ctx when
8572 issuing and attempting to decrypt tickets in case it has changed during
8573 servername handling. Use a non-zero length session ID when attempting
8574 stateless session resumption: this makes it possible to determine if
8575 a resumption has occurred immediately after receiving server hello
8576 (several places in OpenSSL subtly assume this) instead of later in
8581 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8582 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8583 fixes for a few places where the return code is not checked
8586 *Julia Lawall <julia@diku.dk>*
8588 * Add --strict-warnings option to Configure script to include devteam
8589 warnings in other configurations.
8593 * Add support for --libdir option and LIBDIR variable in makefiles. This
8594 makes it possible to install openssl libraries in locations which
8595 have names other than "lib", for example "/usr/lib64" which some
8598 *Steve Henson, based on patch from Jeremy Utley*
8600 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8601 X690 8.9.12 and can produce some misleading textual output of OIDs.
8603 *Steve Henson, reported by Dan Kaminsky*
8605 * Delete MD2 from algorithm tables. This follows the recommendation in
8606 several standards that it is not used in new applications due to
8607 several cryptographic weaknesses. For binary compatibility reasons
8608 the MD2 API is still compiled in by default.
8612 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8617 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8618 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8621 *Guenter <lists@gknw.net>*
8623 * Fix the server certificate chain building code to use X509_verify_cert(),
8624 it used to have an ad-hoc builder which was unable to cope with anything
8625 other than a simple chain.
8627 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8629 * Don't check self signed certificate signatures in X509_verify_cert()
8630 by default (a flag can override this): it just wastes time without
8631 adding any security. As a useful side effect self signed root CAs
8632 with non-FIPS digests are now usable in FIPS mode.
8636 * In dtls1_process_out_of_seq_message() the check if the current message
8637 is already buffered was missing. For every new message was memory
8638 allocated, allowing an attacker to perform an denial of service attack
8639 with sending out of seq handshake messages until there is no memory
8640 left. Additionally every future message was buffered, even if the
8641 sequence number made no sense and would be part of another handshake.
8642 So only messages with sequence numbers less than 10 in advance will be
8643 buffered. [CVE-2009-1378][]
8645 *Robin Seggelmann, discovered by Daniel Mentz*
8647 * Records are buffered if they arrive with a future epoch to be
8648 processed after finishing the corresponding handshake. There is
8649 currently no limitation to this buffer allowing an attacker to perform
8650 a DOS attack with sending records with future epochs until there is no
8651 memory left. This patch adds the pqueue_size() function to determine
8652 the size of a buffer and limits the record buffer to 100 entries.
8655 *Robin Seggelmann, discovered by Daniel Mentz*
8657 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8658 parent structure is freed. [CVE-2009-1379][]
8662 * Handle non-blocking I/O properly in SSL_shutdown() call.
8664 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8666 * Add `2.5.4.*` OIDs
8668 *Ilya O. <vrghost@gmail.com>*
8670 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8672 * Disable renegotiation completely - this fixes a severe security
8673 problem [CVE-2009-3555][] at the cost of breaking all
8674 renegotiation. Renegotiation can be re-enabled by setting
8675 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8676 run-time. This is really not recommended unless you know what
8681 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8683 * Don't set val to NULL when freeing up structures, it is freed up by
8684 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8685 zeroing past the valid field. [CVE-2009-0789][]
8687 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8689 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8690 checked correctly. This would allow some invalid signed attributes to
8691 appear to verify correctly. [CVE-2009-0591][]
8693 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8695 * Reject UniversalString and BMPString types with invalid lengths. This
8696 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8697 a legal length. [CVE-2009-0590][]
8701 * Set S/MIME signing as the default purpose rather than setting it
8702 unconditionally. This allows applications to override it at the store
8707 * Permit restricted recursion of ASN1 strings. This is needed in practice
8708 to handle some structures.
8712 * Improve efficiency of mem_gets: don't search whole buffer each time
8715 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8717 * New -hex option for openssl rand.
8721 * Print out UTF8String and NumericString when parsing ASN1.
8725 * Support NumericString type for name components.
8729 * Allow CC in the environment to override the automatically chosen
8730 compiler. Note that nothing is done to ensure flags work with the
8735 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8737 * Properly check EVP_VerifyFinal() and similar return values
8740 *Ben Laurie, Bodo Moeller, Google Security Team*
8742 * Enable TLS extensions by default.
8746 * Allow the CHIL engine to be loaded, whether the application is
8747 multithreaded or not. (This does not release the developer from the
8748 obligation to set up the dynamic locking callbacks.)
8750 *Sander Temme <sander@temme.net>*
8752 * Use correct exit code if there is an error in dgst command.
8754 *Steve Henson; problem pointed out by Roland Dirlewanger*
8756 * Tweak Configure so that you need to say "experimental-jpake" to enable
8757 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8761 * Add experimental JPAKE support, including demo authentication in
8762 s_client and s_server.
8766 * Set the comparison function in v3_addr_canonize().
8768 *Rob Austein <sra@hactrn.net>*
8770 * Add support for XMPP STARTTLS in s_client.
8772 *Philip Paeps <philip@freebsd.org>*
8774 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8775 to ensure that even with this option, only ciphersuites in the
8776 server's preference list will be accepted. (Note that the option
8777 applies only when resuming a session, so the earlier behavior was
8778 just about the algorithm choice for symmetric cryptography.)
8782 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8784 * Fix NULL pointer dereference if a DTLS server received
8785 ChangeCipherSpec as first record [CVE-2009-1386][].
8789 * Fix a state transition in s3_srvr.c and d1_srvr.c
8790 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8794 * The fix in 0.9.8c that supposedly got rid of unsafe
8795 double-checked locking was incomplete for RSA blinding,
8796 addressing just one layer of what turns out to have been
8797 doubly unsafe triple-checked locking.
8799 So now fix this for real by retiring the MONT_HELPER macro
8800 in crypto/rsa/rsa_eay.c.
8802 *Bodo Moeller; problem pointed out by Marius Schilder*
8804 * Various precautionary measures:
8806 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8808 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8809 (NB: This would require knowledge of the secret session ticket key
8810 to exploit, in which case you'd be SOL either way.)
8812 - Change bn_nist.c so that it will properly handle input BIGNUMs
8813 outside the expected range.
8815 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8818 *Neel Mehta, Bodo Moeller*
8820 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8821 the load fails. Useful for distros.
8823 *Ben Laurie and the FreeBSD team*
8825 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8829 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8833 * Expand ENGINE to support engine supplied SSL client certificate functions.
8835 This work was sponsored by Logica.
8839 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8840 keystores. Support for SSL/TLS client authentication too.
8841 Not compiled unless enable-capieng specified to Configure.
8843 This work was sponsored by Logica.
8847 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8848 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8849 attribute creation routines such as certificate requests and PKCS#12
8854 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8856 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8857 handshake which could lead to a client crash as found using the
8858 Codenomicon TLS test suite [CVE-2008-1672][]
8860 *Steve Henson, Mark Cox*
8862 * Fix double free in TLS server name extensions which could lead to
8863 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8867 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8869 Clear the error queue to ensure that error entries left from
8870 older function calls do not interfere with the correct operation.
8872 *Lutz Jaenicke, Erik de Castro Lopo*
8874 * Remove root CA certificates of commercial CAs:
8876 The OpenSSL project does not recommend any specific CA and does not
8877 have any policy with respect to including or excluding any CA.
8878 Therefore it does not make any sense to ship an arbitrary selection
8879 of root CA certificates with the OpenSSL software.
8883 * RSA OAEP patches to fix two separate invalid memory reads.
8884 The first one involves inputs when 'lzero' is greater than
8885 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8886 before the beginning of from). The second one involves inputs where
8887 the 'db' section contains nothing but zeroes (there is a one-byte
8888 invalid read after the end of 'db').
8890 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8892 * Partial backport from 0.9.9-dev:
8894 Introduce bn_mul_mont (dedicated Montgomery multiplication
8895 procedure) as a candidate for BIGNUM assembler implementation.
8896 While 0.9.9-dev uses assembler for various architectures, only
8897 x86_64 is available by default here in the 0.9.8 branch, and
8898 32-bit x86 is available through a compile-time setting.
8900 To try the 32-bit x86 assembler implementation, use Configure
8901 option "enable-montasm" (which exists only for this backport).
8903 As "enable-montasm" for 32-bit x86 disclaims code stability
8904 anyway, in this constellation we activate additional code
8905 backported from 0.9.9-dev for further performance improvements,
8906 namely BN_from_montgomery_word. (To enable this otherwise,
8907 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8909 *Andy Polyakov (backport partially by Bodo Moeller)*
8911 * Add TLS session ticket callback. This allows an application to set
8912 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8913 values. This is useful for key rollover for example where several key
8914 sets may exist with different names.
8918 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8919 This was broken until now in 0.9.8 releases, such that the only way
8920 a registered ENGINE could be used (assuming it initialises
8921 successfully on the host) was to explicitly set it as the default
8922 for the relevant algorithms. This is in contradiction with 0.9.7
8923 behaviour and the documentation. With this fix, when an ENGINE is
8924 registered into a given algorithm's table of implementations, the
8925 'uptodate' flag is reset so that auto-discovery will be used next
8926 time a new context for that algorithm attempts to select an
8929 *Ian Lister (tweaked by Geoff Thorpe)*
8931 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8932 implementation in the following ways:
8934 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8937 Lack of BER streaming support means one pass streaming processing is
8938 only supported if data is detached: setting the streaming flag is
8939 ignored for embedded content.
8941 CMS support is disabled by default and must be explicitly enabled
8942 with the enable-cms configuration option.
8946 * Update the GMP engine glue to do direct copies between BIGNUM and
8947 mpz_t when openssl and GMP use the same limb size. Otherwise the
8948 existing "conversion via a text string export" trick is still used.
8950 *Paul Sheer <paulsheer@gmail.com>*
8952 * Zlib compression BIO. This is a filter BIO which compressed and
8953 uncompresses any data passed through it.
8957 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8958 RFC3394 compatible AES key wrapping.
8962 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8963 sets string data without copying. X509_ALGOR_set0() and
8964 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8965 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8966 from an X509_ATTRIBUTE structure optionally checking it occurs only
8967 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8972 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8973 to get the expected BN_FLG_CONSTTIME behavior.
8975 *Bodo Moeller (Google)*
8979 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8980 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8981 - added some more tests to do_tests.pl
8982 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8983 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8984 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8985 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8986 - various changes to netware.pl to enable gcc-cross builds on Win32
8988 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8989 - various changes to fix missing prototype warnings
8990 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8991 - added AES, WHIRLPOOL and CPUID assembler code to build files
8992 - added missing AES assembler make rules to mk1mf.pl
8993 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
8995 *Guenter Knauf <eflash@gmx.net>*
8997 * Implement certificate status request TLS extension defined in RFC3546.
8998 A client can set the appropriate parameters and receive the encoded
8999 OCSP response via a callback. A server can query the supplied parameters
9000 and set the encoded OCSP response in the callback. Add simplified examples
9001 to s_client and s_server.
9005 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9008 + Binary incompatibility of ssl_ctx_st structure
9009 + DTLS interoperation with non-compliant servers
9010 + Don't call get_session_cb() without proposed session
9011 + Fix ia64 assembler code
9013 *Andy Polyakov, Steve Henson*
9015 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9017 * DTLS Handshake overhaul. There were longstanding issues with
9018 OpenSSL DTLS implementation, which were making it impossible for
9019 RFC 4347 compliant client to communicate with OpenSSL server.
9020 Unfortunately just fixing these incompatibilities would "cut off"
9021 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9022 server keeps tolerating non RFC compliant syntax. The opposite is
9023 not true, 0.9.8f client can not communicate with earlier server.
9024 This update even addresses CVE-2007-4995.
9028 * Changes to avoid need for function casts in OpenSSL: some compilers
9029 (gcc 4.2 and later) reject their use.
9030 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9033 * Add RFC4507 support to OpenSSL. This includes the corrections in
9034 RFC4507bis. The encrypted ticket format is an encrypted encoded
9035 SSL_SESSION structure, that way new session features are automatically
9038 If a client application caches session in an SSL_SESSION structure
9039 support is transparent because tickets are now stored in the encoded
9042 The SSL_CTX structure automatically generates keys for ticket
9043 protection in servers so again support should be possible
9044 with no application modification.
9046 If a client or server wishes to disable RFC4507 support then the option
9047 SSL_OP_NO_TICKET can be set.
9049 Add a TLS extension debugging callback to allow the contents of any client
9050 or server extensions to be examined.
9052 This work was sponsored by Google.
9056 * Add initial support for TLS extensions, specifically for the server_name
9057 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9058 have new members for a host name. The SSL data structure has an
9059 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9060 stored in that context to allow for session resumption, even after the
9061 SSL has been switched to a new SSL_CTX in reaction to a client's
9062 server_name extension.
9064 New functions (subject to change):
9066 SSL_get_servername()
9067 SSL_get_servername_type()
9070 New CTRL codes and macros (subject to change):
9072 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9073 - SSL_CTX_set_tlsext_servername_callback()
9074 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9075 - SSL_CTX_set_tlsext_servername_arg()
9076 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9078 openssl s_client has a new '-servername ...' option.
9080 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9081 '-key2 ...', '-servername_fatal' (subject to change). This allows
9082 testing the HostName extension for a specific single host name ('-cert'
9083 and '-key' remain fallbacks for handshakes without HostName
9084 negotiation). If the unrecognized_name alert has to be sent, this by
9085 default is a warning; it becomes fatal with the '-servername_fatal'
9088 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9090 * Add AES and SSE2 assembly language support to VC++ build.
9094 * Mitigate attack on final subtraction in Montgomery reduction.
9098 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9099 (which previously caused an internal error).
9103 * Squeeze another 10% out of IGE mode when in != out.
9107 * AES IGE mode speedup.
9109 *Dean Gaudet (Google)*
9111 * Add the Korean symmetric 128-bit cipher SEED (see
9112 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9113 add SEED ciphersuites from RFC 4162:
9115 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9116 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9117 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9118 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9120 To minimize changes between patchlevels in the OpenSSL 0.9.8
9121 series, SEED remains excluded from compilation unless OpenSSL
9122 is configured with 'enable-seed'.
9124 *KISA, Bodo Moeller*
9126 * Mitigate branch prediction attacks, which can be practical if a
9127 single processor is shared, allowing a spy process to extract
9128 information. For detailed background information, see
9129 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9130 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9131 and Necessary Software Countermeasures"). The core of the change
9132 are new versions BN_div_no_branch() and
9133 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9134 respectively, which are slower, but avoid the security-relevant
9135 conditional branches. These are automatically called by BN_div()
9136 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9137 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9138 remove a conditional branch.
9140 BN_FLG_CONSTTIME is the new name for the previous
9141 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9142 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9143 in the exponent causes BN_mod_exp_mont() to use the alternative
9144 implementation in BN_mod_exp_mont_consttime().) The old name
9145 remains as a deprecated alias.
9147 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9148 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9149 constant-time implementations for more than just exponentiation.
9150 Here too the old name is kept as a deprecated alias.
9152 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9153 the BN_BLINDING structure gets an independent copy of the
9154 modulus. This means that the previous `BIGNUM *m` argument to
9155 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9156 essentially becomes `const BIGNUM *m`, although we can't actually
9157 change this in the header file before 0.9.9. It allows
9158 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9159 enable BN_FLG_CONSTTIME.
9161 *Matthew D Wood (Intel Corp)*
9163 * In the SSL/TLS server implementation, be strict about session ID
9164 context matching (which matters if an application uses a single
9165 external cache for different purposes). Previously,
9166 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9167 set. This did ensure strict client verification, but meant that,
9168 with applications using a single external cache for quite
9169 different requirements, clients could circumvent ciphersuite
9170 restrictions for a given session ID context by starting a session
9171 in a different context.
9175 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9176 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9177 authentication-only ciphersuites.
9181 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9182 not complete and could lead to a possible single byte overflow
9183 [CVE-2007-5135][] [Ben Laurie]
9185 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9187 * Since AES128 and AES256 (and similarly Camellia128 and
9188 Camellia256) share a single mask bit in the logic of
9189 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9190 kludge to work properly if AES128 is available and AES256 isn't
9191 (or if Camellia128 is available and Camellia256 isn't).
9195 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9196 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9197 When a point or a seed is encoded in a BIT STRING, we need to
9198 prevent the removal of trailing zero bits to get the proper DER
9199 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9200 of a NamedBitList, for which trailing 0 bits need to be removed.)
9204 * Have SSL/TLS server implementation tolerate "mismatched" record
9205 protocol version while receiving ClientHello even if the
9206 ClientHello is fragmented. (The server can't insist on the
9207 particular protocol version it has chosen before the ServerHello
9208 message has informed the client about his choice.)
9212 * Add RFC 3779 support.
9214 *Rob Austein for ARIN, Ben Laurie*
9216 * Load error codes if they are not already present instead of using a
9217 static variable. This allows them to be cleanly unloaded and reloaded.
9218 Improve header file function name parsing.
9222 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9223 or CAPABILITY handshake as required by RFCs.
9227 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9229 * Introduce limits to prevent malicious keys being able to
9230 cause a denial of service. [CVE-2006-2940][]
9232 *Steve Henson, Bodo Moeller*
9234 * Fix ASN.1 parsing of certain invalid structures that can result
9235 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9237 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9238 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9240 * Fix SSL client code which could crash if connecting to a
9241 malicious SSLv2 server. [CVE-2006-4343][]
9243 *Tavis Ormandy and Will Drewry, Google Security Team*
9245 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9246 match only those. Before that, "AES256-SHA" would be interpreted
9247 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9248 the same strength classification in 0.9.7h) as we currently only
9249 have a single AES bit in the ciphersuite description bitmap.
9250 That change, however, also applied to ciphersuite strings such as
9251 "RC4-MD5" that intentionally matched multiple ciphersuites --
9252 namely, SSL 2.0 ciphersuites in addition to the more common ones
9253 from SSL 3.0/TLS 1.0.
9255 So we change the selection algorithm again: Naming an explicit
9256 ciphersuite selects this one ciphersuite, and any other similar
9257 ciphersuite (same bitmap) from *other* protocol versions.
9258 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9259 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9261 Since SSL 2.0 does not have any ciphersuites for which the
9262 128/256 bit distinction would be relevant, this works for now.
9263 The proper fix will be to use different bits for AES128 and
9264 AES256, which would have avoided the problems from the beginning;
9265 however, bits are scarce, so we can only do this in a new release
9266 (not just a patchlevel) when we can change the SSL_CIPHER
9267 definition to split the single 'unsigned long mask' bitmap into
9268 multiple values to extend the available space.
9272 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9274 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9275 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9277 * Add AES IGE and biIGE modes.
9281 * Change the Unix randomness entropy gathering to use poll() when
9282 possible instead of select(), since the latter has some
9283 undesirable limitations.
9285 *Darryl Miles via Richard Levitte and Bodo Moeller*
9287 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9288 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9289 cannot be implicitly activated as part of, e.g., the "AES" alias.
9290 However, please upgrade to OpenSSL 0.9.9[-dev] for
9291 non-experimental use of the ECC ciphersuites to get TLS extension
9292 support, which is required for curve and point format negotiation
9293 to avoid potential handshake problems.
9297 * Disable rogue ciphersuites:
9299 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9300 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9301 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9303 The latter two were purportedly from
9304 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9307 Also deactivate the remaining ciphersuites from
9308 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9309 unofficial, and the ID has long expired.
9313 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9314 dual-core machines) and other potential thread-safety issues.
9318 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9319 versions), which is now available for royalty-free use
9320 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9321 Also, add Camellia TLS ciphersuites from RFC 4132.
9323 To minimize changes between patchlevels in the OpenSSL 0.9.8
9324 series, Camellia remains excluded from compilation unless OpenSSL
9325 is configured with 'enable-camellia'.
9329 * Disable the padding bug check when compression is in use. The padding
9330 bug check assumes the first packet is of even length, this is not
9331 necessarily true if compression is enabled and can result in false
9332 positives causing handshake failure. The actual bug test is ancient
9333 code so it is hoped that implementations will either have fixed it by
9334 now or any which still have the bug do not support compression.
9338 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9340 * When applying a cipher rule check to see if string match is an explicit
9341 cipher suite and only match that one cipher suite if it is.
9345 * Link in manifests for VC++ if needed.
9347 *Austin Ziegler <halostatue@gmail.com>*
9349 * Update support for ECC-based TLS ciphersuites according to
9350 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9351 TLS extensions, which are supported starting with the 0.9.9
9352 branch, not in the OpenSSL 0.9.8 branch).
9356 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9357 opaque EVP_CIPHER_CTX handling.
9361 * Fixes and enhancements to zlib compression code. We now only use
9362 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9363 to conform with the standards mentioned here:
9364 <http://www.zlib.net/DLL_FAQ.txt>
9365 Static zlib linking now works on Windows and the new --with-zlib-include
9366 --with-zlib-lib options to Configure can be used to supply the location
9367 of the headers and library. Gracefully handle case where zlib library
9372 * Several fixes and enhancements to the OID generation code. The old code
9373 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9374 handle numbers larger than ULONG_MAX, truncated printing and had a
9375 non standard OBJ_obj2txt() behaviour.
9379 * Add support for building of engines under engine/ as shared libraries
9380 under VC++ build system.
9384 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9385 Hopefully, we will not see any false combination of paths any more.
9389 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9391 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9392 (part of SSL_OP_ALL). This option used to disable the
9393 countermeasure against man-in-the-middle protocol-version
9394 rollback in the SSL 2.0 server implementation, which is a bad
9395 idea. [CVE-2005-2969][]
9397 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9398 for Information Security, National Institute of Advanced Industrial
9399 Science and Technology [AIST], Japan)*
9401 * Add two function to clear and return the verify parameter flags.
9405 * Keep cipherlists sorted in the source instead of sorting them at
9406 runtime, thus removing the need for a lock.
9410 * Avoid some small subgroup attacks in Diffie-Hellman.
9412 *Nick Mathewson and Ben Laurie*
9414 * Add functions for well-known primes.
9418 * Extended Windows CE support.
9420 *Satoshi Nakamura and Andy Polyakov*
9422 * Initialize SSL_METHOD structures at compile time instead of during
9423 runtime, thus removing the need for a lock.
9427 * Make PKCS7_decrypt() work even if no certificate is supplied by
9428 attempting to decrypt each encrypted key in turn. Add support to
9433 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9435 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9438 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9442 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9443 key into the same file any more.
9447 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9451 * Add -utf8 command line and config file option to 'ca'.
9453 *Stefan <stf@udoma.org*
9455 * Removed the macro des_crypt(), as it seems to conflict with some
9456 libraries. Use DES_crypt().
9460 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9461 involves renaming the source and generated shared-libs for
9462 both. The engines will accept the corrected or legacy ids
9463 ('ncipher' and '4758_cca' respectively) when binding. NB,
9464 this only applies when building 'shared'.
9466 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9468 * Add attribute functions to EVP_PKEY structure. Modify
9469 PKCS12_create() to recognize a CSP name attribute and
9470 use it. Make -CSP option work again in pkcs12 utility.
9474 * Add new functionality to the bn blinding code:
9475 - automatic re-creation of the BN_BLINDING parameters after
9476 a fixed number of uses (currently 32)
9477 - add new function for parameter creation
9478 - introduce flags to control the update behaviour of the
9479 BN_BLINDING parameters
9480 - hide BN_BLINDING structure
9481 Add a second BN_BLINDING slot to the RSA structure to improve
9482 performance when a single RSA object is shared among several
9487 * Add support for DTLS.
9489 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9491 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9492 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9496 * Remove buggy and incomplete DH cert support from
9497 ssl/ssl_rsa.c and ssl/s3_both.c
9501 * Use SHA-1 instead of MD5 as the default digest algorithm for
9502 the apps/openssl applications.
9506 * Compile clean with "-Wall -Wmissing-prototypes
9507 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9508 DEBUG_SAFESTACK must also be set.
9512 * Change ./Configure so that certain algorithms can be disabled by default.
9513 The new counterpiece to "no-xxx" is "enable-xxx".
9515 The patented RC5 and MDC2 algorithms will now be disabled unless
9516 "enable-rc5" and "enable-mdc2", respectively, are specified.
9518 (IDEA remains enabled despite being patented. This is because IDEA
9519 is frequently required for interoperability, and there is no license
9520 fee for non-commercial use. As before, "no-idea" can be used to
9521 avoid this algorithm.)
9525 * Add processing of proxy certificates (see RFC 3820). This work was
9526 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9527 EGEE (Enabling Grids for E-science in Europe).
9531 * RC4 performance overhaul on modern architectures/implementations, such
9532 as Intel P4, IA-64 and AMD64.
9536 * New utility extract-section.pl. This can be used specify an alternative
9537 section number in a pod file instead of having to treat each file as
9538 a separate case in Makefile. This can be done by adding two lines to the
9541 =for comment openssl_section:XXX
9543 The blank line is mandatory.
9547 * New arguments -certform, -keyform and -pass for s_client and s_server
9548 to allow alternative format key and certificate files and passphrase
9553 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9554 update associated structures and add various utility functions.
9556 Add new policy related verify parameters, include policy checking in
9557 standard verify code. Enhance 'smime' application with extra parameters
9558 to support policy checking and print out.
9562 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9563 Nehemiah processors. These extensions support AES encryption in hardware
9564 as well as RNG (though RNG support is currently disabled).
9566 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9568 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9572 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9574 *Andy Polyakov and a number of other people*
9576 * Improved PowerPC platform support. Most notably BIGNUM assembler
9577 implementation contributed by IBM.
9579 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9581 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9582 exponent rather than 'unsigned long'. There is a corresponding change to
9583 the new 'rsa_keygen' element of the RSA_METHOD structure.
9585 *Jelte Jansen, Geoff Thorpe*
9587 * Functionality for creating the initial serial number file is now
9588 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9590 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9591 number file to 1, which is bound to cause problems. To avoid
9592 the problems while respecting compatibility between different 0.9.7
9593 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9594 CA.pl for serial number initialization. With the new release 0.9.8,
9595 we can fix the problem directly in the 'ca' utility.)
9599 * Reduced header interdependencies by declaring more opaque objects in
9600 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9601 give fewer recursive includes, which could break lazy source code - so
9602 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9603 developers should define this symbol when building and using openssl to
9604 ensure they track the recommended behaviour, interfaces, [etc], but
9605 backwards-compatible behaviour prevails when this isn't defined.
9609 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9613 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9614 This will generate a random key of the appropriate length based on the
9615 cipher context. The EVP_CIPHER can provide its own random key generation
9616 routine to support keys of a specific form. This is used in the des and
9617 3des routines to generate a key of the correct parity. Update S/MIME
9618 code to use new functions and hence generate correct parity DES keys.
9619 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9620 valid (weak or incorrect parity).
9624 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9625 as looking them up. This is useful when the verified structure may contain
9626 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9627 present unless the new PKCS7_NO_CRL flag is asserted.
9631 * Extend ASN1 oid configuration module. It now additionally accepts the
9634 shortName = some long name, 1.2.3.4
9638 * Reimplemented the BN_CTX implementation. There is now no more static
9639 limitation on the number of variables it can handle nor the depth of the
9640 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9641 information can now expand as required, and rather than having a single
9642 static array of bignums, BN_CTX now uses a linked-list of such arrays
9643 allowing it to expand on demand whilst maintaining the usefulness of
9644 BN_CTX's "bundling".
9648 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9649 to allow all RSA operations to function using a single BN_CTX.
9653 * Preliminary support for certificate policy evaluation and checking. This
9654 is initially intended to pass the tests outlined in "Conformance Testing
9655 of Relying Party Client Certificate Path Processing Logic" v1.07.
9659 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9660 remained unused and not that useful. A variety of other little bignum
9661 tweaks and fixes have also been made continuing on from the audit (see
9666 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9667 associated ASN1, EVP and SSL functions and old ASN1 macros.
9671 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9672 and this should never fail. So the return value from the use of
9673 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9674 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9678 * BN_CTX_get() should return zero-valued bignums, providing the same
9679 initialised value as BN_new().
9681 *Geoff Thorpe, suggested by Ulf Möller*
9683 * Support for inhibitAnyPolicy certificate extension.
9687 * An audit of the BIGNUM code is underway, for which debugging code is
9688 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9689 is considered valid when processing BIGNUMs, and causes execution to
9690 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9691 further steps are taken to deliberately pollute unused data in BIGNUM
9692 structures to try and expose faulty code further on. For now, openssl will
9693 (in its default mode of operation) continue to tolerate the inconsistent
9694 forms that it has tolerated in the past, but authors and packagers should
9695 consider trying openssl and their own applications when compiled with
9696 these debugging symbols defined. It will help highlight potential bugs in
9697 their own code, and will improve the test coverage for OpenSSL itself. At
9698 some point, these tighter rules will become openssl's default to improve
9699 maintainability, though the assert()s and other overheads will remain only
9700 in debugging configurations. See bn.h for more details.
9702 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9704 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9705 that can only be obtained through BN_CTX_new() (which implicitly
9706 initialises it). The presence of this function only made it possible
9707 to overwrite an existing structure (and cause memory leaks).
9711 * Because of the callback-based approach for implementing LHASH as a
9712 template type, lh_insert() adds opaque objects to hash-tables and
9713 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9714 to clean up those corresponding objects before destroying the hash table
9715 (and losing the object pointers). So some over-zealous constifications in
9716 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9717 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9718 prototyped to have "const" restrictions on the object pointers they are
9719 given (and so aren't required to cast them away any more).
9723 * The tmdiff.h API was so ugly and minimal that our own timing utility
9724 (speed) prefers to use its own implementation. The two implementations
9725 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9726 its object type properly exposed (MS_TM) instead of casting to/from
9727 `char *`. This may still change yet if someone realises MS_TM and
9729 aren't necessarily the greatest nomenclatures - but this is what was used
9730 internally to the implementation so I've used that for now.
9734 * Ensure that deprecated functions do not get compiled when
9735 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9736 the self-tests were still using deprecated key-generation functions so
9737 these have been updated also.
9741 * Reorganise PKCS#7 code to separate the digest location functionality
9742 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9743 New function PKCS7_set_digest() to set the digest type for PKCS#7
9744 digestedData type. Add additional code to correctly generate the
9745 digestedData type and add support for this type in PKCS7 initialization
9750 * New function PKCS7_set0_type_other() this initializes a PKCS7
9751 structure of type "other".
9755 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9756 sure the loop does correctly stop and breaking ("division by zero")
9757 modulus operations are not performed. The (pre-generated) prime
9758 table crypto/bn/bn_prime.h was already correct, but it could not be
9759 re-generated on some platforms because of the "division by zero"
9760 situation in the script.
9762 *Ralf S. Engelschall*
9764 * Update support for ECC-based TLS ciphersuites according to
9765 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9766 SHA-1 now is only used for "small" curves (where the
9767 representation of a field element takes up to 24 bytes); for
9768 larger curves, the field element resulting from ECDH is directly
9769 used as premaster secret.
9771 *Douglas Stebila (Sun Microsystems Laboratories)*
9773 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9774 curve secp160r1 to the tests.
9776 *Douglas Stebila (Sun Microsystems Laboratories)*
9778 * Add the possibility to load symbols globally with DSO.
9780 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9782 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9783 control of the error stack.
9787 * Add support for STORE in ENGINE.
9791 * Add the STORE type. The intention is to provide a common interface
9792 to certificate and key stores, be they simple file-based stores, or
9793 HSM-type store, or LDAP stores, or...
9794 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9798 * Add a generic structure called OPENSSL_ITEM. This can be used to
9799 pass a list of arguments to any function as well as provide a way
9800 for a function to pass data back to the caller.
9804 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9805 works like BUF_strdup() but can be used to duplicate a portion of
9806 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9811 * Add the function sk_find_ex() which works like sk_find(), but will
9812 return an index to an element even if an exact match couldn't be
9813 found. The index is guaranteed to point at the element where the
9814 searched-for key would be inserted to preserve sorting order.
9818 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9819 takes an extra flags argument for optional functionality. Currently,
9820 the following flags are defined:
9822 OBJ_BSEARCH_VALUE_ON_NOMATCH
9823 This one gets OBJ_bsearch_ex() to return a pointer to the first
9824 element where the comparing function returns a negative or zero
9827 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9828 This one gets OBJ_bsearch_ex() to return a pointer to the first
9829 element where the comparing function returns zero. This is useful
9830 if there are more than one element where the comparing function
9835 * Make it possible to create self-signed certificates with 'openssl ca'
9836 in such a way that the self-signed certificate becomes part of the
9837 CA database and uses the same mechanisms for serial number generation
9838 as all other certificate signing. The new flag '-selfsign' enables
9839 this functionality. Adapt CA.sh and CA.pl.in.
9843 * Add functionality to check the public key of a certificate request
9844 against a given private. This is useful to check that a certificate
9845 request can be signed by that key (self-signing).
9849 * Make it possible to have multiple active certificates with the same
9850 subject in the CA index file. This is done only if the keyword
9851 'unique_subject' is set to 'no' in the main CA section (default
9852 if 'CA_default') of the configuration file. The value is saved
9853 with the database itself in a separate index attribute file,
9854 named like the index file with '.attr' appended to the name.
9858 * Generate multi-valued AVAs using '+' notation in config files for
9863 * Support for nameConstraints certificate extension.
9867 * Support for policyConstraints certificate extension.
9871 * Support for policyMappings certificate extension.
9875 * Make sure the default DSA_METHOD implementation only uses its
9876 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9877 and change its own handlers to be NULL so as to remove unnecessary
9878 indirection. This lets alternative implementations fallback to the
9879 default implementation more easily.
9883 * Support for directoryName in GeneralName related extensions
9888 * Make it possible to link applications using Makefile.shared.
9889 Make that possible even when linking against static libraries!
9893 * Support for single pass processing for S/MIME signing. This now
9894 means that S/MIME signing can be done from a pipe, in addition
9895 cleartext signing (multipart/signed type) is effectively streaming
9896 and the signed data does not need to be all held in memory.
9898 This is done with a new flag PKCS7_STREAM. When this flag is set
9899 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9900 is done after the data is output (and digests calculated) in
9901 SMIME_write_PKCS7().
9905 * Add full support for -rpath/-R, both in shared libraries and
9906 applications, at least on the platforms where it's known how
9911 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9912 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9913 will now compute a table of multiples of the generator that
9914 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9915 faster (notably in the case of a single point multiplication,
9916 scalar * generator).
9918 *Nils Larsch, Bodo Moeller*
9920 * IPv6 support for certificate extensions. The various extensions
9921 which use the IP:a.b.c.d can now take IPv6 addresses using the
9922 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9927 * Added an ENGINE that implements RSA by performing private key
9928 exponentiations with the GMP library. The conversions to and from
9929 GMP's mpz_t format aren't optimised nor are any montgomery forms
9930 cached, and on x86 it appears OpenSSL's own performance has caught up.
9931 However there are likely to be other architectures where GMP could
9932 provide a boost. This ENGINE is not built in by default, but it can be
9933 specified at Configure time and should be accompanied by the necessary
9934 linker additions, eg;
9935 ./config -DOPENSSL_USE_GMP -lgmp
9939 * "openssl engine" will not display ENGINE/DSO load failure errors when
9940 testing availability of engines with "-t" - the old behaviour is
9941 produced by increasing the feature's verbosity with "-tt".
9945 * ECDSA routines: under certain error conditions uninitialized BN objects
9946 could be freed. Solution: make sure initialization is performed early
9947 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9952 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9953 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9954 software implementations. For DSA and DH, parameter generation can
9955 also be overridden by providing the appropriate method callbacks.
9959 * Change the "progress" mechanism used in key-generation and
9960 primality testing to functions that take a new BN_GENCB pointer in
9961 place of callback/argument pairs. The new API functions have `_ex`
9962 postfixes and the older functions are reimplemented as wrappers for
9963 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9964 declarations of the old functions to help (graceful) attempts to
9965 migrate to the new functions. Also, the new key-generation API
9966 functions operate on a caller-supplied key-structure and return
9967 success/failure rather than returning a key or NULL - this is to
9968 help make "keygen" another member function of RSA_METHOD etc.
9970 Example for using the new callback interface:
9972 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9976 BN_GENCB_set(&my_cb, my_callback, my_arg);
9978 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9979 /* For the meaning of a, b in calls to my_callback(), see the
9980 * documentation of the function that calls the callback.
9981 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9982 * my_callback should return 1 if it wants BN_is_prime_ex()
9983 * to continue, or 0 to stop.
9988 * Change the ZLIB compression method to be stateful, and make it
9989 available to TLS with the number defined in
9990 draft-ietf-tls-compression-04.txt.
9994 * Add the ASN.1 structures and functions for CertificatePair, which
9995 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
9997 CertificatePair ::= SEQUENCE {
9998 forward [0] Certificate OPTIONAL,
9999 reverse [1] Certificate OPTIONAL,
10000 -- at least one of the pair shall be present -- }
10002 Also implement the PEM functions to read and write certificate
10003 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10005 This needed to be defined, mostly for the sake of the LDAP
10006 attribute crossCertificatePair, but may prove useful elsewhere as
10011 * Make it possible to inhibit symlinking of shared libraries in
10012 Makefile.shared, for Cygwin's sake.
10016 * Extend the BIGNUM API by creating a function
10017 void BN_set_negative(BIGNUM *a, int neg);
10018 and a macro that behave like
10019 int BN_is_negative(const BIGNUM *a);
10021 to avoid the need to access 'a->neg' directly in applications.
10025 * Implement fast modular reduction for pseudo-Mersenne primes
10026 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10027 EC_GROUP_new_curve_GFp() will now automatically use this
10030 *Nils Larsch <nla@trustcenter.de>*
10032 * Add new lock type (CRYPTO_LOCK_BN).
10036 * Change the ENGINE framework to automatically load engines
10037 dynamically from specific directories unless they could be
10038 found to already be built in or loaded. Move all the
10039 current engines except for the cryptodev one to a new
10040 directory engines/.
10041 The engines in engines/ are built as shared libraries if
10042 the "shared" options was given to ./Configure or ./config.
10043 Otherwise, they are inserted in libcrypto.a.
10044 /usr/local/ssl/engines is the default directory for dynamic
10045 engines, but that can be overridden at configure time through
10046 the usual use of --prefix and/or --openssldir, and at run
10047 time with the environment variable OPENSSL_ENGINES.
10049 *Geoff Thorpe and Richard Levitte*
10051 * Add Makefile.shared, a helper makefile to build shared
10052 libraries. Adapt Makefile.org.
10056 * Add version info to Win32 DLLs.
10058 *Peter 'Luna' Runestig" <peter@runestig.com>*
10060 * Add new 'medium level' PKCS#12 API. Certificates and keys
10061 can be added using this API to created arbitrary PKCS#12
10062 files while avoiding the low level API.
10064 New options to PKCS12_create(), key or cert can be NULL and
10065 will then be omitted from the output file. The encryption
10066 algorithm NIDs can be set to -1 for no encryption, the mac
10067 iteration count can be set to 0 to omit the mac.
10069 Enhance pkcs12 utility by making the -nokeys and -nocerts
10070 options work when creating a PKCS#12 file. New option -nomac
10071 to omit the mac, NONE can be set for an encryption algorithm.
10072 New code is modified to use the enhanced PKCS12_create()
10073 instead of the low level API.
10077 * Extend ASN1 encoder to support indefinite length constructed
10078 encoding. This can output sequences tags and octet strings in
10079 this form. Modify pk7_asn1.c to support indefinite length
10080 encoding. This is experimental and needs additional code to
10081 be useful, such as an ASN1 bio and some enhanced streaming
10084 Extend template encode functionality so that tagging is passed
10085 down to the template encoder.
10089 * Let 'openssl req' fail if an argument to '-newkey' is not
10090 recognized instead of using RSA as a default.
10094 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10095 As these are not official, they are not included in "ALL";
10096 the "ECCdraft" ciphersuite group alias can be used to select them.
10098 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10100 * Add ECDH engine support.
10102 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10104 * Add ECDH in new directory crypto/ecdh/.
10106 *Douglas Stebila (Sun Microsystems Laboratories)*
10108 * Let BN_rand_range() abort with an error after 100 iterations
10109 without success (which indicates a broken PRNG).
10113 * Change BN_mod_sqrt() so that it verifies that the input value
10114 is really the square of the return value. (Previously,
10115 BN_mod_sqrt would show GIGO behaviour.)
10119 * Add named elliptic curves over binary fields from X9.62, SECG,
10120 and WAP/WTLS; add OIDs that were still missing.
10122 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10124 * Extend the EC library for elliptic curves over binary fields
10125 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10128 EC_GF2m_simple_method
10132 EC_GROUP_new_curve_GF2m
10133 EC_GROUP_set_curve_GF2m
10134 EC_GROUP_get_curve_GF2m
10135 EC_POINT_set_affine_coordinates_GF2m
10136 EC_POINT_get_affine_coordinates_GF2m
10137 EC_POINT_set_compressed_coordinates_GF2m
10139 Point compression for binary fields is disabled by default for
10140 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10143 As binary polynomials are represented as BIGNUMs, various members
10144 of the EC_GROUP and EC_POINT data structures can be shared
10145 between the implementations for prime fields and binary fields;
10146 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10147 are essentially identical to their `..._GFp` counterparts.
10148 (For simplicity, the `..._GFp` prefix has been dropped from
10149 various internal method names.)
10151 An internal 'field_div' method (similar to 'field_mul' and
10152 'field_sqr') has been added; this is used only for binary fields.
10154 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10156 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10157 through methods ('mul', 'precompute_mult').
10159 The generic implementations (now internally called 'ec_wNAF_mul'
10160 and 'ec_wNAF_precomputed_mult') remain the default if these
10161 methods are undefined.
10163 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10165 * New function EC_GROUP_get_degree, which is defined through
10166 EC_METHOD. For curves over prime fields, this returns the bit
10167 length of the modulus.
10169 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10171 * New functions EC_GROUP_dup, EC_POINT_dup.
10172 (These simply call ..._new and ..._copy).
10174 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10176 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10177 Polynomials are represented as BIGNUMs (where the sign bit is not
10178 used) in the following functions [macros]:
10181 BN_GF2m_sub [= BN_GF2m_add]
10182 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10183 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10184 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10186 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10187 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10188 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10189 BN_GF2m_cmp [= BN_ucmp]
10191 (Note that only the 'mod' functions are actually for fields GF(2^m).
10192 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10194 For some functions, an the irreducible polynomial defining a
10195 field can be given as an 'unsigned int[]' with strictly
10196 decreasing elements giving the indices of those bits that are set;
10197 i.e., p[] represents the polynomial
10198 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10200 p[0] > p[1] > ... > p[k] = 0.
10201 This applies to the following functions:
10204 BN_GF2m_mod_mul_arr
10205 BN_GF2m_mod_sqr_arr
10206 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10207 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10208 BN_GF2m_mod_exp_arr
10209 BN_GF2m_mod_sqrt_arr
10210 BN_GF2m_mod_solve_quad_arr
10214 Conversion can be performed by the following functions:
10219 bntest.c has additional tests for binary polynomial arithmetic.
10221 Two implementations for BN_GF2m_mod_div() are available.
10222 The default algorithm simply uses BN_GF2m_mod_inv() and
10223 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10224 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10225 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10227 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10229 * Add new error code 'ERR_R_DISABLED' that can be used when some
10230 functionality is disabled at compile-time.
10232 *Douglas Stebila <douglas.stebila@sun.com>*
10234 * Change default behaviour of 'openssl asn1parse' so that more
10235 information is visible when viewing, e.g., a certificate:
10237 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10238 mode the content of non-printable OCTET STRINGs is output in a
10239 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10240 avoid the appearance of a printable string.
10242 *Nils Larsch <nla@trustcenter.de>*
10244 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10246 EC_GROUP_set_asn1_flag()
10247 EC_GROUP_get_asn1_flag()
10248 EC_GROUP_set_point_conversion_form()
10249 EC_GROUP_get_point_conversion_form()
10250 These control ASN1 encoding details:
10251 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10252 has been set to OPENSSL_EC_NAMED_CURVE.
10253 - Points are encoded in uncompressed form by default; options for
10254 asn1_for are as for point2oct, namely
10255 POINT_CONVERSION_COMPRESSED
10256 POINT_CONVERSION_UNCOMPRESSED
10257 POINT_CONVERSION_HYBRID
10259 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10261 EC_GROUP_set_seed()
10262 EC_GROUP_get0_seed()
10263 EC_GROUP_get_seed_len()
10264 This is used only for ASN1 purposes (so far).
10266 *Nils Larsch <nla@trustcenter.de>*
10268 * Add 'field_type' member to EC_METHOD, which holds the NID
10269 of the appropriate field type OID. The new function
10270 EC_METHOD_get_field_type() returns this value.
10272 *Nils Larsch <nla@trustcenter.de>*
10275 EC_POINT_point2bn()
10276 EC_POINT_bn2point()
10277 EC_POINT_point2hex()
10278 EC_POINT_hex2point()
10279 providing useful interfaces to EC_POINT_point2oct() and
10280 EC_POINT_oct2point().
10282 *Nils Larsch <nla@trustcenter.de>*
10284 * Change internals of the EC library so that the functions
10285 EC_GROUP_set_generator()
10286 EC_GROUP_get_generator()
10287 EC_GROUP_get_order()
10288 EC_GROUP_get_cofactor()
10289 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10290 to methods, which would lead to unnecessary code duplication when
10291 adding different types of curves.
10293 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10295 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10296 arithmetic, and such that modified wNAFs are generated
10297 (which avoid length expansion in many cases).
10301 * Add a function EC_GROUP_check_discriminant() (defined via
10302 EC_METHOD) that verifies that the curve discriminant is non-zero.
10304 Add a function EC_GROUP_check() that makes some sanity tests
10305 on a EC_GROUP, its generator and order. This includes
10306 EC_GROUP_check_discriminant().
10308 *Nils Larsch <nla@trustcenter.de>*
10310 * Add ECDSA in new directory crypto/ecdsa/.
10312 Add applications 'openssl ecparam' and 'openssl ecdsa'
10313 (these are based on 'openssl dsaparam' and 'openssl dsa').
10315 ECDSA support is also included in various other files across the
10316 library. Most notably,
10317 - 'openssl req' now has a '-newkey ecdsa:file' option;
10318 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10319 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10320 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10321 them suitable for ECDSA where domain parameters must be
10322 extracted before the specific public key;
10323 - ECDSA engine support has been added.
10325 *Nils Larsch <nla@trustcenter.de>*
10327 * Include some named elliptic curves, and add OIDs from X9.62,
10328 SECG, and WAP/WTLS. Each curve can be obtained from the new
10330 EC_GROUP_new_by_curve_name(),
10331 and the list of available named curves can be obtained with
10332 EC_get_builtin_curves().
10333 Also add a 'curve_name' member to EC_GROUP objects, which can be
10335 EC_GROUP_set_curve_name()
10336 EC_GROUP_get_curve_name()
10338 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10340 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10341 was actually never needed) and in BN_mul(). The removal in BN_mul()
10342 required a small change in bn_mul_part_recursive() and the addition
10343 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10344 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10345 bn_sub_words() and bn_add_words() except they take arrays with
10350 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10352 * Cleanse PEM buffers before freeing them since they may contain
10355 *Benjamin Bennett <ben@psc.edu>*
10357 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10358 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10359 authentication-only ciphersuites.
10363 * Since AES128 and AES256 share a single mask bit in the logic of
10364 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10365 kludge to work properly if AES128 is available and AES256 isn't.
10369 * Expand security boundary to match 1.1.1 module.
10373 * Remove redundant features: hash file source, editing of test vectors
10374 modify fipsld to use external fips_premain.c signature.
10378 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10379 run algorithm test programs.
10383 * Make algorithm test programs more tolerant of whitespace.
10387 * Have SSL/TLS server implementation tolerate "mismatched" record
10388 protocol version while receiving ClientHello even if the
10389 ClientHello is fragmented. (The server can't insist on the
10390 particular protocol version it has chosen before the ServerHello
10391 message has informed the client about his choice.)
10395 * Load error codes if they are not already present instead of using a
10396 static variable. This allows them to be cleanly unloaded and reloaded.
10400 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10402 * Introduce limits to prevent malicious keys being able to
10403 cause a denial of service. [CVE-2006-2940][]
10405 *Steve Henson, Bodo Moeller*
10407 * Fix ASN.1 parsing of certain invalid structures that can result
10408 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10410 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10411 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10413 * Fix SSL client code which could crash if connecting to a
10414 malicious SSLv2 server. [CVE-2006-4343][]
10416 *Tavis Ormandy and Will Drewry, Google Security Team*
10418 * Change ciphersuite string processing so that an explicit
10419 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10420 will no longer include "AES128-SHA"), and any other similar
10421 ciphersuite (same bitmap) from *other* protocol versions (so that
10422 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10423 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10424 changes from 0.9.8b and 0.9.8d.
10428 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10430 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10431 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10433 * Change the Unix randomness entropy gathering to use poll() when
10434 possible instead of select(), since the latter has some
10435 undesirable limitations.
10437 *Darryl Miles via Richard Levitte and Bodo Moeller*
10439 * Disable rogue ciphersuites:
10441 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10442 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10443 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10445 The latter two were purportedly from
10446 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10449 Also deactivate the remaining ciphersuites from
10450 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10451 unofficial, and the ID has long expired.
10455 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10456 dual-core machines) and other potential thread-safety issues.
10460 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10462 * Adapt fipsld and the build system to link against the validated FIPS
10463 module in FIPS mode.
10467 * Fixes for VC++ 2005 build under Windows.
10471 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10472 from a Windows bash shell such as MSYS. It is autodetected from the
10473 "config" script when run from a VC++ environment. Modify standard VC++
10474 build to use fipscanister.o from the GNU make build.
10478 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10480 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10481 The value now differs depending on if you build for FIPS or not.
10482 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10483 safely run with a non-FIPSed libcrypto, as it may crash because of
10484 the difference induced by this change.
10488 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10490 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10491 (part of SSL_OP_ALL). This option used to disable the
10492 countermeasure against man-in-the-middle protocol-version
10493 rollback in the SSL 2.0 server implementation, which is a bad
10494 idea. [CVE-2005-2969][]
10496 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10497 for Information Security, National Institute of Advanced Industrial
10498 Science and Technology [AIST, Japan)]*
10500 * Minimal support for X9.31 signatures and PSS padding modes. This is
10501 mainly for FIPS compliance and not fully integrated at this stage.
10505 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10506 the exponentiation using a fixed-length exponent. (Otherwise,
10507 the information leaked through timing could expose the secret key
10508 after many signatures; cf. Bleichenbacher's attack on DSA with
10513 * Make a new fixed-window mod_exp implementation the default for
10514 RSA, DSA, and DH private-key operations so that the sequence of
10515 squares and multiplies and the memory access pattern are
10516 independent of the particular secret key. This will mitigate
10517 cache-timing and potential related attacks.
10519 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10520 and this is automatically used by BN_mod_exp_mont() if the new flag
10521 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10522 will use this BN flag for private exponents unless the flag
10523 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10524 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10526 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10528 * Change the client implementation for SSLv23_method() and
10529 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10530 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10531 (Previously, the SSL 2.0 backwards compatible Client Hello
10532 message format would be used even with SSL_OP_NO_SSLv2.)
10536 * Add support for smime-type MIME parameter in S/MIME messages which some
10541 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10542 a threadsafe manner. Modify rsa code to use new function and add calls
10543 to dsa and dh code (which had race conditions before).
10547 * Include the fixed error library code in the C error file definitions
10548 instead of fixing them up at runtime. This keeps the error code
10549 structures constant.
10553 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10555 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10558 * Fixes for newer kerberos headers. NB: the casts are needed because
10559 the 'length' field is signed on one version and unsigned on another
10560 with no (?) obvious way to tell the difference, without these VC++
10561 complains. Also the "definition" of FAR (blank) is no longer included
10562 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10563 some needed definitions.
10567 * Undo Cygwin change.
10571 * Added support for proxy certificates according to RFC 3820.
10572 Because they may be a security thread to unaware applications,
10573 they must be explicitly allowed in run-time. See
10574 docs/HOWTO/proxy_certificates.txt for further information.
10578 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10580 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10581 server and client random values. Previously
10582 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10583 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10585 This change has negligible security impact because:
10587 1. Server and client random values still have 24 bytes of pseudo random
10590 2. Server and client random values are sent in the clear in the initial
10593 3. The master secret is derived using the premaster secret (48 bytes in
10594 size for static RSA ciphersuites) as well as client server and random
10597 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10600 *Stephen Henson, reported by UK NISCC*
10602 * Use Windows randomness collection on Cygwin.
10606 * Fix hang in EGD/PRNGD query when communication socket is closed
10607 prematurely by EGD/PRNGD.
10609 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10611 * Prompt for pass phrases when appropriate for PKCS12 input format.
10615 * Back-port of selected performance improvements from development
10616 branch, as well as improved support for PowerPC platforms.
10620 * Add lots of checks for memory allocation failure, error codes to indicate
10621 failure and freeing up memory if a failure occurs.
10623 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10625 * Add new -passin argument to dgst.
10629 * Perform some character comparisons of different types in X509_NAME_cmp:
10630 this is needed for some certificates that re-encode DNs into UTF8Strings
10631 (in violation of RFC3280) and can't or won't issue name rollover
10636 * Make an explicit check during certificate validation to see that
10637 the CA setting in each certificate on the chain is correct. As a
10638 side effect always do the following basic checks on extensions,
10639 not just when there's an associated purpose to the check:
10641 - if there is an unhandled critical extension (unless the user
10642 has chosen to ignore this fault)
10643 - if the path length has been exceeded (if one is set at all)
10644 - that certain extensions fit the associated purpose (if one has
10649 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10651 * Avoid a race condition when CRLs are checked in a multi threaded
10652 environment. This would happen due to the reordering of the revoked
10653 entries during signature checking and serial number lookup. Now the
10654 encoding is cached and the serial number sort performed under a lock.
10655 Add new STACK function sk_is_sorted().
10659 * Add Delta CRL to the extension code.
10663 * Various fixes to s3_pkt.c so alerts are sent properly.
10665 *David Holmes <d.holmes@f5.com>*
10667 * Reduce the chances of duplicate issuer name and serial numbers (in
10668 violation of RFC3280) using the OpenSSL certificate creation utilities.
10669 This is done by creating a random 64 bit value for the initial serial
10670 number when a serial number file is created or when a self signed
10671 certificate is created using 'openssl req -x509'. The initial serial
10672 number file is created using 'openssl x509 -next_serial' in CA.pl
10673 rather than being initialized to 1.
10677 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10679 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10680 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10682 *Joe Orton, Steve Henson*
10684 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10687 *Joe Orton, Steve Henson*
10689 * Make it possible to have multiple active certificates with the same
10690 subject in the CA index file. This is done only if the keyword
10691 'unique_subject' is set to 'no' in the main CA section (default
10692 if 'CA_default') of the configuration file. The value is saved
10693 with the database itself in a separate index attribute file,
10694 named like the index file with '.attr' appended to the name.
10698 * X509 verify fixes. Disable broken certificate workarounds when
10699 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10700 keyUsage extension present. Don't accept CRLs with unhandled critical
10701 extensions: since verify currently doesn't process CRL extensions this
10702 rejects a CRL with *any* critical extensions. Add new verify error codes
10707 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10708 A clarification of RFC2560 will require the use of OCTET STRINGs and
10709 some implementations cannot handle the current raw format. Since OpenSSL
10710 copies and compares OCSP nonces as opaque blobs without any attempt at
10711 parsing them this should not create any compatibility issues.
10715 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10716 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10717 this HMAC (and other) operations are several times slower than OpenSSL
10722 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10724 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10726 * Use the correct content when signing type "other".
10730 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10732 * Fix various bugs revealed by running the NISCC test suite:
10734 Stop out of bounds reads in the ASN1 code when presented with
10735 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10737 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10739 If verify callback ignores invalid public key errors don't try to check
10740 certificate signature with the NULL public key.
10744 * New -ignore_err option in ocsp application to stop the server
10745 exiting on the first error in a request.
10749 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10750 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10755 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10756 extra data after the compression methods not only for TLS 1.0
10757 but also for SSL 3.0 (as required by the specification).
10759 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10761 * Change X509_certificate_type() to mark the key as exported/exportable
10762 when it's 512 *bits* long, not 512 bytes.
10766 * Change AES_cbc_encrypt() so it outputs exact multiple of
10767 blocks during encryption.
10771 * Various fixes to base64 BIO and non blocking I/O. On write
10772 flushes were not handled properly if the BIO retried. On read
10773 data was not being buffered properly and had various logic bugs.
10774 This also affects blocking I/O when the data being decoded is a
10779 * Various S/MIME bugfixes and compatibility changes:
10780 output correct application/pkcs7 MIME type if
10781 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10782 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10783 of files as .eml work). Correctly handle very long lines in MIME
10788 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10790 * Countermeasure against the Klima-Pokorny-Rosa extension of
10791 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10792 a protocol version number mismatch like a decryption error
10793 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10797 * Turn on RSA blinding by default in the default implementation
10798 to avoid a timing attack. Applications that don't want it can call
10799 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10800 They would be ill-advised to do so in most cases.
10802 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10804 * Change RSA blinding code so that it works when the PRNG is not
10805 seeded (in this case, the secret RSA exponent is abused as
10806 an unpredictable seed -- if it is not unpredictable, there
10807 is no point in blinding anyway). Make RSA blinding thread-safe
10808 by remembering the creator's thread ID in rsa->blinding and
10809 having all other threads use local one-time blinding factors
10810 (this requires more computation than sharing rsa->blinding, but
10811 avoids excessive locking; and if an RSA object is not shared
10812 between threads, blinding will still be very fast).
10816 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10817 ENGINE as defaults for all supported algorithms irrespective of
10818 the 'flags' parameter. 'flags' is now honoured, so applications
10819 should make sure they are passing it correctly.
10823 * Target "mingw" now allows native Windows code to be generated in
10824 the Cygwin environment as well as with the MinGW compiler.
10828 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10830 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10831 via timing by performing a MAC computation even if incorrect
10832 block cipher padding has been found. This is a countermeasure
10833 against active attacks where the attacker has to distinguish
10834 between bad padding and a MAC verification error. [CVE-2003-0078][]
10836 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10837 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10838 Martin Vuagnoux (EPFL, Ilion)*
10840 * Make the no-err option work as intended. The intention with no-err
10841 is not to have the whole error stack handling routines removed from
10842 libcrypto, it's only intended to remove all the function name and
10843 reason texts, thereby removing some of the footprint that may not
10844 be interesting if those errors aren't displayed anyway.
10846 NOTE: it's still possible for any application or module to have its
10847 own set of error texts inserted. The routines are there, just not
10848 used by default when no-err is given.
10852 * Add support for FreeBSD on IA64.
10854 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10856 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10857 Kerberos function mit_des_cbc_cksum(). Before this change,
10858 the value returned by DES_cbc_cksum() was like the one from
10859 mit_des_cbc_cksum(), except the bytes were swapped.
10861 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10863 * Allow an application to disable the automatic SSL chain building.
10864 Before this a rather primitive chain build was always performed in
10865 ssl3_output_cert_chain(): an application had no way to send the
10866 correct chain if the automatic operation produced an incorrect result.
10868 Now the chain builder is disabled if either:
10870 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10872 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10874 The reasoning behind this is that an application would not want the
10875 auto chain building to take place if extra chain certificates are
10876 present and it might also want a means of sending no additional
10877 certificates (for example the chain has two certificates and the
10882 * Add the possibility to build without the ENGINE framework.
10884 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10886 * Under Win32 gmtime() can return NULL: check return value in
10887 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10891 * DSA routines: under certain error conditions uninitialized BN objects
10892 could be freed. Solution: make sure initialization is performed early
10893 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10894 Nils Larsch <nla@trustcenter.de> via PR#459)
10898 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10899 checked on reconnect on the client side, therefore session resumption
10900 could still fail with a "ssl session id is different" error. This
10901 behaviour is masked when SSL_OP_ALL is used due to
10902 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10903 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10904 followup to PR #377.
10908 * IA-32 assembler support enhancements: unified ELF targets, support
10909 for SCO/Caldera platforms, fix for Cygwin shared build.
10913 * Add support for FreeBSD on sparc64. As a consequence, support for
10914 FreeBSD on non-x86 processors is separate from x86 processors on
10915 the config script, much like the NetBSD support.
10917 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10919 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
10921 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10924 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10925 code (06) was taken as the first octet of the session ID and the last
10926 octet was ignored consequently. As a result SSLv2 client side session
10927 caching could not have worked due to the session ID mismatch between
10929 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10934 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10935 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10940 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10941 seems that in spite of existing for more than a year, many application
10942 author have done nothing to provide the necessary callbacks, which
10943 means that this particular engine will not work properly anywhere.
10944 This is a very unfortunate situation which forces us, in the name
10945 of usability, to give the hw_ncipher.c a static lock, which is part
10947 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10948 appear in 0.9.8 or later. We EXPECT application authors to have
10949 dealt properly with this when 0.9.8 is released (unless we actually
10950 make such changes in the libcrypto locking code that changes will
10951 have to be made anyway).
10955 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10956 octets have been read, EOF or an error occurs. Without this change
10957 some truncated ASN1 structures will not produce an error.
10961 * Disable Heimdal support, since it hasn't been fully implemented.
10962 Still give the possibility to force the use of Heimdal, but with
10963 warnings and a request that patches get sent to openssl-dev.
10967 * Add the VC-CE target, introduce the WINCE sysname, and add
10968 INSTALL.WCE and appropriate conditionals to make it build.
10970 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10972 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10973 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10974 edit numbers of the version.
10976 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10978 * Introduce safe string copy and catenation functions
10979 (BUF_strlcpy() and BUF_strlcat()).
10981 *Ben Laurie (CHATS) and Richard Levitte*
10983 * Avoid using fixed-size buffers for one-line DNs.
10985 *Ben Laurie (CHATS)*
10987 * Add BUF_MEM_grow_clean() to avoid information leakage when
10988 resizing buffers containing secrets, and use where appropriate.
10990 *Ben Laurie (CHATS)*
10992 * Avoid using fixed size buffers for configuration file location.
10994 *Ben Laurie (CHATS)*
10996 * Avoid filename truncation for various CA files.
10998 *Ben Laurie (CHATS)*
11000 * Use sizeof in preference to magic numbers.
11002 *Ben Laurie (CHATS)*
11004 * Avoid filename truncation in cert requests.
11006 *Ben Laurie (CHATS)*
11008 * Add assertions to check for (supposedly impossible) buffer
11011 *Ben Laurie (CHATS)*
11013 * Don't cache truncated DNS entries in the local cache (this could
11014 potentially lead to a spoofing attack).
11016 *Ben Laurie (CHATS)*
11018 * Fix various buffers to be large enough for hex/decimal
11019 representations in a platform independent manner.
11021 *Ben Laurie (CHATS)*
11023 * Add CRYPTO_realloc_clean() to avoid information leakage when
11024 resizing buffers containing secrets, and use where appropriate.
11026 *Ben Laurie (CHATS)*
11028 * Add BIO_indent() to avoid much slightly worrying code to do
11031 *Ben Laurie (CHATS)*
11033 * Convert sprintf()/BIO_puts() to BIO_printf().
11035 *Ben Laurie (CHATS)*
11037 * buffer_gets() could terminate with the buffer only half
11040 *Ben Laurie (CHATS)*
11042 * Add assertions to prevent user-supplied crypto functions from
11043 overflowing internal buffers by having large block sizes, etc.
11045 *Ben Laurie (CHATS)*
11047 * New OPENSSL_assert() macro (similar to assert(), but enabled
11050 *Ben Laurie (CHATS)*
11052 * Eliminate unused copy of key in RC4.
11054 *Ben Laurie (CHATS)*
11056 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11058 *Ben Laurie (CHATS)*
11060 * Fix off-by-one error in EGD path.
11062 *Ben Laurie (CHATS)*
11064 * If RANDFILE path is too long, ignore instead of truncating.
11066 *Ben Laurie (CHATS)*
11068 * Eliminate unused and incorrectly sized X.509 structure
11071 *Ben Laurie (CHATS)*
11073 * Eliminate unused and dangerous function knumber().
11075 *Ben Laurie (CHATS)*
11077 * Eliminate unused and dangerous structure, KSSL_ERR.
11079 *Ben Laurie (CHATS)*
11081 * Protect against overlong session ID context length in an encoded
11082 session object. Since these are local, this does not appear to be
11085 *Ben Laurie (CHATS)*
11087 * Change from security patch (see 0.9.6e below) that did not affect
11088 the 0.9.6 release series:
11090 Remote buffer overflow in SSL3 protocol - an attacker could
11091 supply an oversized master key in Kerberos-enabled versions.
11094 *Ben Laurie (CHATS)*
11096 * Change the SSL kerb5 codes to match RFC 2712.
11100 * Make -nameopt work fully for req and add -reqopt switch.
11102 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11104 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11106 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11108 * Make sure tests can be performed even if the corresponding algorithms
11109 have been removed entirely. This was also the last step to make
11110 OpenSSL compilable with DJGPP under all reasonable conditions.
11112 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11114 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11115 to allow version independent disabling of normally unselected ciphers,
11116 which may be activated as a side-effect of selecting a single cipher.
11118 (E.g., cipher list string "RSA" enables ciphersuites that are left
11119 out of "ALL" because they do not provide symmetric encryption.
11120 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11122 *Lutz Jaenicke, Bodo Moeller*
11124 * Add appropriate support for separate platform-dependent build
11125 directories. The recommended way to make a platform-dependent
11126 build directory is the following (tested on Linux), maybe with
11129 # Place yourself outside of the OpenSSL source tree. In
11130 # this example, the environment variable OPENSSL_SOURCE
11131 # is assumed to contain the absolute OpenSSL source directory.
11132 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11133 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11134 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11135 mkdir -p `dirname $F`
11136 ln -s $OPENSSL_SOURCE/$F $F
11139 To be absolutely sure not to disturb the source tree, a "make clean"
11140 is a good thing. If it isn't successful, don't worry about it,
11141 it probably means the source directory is very clean.
11145 * Make sure any ENGINE control commands make local copies of string
11146 pointers passed to them whenever necessary. Otherwise it is possible
11147 the caller may have overwritten (or deallocated) the original string
11148 data when a later ENGINE operation tries to use the stored values.
11150 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11152 * Improve diagnostics in file reading and command-line digests.
11154 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11156 * Add AES modes CFB and OFB to the object database. Correct an
11157 error in AES-CFB decryption.
11161 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11162 allows existing EVP_CIPHER_CTX structures to be reused after
11163 calling `EVP_*Final()`. This behaviour is used by encryption
11164 BIOs and some applications. This has the side effect that
11165 applications must explicitly clean up cipher contexts with
11166 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11170 * Check the values of dna and dnb in bn_mul_recursive before calling
11171 bn_mul_comba (a non zero value means the a or b arrays do not contain
11172 n2 elements) and fallback to bn_mul_normal if either is not zero.
11176 * Fix escaping of non-ASCII characters when using the -subj option
11177 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11181 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11182 form for "surname", serialNumber has no short form.
11183 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11184 therefore remove "mail" short name for "internet 7".
11185 The OID for unique identifiers in X509 certificates is
11186 x500UniqueIdentifier, not uniqueIdentifier.
11187 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11191 * Add an "init" command to the ENGINE config module and auto initialize
11192 ENGINEs. Without any "init" command the ENGINE will be initialized
11193 after all ctrl commands have been executed on it. If init=1 the
11194 ENGINE is initialized at that point (ctrls before that point are run
11195 on the uninitialized ENGINE and after on the initialized one). If
11196 init=0 then the ENGINE will not be initialized at all.
11200 * Fix the 'app_verify_callback' interface so that the user-defined
11201 argument is actually passed to the callback: In the
11202 SSL_CTX_set_cert_verify_callback() prototype, the callback
11203 declaration has been changed from
11206 int (*cb)(X509_STORE_CTX *,void *);
11207 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11208 i=s->ctx->app_verify_callback(&ctx)
11209 has been changed into
11210 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11212 To update applications using SSL_CTX_set_cert_verify_callback(),
11213 a dummy argument can be added to their callback functions.
11215 *D. K. Smetters <smetters@parc.xerox.com>*
11217 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11219 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11221 * Add and OPENSSL_LOAD_CONF define which will cause
11222 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11223 This allows older applications to transparently support certain
11224 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11225 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11226 load the config file and OPENSSL_add_all_algorithms_conf() which will
11227 always load it have also been added.
11231 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11232 Adjust NIDs and EVP layer.
11234 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11236 * Config modules support in openssl utility.
11238 Most commands now load modules from the config file,
11239 though in a few (such as version) this isn't done
11240 because it couldn't be used for anything.
11242 In the case of ca and req the config file used is
11243 the same as the utility itself: that is the -config
11244 command line option can be used to specify an
11249 * Move default behaviour from OPENSSL_config(). If appname is NULL
11250 use "openssl_conf" if filename is NULL use default openssl config file.
11254 * Add an argument to OPENSSL_config() to allow the use of an alternative
11255 config section name. Add a new flag to tolerate a missing config file
11256 and move code to CONF_modules_load_file().
11260 * Support for crypto accelerator cards from Accelerated Encryption
11261 Processing, www.aep.ie. (Use engine 'aep')
11262 The support was copied from 0.9.6c [engine] and adapted/corrected
11263 to work with the new engine framework.
11265 *AEP Inc. and Richard Levitte*
11267 * Support for SureWare crypto accelerator cards from Baltimore
11268 Technologies. (Use engine 'sureware')
11269 The support was copied from 0.9.6c [engine] and adapted
11270 to work with the new engine framework.
11274 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11275 make the newer ENGINE framework commands for the CHIL engine work.
11277 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11279 * Make it possible to produce shared libraries on ReliantUNIX.
11281 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11283 * Add the configuration target debug-linux-ppro.
11284 Make 'openssl rsa' use the general key loading routines
11285 implemented in apps.c, and make those routines able to
11286 handle the key format FORMAT_NETSCAPE and the variant
11289 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11291 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11293 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11295 * Add -keyform to rsautl, and document -engine.
11297 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11299 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11300 BIO_R_NO_SUCH_FILE error code rather than the generic
11301 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11305 * Add new functions
11306 ERR_peek_last_error
11307 ERR_peek_last_error_line
11308 ERR_peek_last_error_line_data.
11309 These are similar to
11311 ERR_peek_error_line
11312 ERR_peek_error_line_data,
11313 but report on the latest error recorded rather than the first one
11314 still in the error queue.
11316 *Ben Laurie, Bodo Moeller*
11318 * default_algorithms option in ENGINE config module. This allows things
11320 default_algorithms = ALL
11321 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11325 * Preliminary ENGINE config module.
11329 * New experimental application configuration code.
11333 * Change the AES code to follow the same name structure as all other
11334 symmetric ciphers, and behave the same way. Move everything to
11335 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11337 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11339 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11341 *Ben Laurie and Theo de Raadt*
11343 * Add option to output public keys in req command.
11345 *Massimiliano Pala madwolf@openca.org*
11347 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11348 (up to about 10% better than before for P-192 and P-224).
11352 * New functions/macros
11354 SSL_CTX_set_msg_callback(ctx, cb)
11355 SSL_CTX_set_msg_callback_arg(ctx, arg)
11356 SSL_set_msg_callback(ssl, cb)
11357 SSL_set_msg_callback_arg(ssl, arg)
11359 to request calling a callback function
11361 void cb(int write_p, int version, int content_type,
11362 const void *buf, size_t len, SSL *ssl, void *arg)
11364 whenever a protocol message has been completely received
11365 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11366 protocol version according to which the SSL library interprets
11367 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11368 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11369 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11370 specification (change_cipher_spec(20), alert(21), handshake(22)).
11371 'buf' and 'len' point to the actual message, 'ssl' to the
11372 SSL object, and 'arg' is the application-defined value set by
11373 SSL[_CTX]_set_msg_callback_arg().
11375 'openssl s_client' and 'openssl s_server' have new '-msg' options
11376 to enable a callback that displays all protocol messages.
11380 * Change the shared library support so shared libraries are built as
11381 soon as the corresponding static library is finished, and thereby get
11382 openssl and the test programs linked against the shared library.
11383 This still only happens when the keyword "shard" has been given to
11384 the configuration scripts.
11386 NOTE: shared library support is still an experimental thing, and
11387 backward binary compatibility is still not guaranteed.
11389 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11391 * Add support for Subject Information Access extension.
11393 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11395 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11396 additional bytes when new memory had to be allocated, not just
11397 when reusing an existing buffer.
11401 * New command line and configuration option 'utf8' for the req command.
11402 This allows field values to be specified as UTF8 strings.
11406 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11407 runs for the former and machine-readable output for the latter.
11411 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11412 of the e-mail address in the DN (i.e., it will go into a certificate
11413 extension only). The new configuration file option 'email_in_dn = no'
11414 has the same effect.
11416 *Massimiliano Pala madwolf@openca.org*
11418 * Change all functions with names starting with `des_` to be starting
11419 with `DES_` instead. Add wrappers that are compatible with libdes,
11420 but are named `_ossl_old_des_*`. Finally, add macros that map the
11421 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11422 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11423 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11426 Since we provide two compatibility mappings, the user needs to
11427 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11428 compatibility is desired. The default (i.e., when that macro
11429 isn't defined) is OpenSSL 0.9.6c compatibility.
11431 There are also macros that enable and disable the support of old
11432 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11433 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11434 are defined, the default will apply: to support the old des routines.
11436 In either case, one must include openssl/des.h to get the correct
11437 definitions. Do not try to just include openssl/des_old.h, that
11440 NOTE: This is a major break of an old API into a new one. Software
11441 authors are encouraged to switch to the `DES_` style functions. Some
11442 time in the future, des_old.h and the libdes compatibility functions
11443 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11444 default), and then completely removed.
11448 * Test for certificates which contain unsupported critical extensions.
11449 If such a certificate is found during a verify operation it is
11450 rejected by default: this behaviour can be overridden by either
11451 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11452 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11453 X509_supported_extension() has also been added which returns 1 if a
11454 particular extension is supported.
11458 * Modify the behaviour of EVP cipher functions in similar way to digests
11459 to retain compatibility with existing code.
11463 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11464 compatibility with existing code. In particular the 'ctx' parameter does
11465 not have to be to be initialized before the call to EVP_DigestInit() and
11466 it is tidied up after a call to EVP_DigestFinal(). New function
11467 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11468 EVP_MD_CTX_copy() changed to not require the destination to be
11469 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11470 requires the destination to be valid.
11472 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11473 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11477 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11478 so that complete 'Handshake' protocol structures are kept in memory
11479 instead of overwriting 'msg_type' and 'length' with 'body' data.
11483 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11485 *Massimo Santin via Richard Levitte*
11487 * Major restructuring to the underlying ENGINE code. This includes
11488 reduction of linker bloat, separation of pure "ENGINE" manipulation
11489 (initialisation, etc) from functionality dealing with implementations
11490 of specific crypto interfaces. This change also introduces integrated
11491 support for symmetric ciphers and digest implementations - so ENGINEs
11492 can now accelerate these by providing EVP_CIPHER and EVP_MD
11493 implementations of their own. This is detailed in crypto/engine/README
11494 as it couldn't be adequately described here. However, there are a few
11495 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11496 were changed in the original introduction of ENGINE code have now
11497 reverted back - the hooking from this code to ENGINE is now a good
11498 deal more passive and at run-time, operations deal directly with
11499 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11500 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11501 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11502 they were not being used by the framework as there is no concept of a
11503 BIGNUM_METHOD and they could not be generalised to the new
11504 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11505 ENGINE_cpy() has been removed as it cannot be consistently defined in
11510 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11514 * Change mkdef.pl to sort symbols that get the same entry number,
11515 and make sure the automatically generated functions `ERR_load_*`
11516 become part of libeay.num as well.
11520 * New function SSL_renegotiate_pending(). This returns true once
11521 renegotiation has been requested (either SSL_renegotiate() call
11522 or HelloRequest/ClientHello received from the peer) and becomes
11523 false once a handshake has been completed.
11524 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11525 sends a HelloRequest, but does not ensure that a handshake takes
11526 place. SSL_renegotiate_pending() is useful for checking if the
11527 client has followed the request.)
11531 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11532 By default, clients may request session resumption even during
11533 renegotiation (if session ID contexts permit); with this option,
11534 session resumption is possible only in the first handshake.
11536 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11537 more bits available for options that should not be part of
11538 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11542 * Add some demos for certificate and certificate request creation.
11546 * Make maximum certificate chain size accepted from the peer application
11547 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11548 "Douglas E. Engert" <deengert@anl.gov>.
11552 * Add support for shared libraries for Unixware-7
11553 (Boyd Lynn Gerber <gerberb@zenez.com>).
11557 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11558 be done prior to destruction. Use this to unload error strings from
11559 ENGINEs that load their own error strings. NB: This adds two new API
11560 functions to "get" and "set" this destroy handler in an ENGINE.
11564 * Alter all existing ENGINE implementations (except "openssl" and
11565 "openbsd") to dynamically instantiate their own error strings. This
11566 makes them more flexible to be built both as statically-linked ENGINEs
11567 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11568 Also, add stub code to each that makes building them as self-contained
11569 shared-libraries easier (see README.ENGINE).
11573 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11574 implementations into applications that are completely implemented in
11575 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11576 commands that can be used to configure what shared-library to load and
11577 to control aspects of the way it is handled. Also, made an update to
11578 the README.ENGINE file that brings its information up-to-date and
11579 provides some information and instructions on the "dynamic" ENGINE
11580 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11584 * Make it possible to unload ranges of ERR strings with a new
11585 "ERR_unload_strings" function.
11589 * Add a copy() function to EVP_MD.
11593 * Make EVP_MD routines take a context pointer instead of just the
11594 md_data void pointer.
11598 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11599 that the digest can only process a single chunk of data
11600 (typically because it is provided by a piece of
11601 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11602 is only going to provide a single chunk of data, and hence the
11603 framework needn't accumulate the data for oneshot drivers.
11607 * As with "ERR", make it possible to replace the underlying "ex_data"
11608 functions. This change also alters the storage and management of global
11609 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11610 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11611 index counters. The API functions that use this state have been changed
11612 to take a "class_index" rather than pointers to the class's local STACK
11613 and counter, and there is now an API function to dynamically create new
11614 classes. This centralisation allows us to (a) plug a lot of the
11615 thread-safety problems that existed, and (b) makes it possible to clean
11616 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11617 such data would previously have always leaked in application code and
11618 workarounds were in place to make the memory debugging turn a blind eye
11619 to it. Application code that doesn't use this new function will still
11620 leak as before, but their memory debugging output will announce it now
11621 rather than letting it slide.
11623 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11624 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11625 has a return value to indicate success or failure.
11629 * Make it possible to replace the underlying "ERR" functions such that the
11630 global state (2 LHASH tables and 2 locks) is only used by the "default"
11631 implementation. This change also adds two functions to "get" and "set"
11632 the implementation prior to it being automatically set the first time
11633 any other ERR function takes place. Ie. an application can call "get",
11634 pass the return value to a module it has just loaded, and that module
11635 can call its own "set" function using that value. This means the
11636 module's "ERR" operations will use (and modify) the error state in the
11637 application and not in its own statically linked copy of OpenSSL code.
11641 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11642 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11643 the operation, and provides a more encapsulated way for external code
11644 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11645 to use these functions rather than manually incrementing the counts.
11647 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11651 * Add EVP test program.
11655 * Add symmetric cipher support to ENGINE. Expect the API to change!
11659 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11660 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11661 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11662 These allow a CRL to be built without having to access X509_CRL fields
11663 directly. Modify 'ca' application to use new functions.
11667 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11668 bug workarounds. Rollback attack detection is a security feature.
11669 The problem will only arise on OpenSSL servers when TLSv1 is not
11670 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11671 Software authors not wanting to support TLSv1 will have special reasons
11672 for their choice and can explicitly enable this option.
11674 *Bodo Moeller, Lutz Jaenicke*
11676 * Rationalise EVP so it can be extended: don't include a union of
11677 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11678 (similar to those existing for EVP_CIPHER_CTX).
11683 EVP_MD_CTX_init(&md); /* new function call */
11684 EVP_DigestInit(&md, EVP_sha1());
11685 EVP_DigestUpdate(&md, in, len);
11686 EVP_DigestFinal(&md, out, NULL);
11687 EVP_MD_CTX_cleanup(&md); /* new function call */
11691 * Make DES key schedule conform to the usual scheme, as well as
11692 correcting its structure. This means that calls to DES functions
11693 now have to pass a pointer to a des_key_schedule instead of a
11694 plain des_key_schedule (which was actually always a pointer
11697 des_key_schedule ks;
11699 des_set_key_checked(..., &ks);
11700 des_ncbc_encrypt(..., &ks, ...);
11702 (Note that a later change renames 'des_...' into 'DES_...'.)
11706 * Initial reduction of linker bloat: the use of some functions, such as
11707 PEM causes large amounts of unused functions to be linked in due to
11708 poor organisation. For example pem_all.c contains every PEM function
11709 which has a knock on effect of linking in large amounts of (unused)
11710 ASN1 code. Grouping together similar functions and splitting unrelated
11711 functions prevents this.
11715 * Cleanup of EVP macros.
11719 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11720 correct `_ecb suffix`.
11724 * Add initial OCSP responder support to ocsp application. The
11725 revocation information is handled using the text based index
11726 use by the ca application. The responder can either handle
11727 requests generated internally, supplied in files (for example
11728 via a CGI script) or using an internal minimal server.
11732 * Add configuration choices to get zlib compression for TLS.
11736 * Changes to Kerberos SSL for RFC 2712 compliance:
11737 1. Implemented real KerberosWrapper, instead of just using
11738 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11739 2. Implemented optional authenticator field of KerberosWrapper.
11741 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11742 and authenticator structs; see crypto/krb5/.
11744 Generalized Kerberos calls to support multiple Kerberos libraries.
11745 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11746 via Richard Levitte*
11748 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11749 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11750 values for each of the key sizes rather than having just
11751 parameters (and 'speed' generating keys each time).
11755 * Speed up EVP routines.
11758 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11759 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11760 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11761 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11763 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11764 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11765 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11768 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11770 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11774 * Added the OS2-EMX target.
11776 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11778 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11779 to support NCONF routines in extension code. New function CONF_set_nconf()
11780 to allow functions which take an NCONF to also handle the old LHASH
11781 structure: this means that the old CONF compatible routines can be
11782 retained (in particular wrt extensions) without having to duplicate the
11783 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11787 * Enhance the general user interface with mechanisms for inner control
11788 and with possibilities to have yes/no kind of prompts.
11792 * Change all calls to low level digest routines in the library and
11793 applications to use EVP. Add missing calls to HMAC_cleanup() and
11794 don't assume HMAC_CTX can be copied using memcpy().
11796 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11798 * Add the possibility to control engines through control names but with
11799 arbitrary arguments instead of just a string.
11800 Change the key loaders to take a UI_METHOD instead of a callback
11801 function pointer. NOTE: this breaks binary compatibility with earlier
11802 versions of OpenSSL [engine].
11803 Adapt the nCipher code for these new conditions and add a card insertion
11808 * Enhance the general user interface with mechanisms to better support
11809 dialog box interfaces, application-defined prompts, the possibility
11810 to use defaults (for example default passwords from somewhere else)
11811 and interrupts/cancellations.
11815 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11816 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11820 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11821 tidy up some unnecessarily weird code in 'sk_new()').
11823 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11825 * Change the key loading routines for ENGINEs to use the same kind
11826 callback (pem_password_cb) as all other routines that need this
11831 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11832 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11833 than this minimum value is recommended.
11837 * New random seeder for OpenVMS, using the system process statistics
11838 that are easily reachable.
11842 * Windows apparently can't transparently handle global
11843 variables defined in DLLs. Initialisations such as:
11845 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11847 won't compile. This is used by the any applications that need to
11848 declare their own ASN1 modules. This was fixed by adding the option
11849 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11850 needed for static libraries under Win32.
11854 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11855 setting of purpose and trust fields. New X509_STORE trust and
11856 purpose functions and tidy up setting in other SSL functions.
11860 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11861 structure. These are inherited by X509_STORE_CTX when it is
11862 initialised. This allows various defaults to be set in the
11863 X509_STORE structure (such as flags for CRL checking and custom
11864 purpose or trust settings) for functions which only use X509_STORE_CTX
11865 internally such as S/MIME.
11867 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11868 trust settings if they are not set in X509_STORE. This allows X509_STORE
11869 purposes and trust (in S/MIME for example) to override any set by default.
11871 Add command line options for CRL checking to smime, s_client and s_server
11876 * Initial CRL based revocation checking. If the CRL checking flag(s)
11877 are set then the CRL is looked up in the X509_STORE structure and
11878 its validity and signature checked, then if the certificate is found
11879 in the CRL the verify fails with a revoked error.
11881 Various new CRL related callbacks added to X509_STORE_CTX structure.
11883 Command line options added to 'verify' application to support this.
11885 This needs some additional work, such as being able to handle multiple
11886 CRLs with different times, extension based lookup (rather than just
11887 by subject name) and ultimately more complete V2 CRL extension
11892 * Add a general user interface API (crypto/ui/). This is designed
11893 to replace things like des_read_password and friends (backward
11894 compatibility functions using this new API are provided).
11895 The purpose is to remove prompting functions from the DES code
11896 section as well as provide for prompting through dialog boxes in
11897 a window system and the like.
11901 * Add "ex_data" support to ENGINE so implementations can add state at a
11902 per-structure level rather than having to store it globally.
11906 * Make it possible for ENGINE structures to be copied when retrieved by
11907 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11908 This causes the "original" ENGINE structure to act like a template,
11909 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11910 operational state can be localised to each ENGINE structure, despite the
11911 fact they all share the same "methods". New ENGINE structures returned in
11912 this case have no functional references and the return value is the single
11913 structural reference. This matches the single structural reference returned
11914 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11919 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11920 needs to match any other type at all we need to manually clear the
11925 * Changes to the "openssl engine" utility to include;
11926 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11927 about an ENGINE's available control commands.
11928 - executing control commands from command line arguments using the
11929 '-pre' and '-post' switches. '-post' is only used if '-t' is
11930 specified and the ENGINE is successfully initialised. The syntax for
11931 the individual commands are colon-separated, for example;
11932 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11936 * New dynamic control command support for ENGINEs. ENGINEs can now
11937 declare their own commands (numbers), names (strings), descriptions,
11938 and input types for run-time discovery by calling applications. A
11939 subset of these commands are implicitly classed as "executable"
11940 depending on their input type, and only these can be invoked through
11941 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11942 can be based on user input, config files, etc). The distinction is
11943 that "executable" commands cannot return anything other than a boolean
11944 result and can only support numeric or string input, whereas some
11945 discoverable commands may only be for direct use through
11946 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11947 pointers, or other custom uses. The "executable" commands are to
11948 support parameterisations of ENGINE behaviour that can be
11949 unambiguously defined by ENGINEs and used consistently across any
11950 OpenSSL-based application. Commands have been added to all the
11951 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11952 control over shared-library paths without source code alterations.
11956 * Changed all ENGINE implementations to dynamically allocate their
11957 ENGINEs rather than declaring them statically. Apart from this being
11958 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11959 this also allows the implementations to compile without using the
11960 internal engine_int.h header.
11964 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11965 'const' value. Any code that should be able to modify a RAND_METHOD
11966 should already have non-const pointers to it (ie. they should only
11967 modify their own ones).
11971 * Made a variety of little tweaks to the ENGINE code.
11972 - "atalla" and "ubsec" string definitions were moved from header files
11973 to C code. "nuron" string definitions were placed in variables
11974 rather than hard-coded - allowing parameterisation of these values
11975 later on via ctrl() commands.
11976 - Removed unused "#if 0"'d code.
11977 - Fixed engine list iteration code so it uses ENGINE_free() to release
11978 structural references.
11979 - Constified the RAND_METHOD element of ENGINE structures.
11980 - Constified various get/set functions as appropriate and added
11981 missing functions (including a catch-all ENGINE_cpy that duplicates
11982 all ENGINE values onto a new ENGINE except reference counts/state).
11983 - Removed NULL parameter checks in get/set functions. Setting a method
11984 or function to NULL is a way of cancelling out a previously set
11985 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11986 and doesn't justify the extra error symbols and code.
11987 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
11988 flags from engine_int.h to engine.h.
11989 - Changed prototypes for ENGINE handler functions (init(), finish(),
11990 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
11994 * Implement binary inversion algorithm for BN_mod_inverse in addition
11995 to the algorithm using long division. The binary algorithm can be
11996 used only if the modulus is odd. On 32-bit systems, it is faster
11997 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
11998 roughly 5-15% for 256-bit moduli), so we use it only for moduli
11999 up to 450 bits. In 64-bit environments, the binary algorithm
12000 appears to be advantageous for much longer moduli; here we use it
12001 for moduli up to 2048 bits.
12005 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12006 could not support the combine flag in choice fields.
12010 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12011 extensions from a certificate request to the certificate.
12015 * Allow multiple 'certopt' and 'nameopt' options to be separated
12016 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12017 file: this allows the display of the certificate about to be
12018 signed to be customised, to allow certain fields to be included
12019 or excluded and extension details. The old system didn't display
12020 multicharacter strings properly, omitted fields not in the policy
12021 and couldn't display additional details such as extensions.
12025 * Function EC_POINTs_mul for multiple scalar multiplication
12026 of an arbitrary number of elliptic curve points
12027 \sum scalars[i]*points[i],
12028 optionally including the generator defined for the EC_GROUP:
12029 scalar*generator + \sum scalars[i]*points[i].
12031 EC_POINT_mul is a simple wrapper function for the typical case
12032 that the point list has just one item (besides the optional
12037 * First EC_METHODs for curves over GF(p):
12039 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12040 operations and provides various method functions that can also
12041 operate with faster implementations of modular arithmetic.
12043 EC_GFp_mont_method() reuses most functions that are part of
12044 EC_GFp_simple_method, but uses Montgomery arithmetic.
12046 *Bodo Moeller; point addition and point doubling
12047 implementation directly derived from source code provided by
12048 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12050 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12051 crypto/ec/ec_lib.c):
12053 Curves are EC_GROUP objects (with an optional group generator)
12054 based on EC_METHODs that are built into the library.
12056 Points are EC_POINT objects based on EC_GROUP objects.
12058 Most of the framework would be able to handle curves over arbitrary
12059 finite fields, but as there are no obvious types for fields other
12060 than GF(p), some functions are limited to that for now.
12064 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12065 that the file contains a complete HTTP response.
12069 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12070 change the def and num file printf format specifier from "%-40sXXX"
12071 to "%-39s XXX". The latter will always guarantee a space after the
12072 field while the former will cause them to run together if the field
12073 is 40 of more characters long.
12077 * Constify the cipher and digest 'method' functions and structures
12078 and modify related functions to take constant EVP_MD and EVP_CIPHER
12083 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12084 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12088 * Modify `EVP_Digest*()` routines so they now return values. Although the
12089 internal software routines can never fail additional hardware versions
12094 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12096 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12097 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12100 ERR_R_NESTED_ASN1_ERROR
12102 ERR_R_MISSING_ASN1_EOS
12103 were 4 .. 9, conflicting with
12104 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12106 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12107 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12109 Add new error code 'ERR_R_INTERNAL_ERROR'.
12113 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12118 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12119 sets the subject name for a new request or supersedes the
12120 subject name in a given request. Formats that can be parsed are
12121 'CN=Some Name, OU=myOU, C=IT'
12123 'CN=Some Name/OU=myOU/C=IT'.
12125 Add options '-batch' and '-verbose' to 'openssl req'.
12127 *Massimiliano Pala <madwolf@hackmasters.net>*
12129 * Introduce the possibility to access global variables through
12130 functions on platform were that's the best way to handle exporting
12131 global variables in shared libraries. To enable this functionality,
12132 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12133 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12134 is normally done by Configure or something similar).
12136 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12137 in the source file (foo.c) like this:
12139 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12140 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12142 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12143 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12145 OPENSSL_DECLARE_GLOBAL(int,foo);
12146 #define foo OPENSSL_GLOBAL_REF(foo)
12147 OPENSSL_DECLARE_GLOBAL(double,bar);
12148 #define bar OPENSSL_GLOBAL_REF(bar)
12150 The #defines are very important, and therefore so is including the
12151 header file everywhere where the defined globals are used.
12153 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12154 of ASN.1 items, but that structure is a bit different.
12156 The largest change is in util/mkdef.pl which has been enhanced with
12157 better and easier to understand logic to choose which symbols should
12158 go into the Windows .def files as well as a number of fixes and code
12159 cleanup (among others, algorithm keywords are now sorted
12160 lexicographically to avoid constant rewrites).
12164 * In BN_div() keep a copy of the sign of 'num' before writing the
12165 result to 'rm' because if rm==num the value will be overwritten
12166 and produce the wrong result if 'num' is negative: this caused
12167 problems with BN_mod() and BN_nnmod().
12171 * Function OCSP_request_verify(). This checks the signature on an
12172 OCSP request and verifies the signer certificate. The signer
12173 certificate is just checked for a generic purpose and OCSP request
12178 * Add OCSP_check_validity() function to check the validity of OCSP
12179 responses. OCSP responses are prepared in real time and may only
12180 be a few seconds old. Simply checking that the current time lies
12181 between thisUpdate and nextUpdate max reject otherwise valid responses
12182 caused by either OCSP responder or client clock inaccuracy. Instead
12183 we allow thisUpdate and nextUpdate to fall within a certain period of
12184 the current time. The age of the response can also optionally be
12185 checked. Two new options -validity_period and -status_age added to
12190 * If signature or public key algorithm is unrecognized print out its
12191 OID rather that just UNKNOWN.
12195 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12196 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12197 ID to be generated from the issuer certificate alone which can then be
12198 passed to OCSP_id_issuer_cmp().
12202 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12203 ASN1 modules to export functions returning ASN1_ITEM pointers
12204 instead of the ASN1_ITEM structures themselves. This adds several
12205 new macros which allow the underlying ASN1 function/structure to
12206 be accessed transparently. As a result code should not use ASN1_ITEM
12207 references directly (such as &X509_it) but instead use the relevant
12208 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12209 use of the new ASN1 code on platforms where exporting structures
12210 is problematical (for example in shared libraries) but exporting
12211 functions returning pointers to structures is not.
12215 * Add support for overriding the generation of SSL/TLS session IDs.
12216 These callbacks can be registered either in an SSL_CTX or per SSL.
12217 The purpose of this is to allow applications to control, if they wish,
12218 the arbitrary values chosen for use as session IDs, particularly as it
12219 can be useful for session caching in multiple-server environments. A
12220 command-line switch for testing this (and any client code that wishes
12221 to use such a feature) has been added to "s_server".
12223 *Geoff Thorpe, Lutz Jaenicke*
12225 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12226 of the form '#if defined(...) || defined(...) || ...' and
12227 '#if !defined(...) && !defined(...) && ...'. This also avoids
12228 the growing number of special cases it was previously handling.
12232 * Make all configuration macros available for application by making
12233 sure they are available in opensslconf.h, by giving them names starting
12234 with `OPENSSL_` to avoid conflicts with other packages and by making
12235 sure e_os2.h will cover all platform-specific cases together with
12237 Additionally, it is now possible to define configuration/platform-
12238 specific names (called "system identities"). In the C code, these
12239 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12240 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12241 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12246 * New option -set_serial to 'req' and 'x509' this allows the serial
12247 number to use to be specified on the command line. Previously self
12248 signed certificates were hard coded with serial number 0 and the
12249 CA options of 'x509' had to use a serial number in a file which was
12254 * New options to 'ca' utility to support V2 CRL entry extensions.
12255 Currently CRL reason, invalidity date and hold instruction are
12256 supported. Add new CRL extensions to V3 code and some new objects.
12260 * New function EVP_CIPHER_CTX_set_padding() this is used to
12261 disable standard block padding (aka PKCS#5 padding) in the EVP
12262 API, which was previously mandatory. This means that the data is
12263 not padded in any way and so the total length much be a multiple
12264 of the block size, otherwise an error occurs.
12268 * Initial (incomplete) OCSP SSL support.
12272 * New function OCSP_parse_url(). This splits up a URL into its host,
12273 port and path components: primarily to parse OCSP URLs. New -url
12274 option to ocsp utility.
12278 * New nonce behavior. The return value of OCSP_check_nonce() now
12279 reflects the various checks performed. Applications can decide
12280 whether to tolerate certain situations such as an absent nonce
12281 in a response when one was present in a request: the ocsp application
12282 just prints out a warning. New function OCSP_add1_basic_nonce()
12283 this is to allow responders to include a nonce in a response even if
12284 the request is nonce-less.
12288 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12289 skipped when using openssl x509 multiple times on a single input file,
12290 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12294 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12295 set string type: to handle setting ASN1_TIME structures. Fix ca
12296 utility to correctly initialize revocation date of CRLs.
12300 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12301 the clients preferred ciphersuites and rather use its own preferences.
12302 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12303 Internet Explorer by ensuring unchanged hash method during stepup.
12304 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12308 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12309 to aes and add a new 'exist' option to print out symbols that don't
12314 * Additional options to ocsp utility to allow flags to be set and
12315 additional certificates supplied.
12319 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12320 OCSP client a number of certificate to only verify the response
12325 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12326 handle the new API. Currently only ECB, CBC modes supported. Add new
12329 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12330 Encryption Standard (AES) Ciphersuites for Transport Layer
12331 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12332 not enabled by default and were not part of the "ALL" ciphersuite
12333 alias because they were not yet official; they could be
12334 explicitly requested by specifying the "AESdraft" ciphersuite
12335 group alias. In the final release of OpenSSL 0.9.7, the group
12336 alias is called "AES" and is part of "ALL".)
12338 *Ben Laurie, Steve Henson, Bodo Moeller*
12340 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12341 request to response.
12345 * Functions for OCSP responders. OCSP_request_onereq_count(),
12346 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12347 extract information from a certificate request. OCSP_response_create()
12348 creates a response and optionally adds a basic response structure.
12349 OCSP_basic_add1_status() adds a complete single response to a basic
12350 response and returns the OCSP_SINGLERESP structure just added (to allow
12351 extensions to be included for example). OCSP_basic_add1_cert() adds a
12352 certificate to a basic response and OCSP_basic_sign() signs a basic
12353 response with various flags. New helper functions ASN1_TIME_check()
12354 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12355 (converts ASN1_TIME to GeneralizedTime).
12359 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12360 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12361 structure from a certificate. X509_pubkey_digest() digests the public_key
12362 contents: this is used in various key identifiers.
12366 * Make sk_sort() tolerate a NULL argument.
12368 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12370 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12371 passed by the function are trusted implicitly. If any of them signed the
12372 response then it is assumed to be valid and is not verified.
12376 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12377 to data. This was previously part of the PKCS7 ASN1 code. This
12378 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12379 *Steve Henson, reported by Kenneth R. Robinette
12380 <support@securenetterm.com>*
12382 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12383 routines: without these tracing memory leaks is very painful.
12384 Fix leaks in PKCS12 and PKCS7 routines.
12388 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12389 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12390 effectively meant GeneralizedTime would never be used. Now it
12391 is initialised to -1 but X509_time_adj() now has to check the value
12392 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12393 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12394 *Steve Henson, reported by Kenneth R. Robinette
12395 <support@securenetterm.com>*
12397 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12398 result in a zero length in the ASN1_INTEGER structure which was
12399 not consistent with the structure when d2i_ASN1_INTEGER() was used
12400 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12401 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12402 where it did not print out a minus for negative ASN1_INTEGER.
12406 * Add summary printout to ocsp utility. The various functions which
12407 convert status values to strings have been renamed to:
12408 OCSP_response_status_str(), OCSP_cert_status_str() and
12409 OCSP_crl_reason_str() and are no longer static. New options
12410 to verify nonce values and to disable verification. OCSP response
12411 printout format cleaned up.
12415 * Add additional OCSP certificate checks. These are those specified
12416 in RFC2560. This consists of two separate checks: the CA of the
12417 certificate being checked must either be the OCSP signer certificate
12418 or the issuer of the OCSP signer certificate. In the latter case the
12419 OCSP signer certificate must contain the OCSP signing extended key
12420 usage. This check is performed by attempting to match the OCSP
12421 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12422 in the OCSP_CERTID structures of the response.
12426 * Initial OCSP certificate verification added to OCSP_basic_verify()
12427 and related routines. This uses the standard OpenSSL certificate
12428 verify routines to perform initial checks (just CA validity) and
12429 to obtain the certificate chain. Then additional checks will be
12430 performed on the chain. Currently the root CA is checked to see
12431 if it is explicitly trusted for OCSP signing. This is used to set
12432 a root CA as a global signing root: that is any certificate that
12433 chains to that CA is an acceptable OCSP signing certificate.
12437 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12438 extensions from a separate configuration file.
12439 As when reading extensions from the main configuration file,
12440 the '-extensions ...' option may be used for specifying the
12443 *Massimiliano Pala <madwolf@comune.modena.it>*
12445 * New OCSP utility. Allows OCSP requests to be generated or
12446 read. The request can be sent to a responder and the output
12447 parsed, outputted or printed in text form. Not complete yet:
12448 still needs to check the OCSP response validity.
12452 * New subcommands for 'openssl ca':
12453 `openssl ca -status <serial>` prints the status of the cert with
12454 the given serial number (according to the index file).
12455 `openssl ca -updatedb` updates the expiry status of certificates
12458 *Massimiliano Pala <madwolf@comune.modena.it>*
12460 * New '-newreq-nodes' command option to CA.pl. This is like
12461 '-newreq', but calls 'openssl req' with the '-nodes' option
12462 so that the resulting key is not encrypted.
12464 *Damien Miller <djm@mindrot.org>*
12466 * New configuration for the GNU Hurd.
12468 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12470 * Initial code to implement OCSP basic response verify. This
12471 is currently incomplete. Currently just finds the signer's
12472 certificate and verifies the signature on the response.
12476 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12477 value of OPENSSLDIR. This is available via the new '-d' option
12478 to 'openssl version', and is also included in 'openssl version -a'.
12482 * Allowing defining memory allocation callbacks that will be given
12483 file name and line number information in additional arguments
12484 (a `const char*` and an int). The basic functionality remains, as
12485 well as the original possibility to just replace malloc(),
12486 realloc() and free() by functions that do not know about these
12487 additional arguments. To register and find out the current
12488 settings for extended allocation functions, the following
12489 functions are provided:
12491 CRYPTO_set_mem_ex_functions
12492 CRYPTO_set_locked_mem_ex_functions
12493 CRYPTO_get_mem_ex_functions
12494 CRYPTO_get_locked_mem_ex_functions
12496 These work the same way as CRYPTO_set_mem_functions and friends.
12497 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12498 extended allocation function is enabled.
12499 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12500 a conventional allocation function is enabled.
12502 *Richard Levitte, Bodo Moeller*
12504 * Finish off removing the remaining LHASH function pointer casts.
12505 There should no longer be any prototype-casting required when using
12506 the LHASH abstraction, and any casts that remain are "bugs". See
12507 the callback types and macros at the head of lhash.h for details
12508 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12512 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12513 If /dev/[u]random devices are not available or do not return enough
12514 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12516 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12517 /etc/entropy will be queried once each in this sequence, querying stops
12518 when enough entropy was collected without querying more sockets.
12522 * Change the Unix RAND_poll() variant to be able to poll several
12523 random devices, as specified by DEVRANDOM, until a sufficient amount
12524 of data has been collected. We spend at most 10 ms on each file
12525 (select timeout) and read in non-blocking mode. DEVRANDOM now
12526 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12527 (previously it was just the string "/dev/urandom"), so on typical
12528 platforms the 10 ms delay will never occur.
12529 Also separate out the Unix variant to its own file, rand_unix.c.
12530 For VMS, there's a currently-empty rand_vms.c.
12534 * Move OCSP client related routines to ocsp_cl.c. These
12535 provide utility functions which an application needing
12536 to issue a request to an OCSP responder and analyse the
12537 response will typically need: as opposed to those which an
12538 OCSP responder itself would need which will be added later.
12540 OCSP_request_sign() signs an OCSP request with an API similar
12541 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12542 response. OCSP_response_get1_basic() extracts basic response
12543 from response. OCSP_resp_find_status(): finds and extracts status
12544 information from an OCSP_CERTID structure (which will be created
12545 when the request structure is built). These are built from lower
12546 level functions which work on OCSP_SINGLERESP structures but
12547 won't normally be used unless the application wishes to examine
12548 extensions in the OCSP response for example.
12550 Replace nonce routines with a pair of functions.
12551 OCSP_request_add1_nonce() adds a nonce value and optionally
12552 generates a random value. OCSP_check_nonce() checks the
12553 validity of the nonce in an OCSP response.
12557 * Change function OCSP_request_add() to OCSP_request_add0_id().
12558 This doesn't copy the supplied OCSP_CERTID and avoids the
12559 need to free up the newly created id. Change return type
12560 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12561 This can then be used to add extensions to the request.
12562 Deleted OCSP_request_new(), since most of its functionality
12563 is now in OCSP_REQUEST_new() (and the case insensitive name
12564 clash) apart from the ability to set the request name which
12565 will be added elsewhere.
12569 * Update OCSP API. Remove obsolete extensions argument from
12570 various functions. Extensions are now handled using the new
12571 OCSP extension code. New simple OCSP HTTP function which
12572 can be used to send requests and parse the response.
12576 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12577 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12578 uses the special reorder version of SET OF to sort the attributes
12579 and reorder them to match the encoded order. This resolves a long
12580 standing problem: a verify on a PKCS7 structure just after signing
12581 it used to fail because the attribute order did not match the
12582 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12583 it uses the received order. This is necessary to tolerate some broken
12584 software that does not order SET OF. This is handled by encoding
12585 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12586 to produce the required SET OF.
12590 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12591 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12592 files to get correct declarations of the ASN.1 item variables.
12596 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12597 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12598 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12599 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12600 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12601 ASN1_ITEM and no wrapper functions.
12605 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12606 replace the old function pointer based I/O routines. Change most of
12607 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12611 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12612 lines, recognize more "algorithms" that can be deselected, and make
12613 it complain about algorithm deselection that isn't recognised.
12617 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12618 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12619 to use new functions. Add NO_ASN1_OLD which can be set to remove
12620 some old style ASN1 functions: this can be used to determine if old
12621 code will still work when these eventually go away.
12625 * New extension functions for OCSP structures, these follow the
12626 same conventions as certificates and CRLs.
12630 * New function X509V3_add1_i2d(). This automatically encodes and
12631 adds an extension. Its behaviour can be customised with various
12632 flags to append, replace or delete. Various wrappers added for
12633 certificates and CRLs.
12637 * Fix to avoid calling the underlying ASN1 print routine when
12638 an extension cannot be parsed. Correct a typo in the
12639 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12643 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12644 entries for variables.
12648 * Add functionality to apps/openssl.c for detecting locking
12649 problems: As the program is single-threaded, all we have
12650 to do is register a locking callback using an array for
12651 storing which locks are currently held by the program.
12655 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12656 SSL_get_ex_data_X509_STORE_idx(), which is used in
12657 ssl_verify_cert_chain() and thus can be called at any time
12658 during TLS/SSL handshakes so that thread-safety is essential.
12659 Unfortunately, the ex_data design is not at all suited
12660 for multi-threaded use, so it probably should be abolished.
12664 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12666 *Broadcom, tweaked and integrated by Geoff Thorpe*
12668 * Move common extension printing code to new function
12669 X509V3_print_extensions(). Reorganise OCSP print routines and
12670 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12674 * New function X509_signature_print() to remove duplication in some
12679 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12680 set (this was treated exactly the same as SET OF previously). This
12681 is used to reorder the STACK representing the structure to match the
12682 encoding. This will be used to get round a problem where a PKCS7
12683 structure which was signed could not be verified because the STACK
12684 order did not reflect the encoded order.
12688 * Reimplement the OCSP ASN1 module using the new code.
12692 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12693 for its ASN1 operations. The old style function pointers still exist
12694 for now but they will eventually go away.
12698 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12699 completely replaces the old ASN1 functionality with a table driven
12700 encoder and decoder which interprets an ASN1_ITEM structure describing
12701 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12702 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12703 has also been converted to the new form.
12707 * Change BN_mod_exp_recp so that negative moduli are tolerated
12708 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12709 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12710 for negative moduli.
12714 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12715 of not touching the result's sign bit.
12719 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12724 * Changed the LHASH code to use prototypes for callbacks, and created
12725 macros to declare and implement thin (optionally static) functions
12726 that provide type-safety and avoid function pointer casting for the
12727 type-specific callbacks.
12731 * Added Kerberos Cipher Suites to be used with TLS, as written in
12733 *Veers Staats <staatsvr@asc.hpc.mil>,
12734 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12736 * Reformat the FAQ so the different questions and answers can be divided
12737 in sections depending on the subject.
12741 * Have the zlib compression code load ZLIB.DLL dynamically under
12746 * New function BN_mod_sqrt for computing square roots modulo a prime
12747 (using the probabilistic Tonelli-Shanks algorithm unless
12748 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12749 be handled deterministically).
12751 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12753 * Make BN_mod_inverse faster by explicitly handling small quotients
12754 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12755 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12759 * New function BN_kronecker.
12763 * Fix BN_gcd so that it works on negative inputs; the result is
12764 positive unless both parameters are zero.
12765 Previously something reasonably close to an infinite loop was
12766 possible because numbers could be growing instead of shrinking
12767 in the implementation of Euclid's algorithm.
12771 * Fix BN_is_word() and BN_is_one() macros to take into account the
12772 sign of the number in question.
12774 Fix BN_is_word(a,w) to work correctly for w == 0.
12776 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12777 because its test if the absolute value of 'a' equals 'w'.
12778 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12779 it exists mostly for use in the implementations of BN_is_zero(),
12780 BN_is_one(), and BN_is_word().
12784 * New function BN_swap.
12788 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12789 the exponentiation functions are more likely to produce reasonable
12790 results on negative inputs.
12794 * Change BN_mod_mul so that the result is always non-negative.
12795 Previously, it could be negative if one of the factors was negative;
12796 I don't think anyone really wanted that behaviour.
12800 * Move `BN_mod_...` functions into new file crypto/bn/bn_mod.c
12801 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12802 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12803 and add new functions:
12812 BN_mod_lshift1_quick
12814 BN_mod_lshift_quick
12816 These functions always generate non-negative results.
12818 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12819 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12821 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12822 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12823 be reduced modulo m.
12825 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12828 The following entry accidentally appeared in the CHANGES file
12829 distributed with OpenSSL 0.9.7. The modifications described in
12830 it do *not* apply to OpenSSL 0.9.7.
12832 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12833 was actually never needed) and in BN_mul(). The removal in BN_mul()
12834 required a small change in bn_mul_part_recursive() and the addition
12835 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12836 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12837 bn_sub_words() and bn_add_words() except they take arrays with
12843 * In 'openssl passwd', verify passwords read from the terminal
12844 unless the '-salt' option is used (which usually means that
12845 verification would just waste user's time since the resulting
12846 hash is going to be compared with some given password hash)
12847 or the new '-noverify' option is used.
12849 This is an incompatible change, but it does not affect
12850 non-interactive use of 'openssl passwd' (passwords on the command
12851 line, '-stdin' option, '-in ...' option) and thus should not
12852 cause any problems.
12856 * Remove all references to RSAref, since there's no more need for it.
12860 * Make DSO load along a path given through an environment variable
12861 (SHLIB_PATH) with shl_load().
12865 * Constify the ENGINE code as a result of BIGNUM constification.
12866 Also constify the RSA code and most things related to it. In a
12867 few places, most notable in the depth of the ASN.1 code, ugly
12868 casts back to non-const were required (to be solved at a later
12873 * Make it so the openssl application has all engines loaded by default.
12877 * Constify the BIGNUM routines a little more.
12881 * Add the following functions:
12883 ENGINE_load_cswift()
12885 ENGINE_load_atalla()
12886 ENGINE_load_nuron()
12887 ENGINE_load_builtin_engines()
12889 That way, an application can itself choose if external engines that
12890 are built-in in OpenSSL shall ever be used or not. The benefit is
12891 that applications won't have to be linked with libdl or other dso
12892 libraries unless it's really needed.
12894 Changed 'openssl engine' to load all engines on demand.
12895 Changed the engine header files to avoid the duplication of some
12896 declarations (they differed!).
12900 * 'openssl engine' can now list capabilities.
12904 * Better error reporting in 'openssl engine'.
12908 * Never call load_dh_param(NULL) in s_server.
12912 * Add engine application. It can currently list engines by name and
12913 identity, and test if they are actually available.
12917 * Improve RPM specification file by forcing symbolic linking and making
12918 sure the installed documentation is also owned by root.root.
12920 *Damien Miller <djm@mindrot.org>*
12922 * Give the OpenSSL applications more possibilities to make use of
12923 keys (public as well as private) handled by engines.
12927 * Add OCSP code that comes from CertCo.
12931 * Add VMS support for the Rijndael code.
12935 * Added untested support for Nuron crypto accelerator.
12939 * Add support for external cryptographic devices. This code was
12940 previously distributed separately as the "engine" branch.
12942 *Geoff Thorpe, Richard Levitte*
12944 * Rework the filename-translation in the DSO code. It is now possible to
12945 have far greater control over how a "name" is turned into a filename
12946 depending on the operating environment and any oddities about the
12947 different shared library filenames on each system.
12951 * Support threads on FreeBSD-elf in Configure.
12955 * Fix for SHA1 assembly problem with MASM: it produces
12956 warnings about corrupt line number information when assembling
12957 with debugging information. This is caused by the overlapping
12960 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12963 NCONF_get_number() has no error checking at all. As a replacement,
12964 NCONF_get_number_e() is defined (`_e` for "error checking") and is
12965 promoted strongly. The old NCONF_get_number is kept around for
12966 binary backward compatibility.
12967 Make it possible for methods to load from something other than a BIO,
12968 by providing a function pointer that is given a name instead of a BIO.
12969 For example, this could be used to load configuration data from an
12974 * Fix for non blocking accept BIOs. Added new I/O special reason
12975 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12976 with non blocking I/O was not possible because no retry code was
12977 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12982 * Added the beginnings of Rijndael support.
12986 * Fix for bug in DirectoryString mask setting. Add support for
12987 X509_NAME_print_ex() in 'req' and X509_print_ex() function
12988 to allow certificate printing to more controllable, additional
12989 'certopt' option to 'x509' to allow new printing options to be
12994 * Clean old EAY MD5 hack from e_os.h.
12998 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13000 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13001 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13003 *Joe Orton, Steve Henson*
13005 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13007 * Fix additional bug revealed by the NISCC test suite:
13009 Stop bug triggering large recursion when presented with
13010 certain ASN.1 tags [CVE-2003-0851][]
13014 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13016 * Fix various bugs revealed by running the NISCC test suite:
13018 Stop out of bounds reads in the ASN1 code when presented with
13019 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13021 If verify callback ignores invalid public key errors don't try to check
13022 certificate signature with the NULL public key.
13026 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13027 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13032 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13033 extra data after the compression methods not only for TLS 1.0
13034 but also for SSL 3.0 (as required by the specification).
13036 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13038 * Change X509_certificate_type() to mark the key as exported/exportable
13039 when it's 512 *bits* long, not 512 bytes.
13043 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13045 * Countermeasure against the Klima-Pokorny-Rosa extension of
13046 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13047 a protocol version number mismatch like a decryption error
13048 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13052 * Turn on RSA blinding by default in the default implementation
13053 to avoid a timing attack. Applications that don't want it can call
13054 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13055 They would be ill-advised to do so in most cases.
13057 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13059 * Change RSA blinding code so that it works when the PRNG is not
13060 seeded (in this case, the secret RSA exponent is abused as
13061 an unpredictable seed -- if it is not unpredictable, there
13062 is no point in blinding anyway). Make RSA blinding thread-safe
13063 by remembering the creator's thread ID in rsa->blinding and
13064 having all other threads use local one-time blinding factors
13065 (this requires more computation than sharing rsa->blinding, but
13066 avoids excessive locking; and if an RSA object is not shared
13067 between threads, blinding will still be very fast).
13071 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13073 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13074 via timing by performing a MAC computation even if incorrect
13075 block cipher padding has been found. This is a countermeasure
13076 against active attacks where the attacker has to distinguish
13077 between bad padding and a MAC verification error. [CVE-2003-0078][]
13079 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13080 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13081 Martin Vuagnoux (EPFL, Ilion)*
13083 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13085 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13086 memory from its contents. This is done with a counter that will
13087 place alternating values in each byte. This can be used to solve
13088 two issues: 1) the removal of calls to memset() by highly optimizing
13089 compilers, and 2) cleansing with other values than 0, since those can
13090 be read through on certain media, for example a swap space on disk.
13094 * Bugfix: client side session caching did not work with external caching,
13095 because the session->cipher setting was not restored when reloading
13096 from the external cache. This problem was masked, when
13097 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13098 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13102 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13103 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13105 *Zeev Lieber <zeev-l@yahoo.com>*
13107 * Undo an undocumented change introduced in 0.9.6e which caused
13108 repeated calls to OpenSSL_add_all_ciphers() and
13109 OpenSSL_add_all_digests() to be ignored, even after calling
13114 * Change the default configuration reader to deal with last line not
13115 being properly terminated.
13119 * Change X509_NAME_cmp() so it applies the special rules on handling
13120 DN values that are of type PrintableString, as well as RDNs of type
13121 emailAddress where the value has the type ia5String.
13123 *stefank@valicert.com via Richard Levitte*
13125 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13126 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13127 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13128 the bitwise-OR of the two for use by the majority of applications
13129 wanting this behaviour, and update the docs. The documented
13130 behaviour and actual behaviour were inconsistent and had been
13131 changing anyway, so this is more a bug-fix than a behavioural
13134 *Geoff Thorpe, diagnosed by Nadav Har'El*
13136 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13137 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13141 * Fix initialization code race conditions in
13142 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13143 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13144 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13145 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13146 ssl2_get_cipher_by_char(),
13147 ssl3_get_cipher_by_char().
13149 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13151 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13152 the cached sessions are flushed, as the remove_cb() might use ex_data
13153 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13154 (see [openssl.org #212]).
13156 *Geoff Thorpe, Lutz Jaenicke*
13158 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13159 length, instead of the encoding length to d2i_ASN1_OBJECT.
13163 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13165 * [In 0.9.6g-engine release:]
13166 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13168 *Lynn Gazis <lgazis@rainbow.com>*
13170 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13172 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13173 and get fix the header length calculation.
13174 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13175 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13177 * Use proper error handling instead of 'assertions' in buffer
13178 overflow checks added in 0.9.6e. This prevents DoS (the
13179 assertions could call abort()).
13181 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13183 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13185 * Add various sanity checks to asn1_get_length() to reject
13186 the ASN1 length bytes if they exceed sizeof(long), will appear
13187 negative or the content length exceeds the length of the
13190 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13192 * Fix cipher selection routines: ciphers without encryption had no flags
13193 for the cipher strength set and where therefore not handled correctly
13194 by the selection routines (PR #130).
13198 * Fix EVP_dsa_sha macro.
13203 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13204 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13205 that was added in OpenSSL 0.9.6d.
13207 As the countermeasure turned out to be incompatible with some
13208 broken SSL implementations, the new option is part of SSL_OP_ALL.
13209 SSL_OP_ALL is usually employed when compatibility with weird SSL
13210 implementations is desired (e.g. '-bugs' option to 's_client' and
13211 's_server'), so the new option is automatically set in many
13216 * Changes in security patch:
13218 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13219 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13220 Air Force Materiel Command, USAF, under agreement number
13223 * Add various sanity checks to asn1_get_length() to reject
13224 the ASN1 length bytes if they exceed sizeof(long), will appear
13225 negative or the content length exceeds the length of the
13226 supplied buffer. [CVE-2002-0659][]
13228 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13230 * Assertions for various potential buffer overflows, not known to
13231 happen in practice.
13233 *Ben Laurie (CHATS)*
13235 * Various temporary buffers to hold ASCII versions of integers were
13236 too small for 64 bit platforms. [CVE-2002-0655][]
13237 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13239 * Remote buffer overflow in SSL3 protocol - an attacker could
13240 supply an oversized session ID to a client. [CVE-2002-0656][]
13242 *Ben Laurie (CHATS)*
13244 * Remote buffer overflow in SSL2 protocol - an attacker could
13245 supply an oversized client master key. [CVE-2002-0656][]
13247 *Ben Laurie (CHATS)*
13249 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13251 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13252 encoded as NULL) with id-dsa-with-sha1.
13254 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13256 * Check various `X509_...()` return values in apps/req.c.
13258 *Nils Larsch <nla@trustcenter.de>*
13260 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13261 an end-of-file condition would erroneously be flagged, when the CRLF
13262 was just at the end of a processed block. The bug was discovered when
13263 processing data through a buffering memory BIO handing the data to a
13264 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13265 <ptsekov@syntrex.com> and Nedelcho Stanev.
13269 * Implement a countermeasure against a vulnerability recently found
13270 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13271 before application data chunks to avoid the use of known IVs
13272 with data potentially chosen by the attacker.
13276 * Fix length checks in ssl3_get_client_hello().
13280 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13281 to prevent ssl3_read_internal() from incorrectly assuming that
13282 ssl3_read_bytes() found application data while handshake
13283 processing was enabled when in fact s->s3->in_read_app_data was
13284 merely automatically cleared during the initial handshake.
13286 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13288 * Fix object definitions for Private and Enterprise: they were not
13289 recognized in their shortname (=lowercase) representation. Extend
13290 obj_dat.pl to issue an error when using undefined keywords instead
13291 of silently ignoring the problem (Svenning Sorensen
13292 <sss@sss.dnsalias.net>).
13296 * Fix DH_generate_parameters() so that it works for 'non-standard'
13297 generators, i.e. generators other than 2 and 5. (Previously, the
13298 code did not properly initialise the 'add' and 'rem' values to
13299 BN_generate_prime().)
13301 In the new general case, we do not insist that 'generator' is
13302 actually a primitive root: This requirement is rather pointless;
13303 a generator of the order-q subgroup is just as good, if not
13308 * Map new X509 verification errors to alerts. Discovered and submitted by
13309 Tom Wu <tom@arcot.com>.
13313 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13314 returning non-zero before the data has been completely received
13315 when using non-blocking I/O.
13317 *Bodo Moeller; problem pointed out by John Hughes*
13319 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13321 *Ben Laurie, Lutz Jaenicke*
13323 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13324 Yoram Zahavi <YoramZ@gilian.com>).
13328 * Add information about CygWin 1.3 and on, and preserve proper
13329 configuration for the versions before that.
13331 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13333 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13334 check whether we deal with a copy of a session and do not delete from
13335 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13336 <izhar@checkpoint.com>.
13340 * Do not store session data into the internal session cache, if it
13341 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13342 flag is set). Proposed by Aslam <aslam@funk.com>.
13346 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13351 * [In 0.9.6d-engine release:]
13352 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13354 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13356 * Add the configuration target linux-s390x.
13358 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13360 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13361 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13362 variable as an indication that a ClientHello message has been
13363 received. As the flag value will be lost between multiple
13364 invocations of ssl3_accept when using non-blocking I/O, the
13365 function may not be aware that a handshake has actually taken
13366 place, thus preventing a new session from being added to the
13369 To avoid this problem, we now set s->new_session to 2 instead of
13370 using a local variable.
13372 *Lutz Jaenicke, Bodo Moeller*
13374 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13375 if the SSL_R_LENGTH_MISMATCH error is detected.
13377 *Geoff Thorpe, Bodo Moeller*
13379 * New 'shared_ldflag' column in Configure platform table.
13383 * Fix EVP_CIPHER_mode macro.
13385 *"Dan S. Camper" <dan@bti.net>*
13387 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13388 type, we must throw them away by setting rr->length to 0.
13390 *D P Chang <dpc@qualys.com>*
13392 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13394 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13395 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13396 worked incorrectly for those cases where range = `10..._2` and
13397 `3*range` is two bits longer than range.)
13401 * Only add signing time to PKCS7 structures if it is not already
13406 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13407 OBJ_ld_ce should be OBJ_id_ce.
13408 Also some ip-pda OIDs in crypto/objects/objects.txt were
13409 incorrect (cf. RFC 3039).
13411 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13413 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13414 returns early because it has nothing to do.
13416 *Andy Schneider <andy.schneider@bjss.co.uk>*
13418 * [In 0.9.6c-engine release:]
13419 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13421 *Andy Schneider <andy.schneider@bjss.co.uk>*
13423 * [In 0.9.6c-engine release:]
13424 Add support for Cryptographic Appliance's keyserver technology.
13425 (Use engine 'keyclient')
13427 *Cryptographic Appliances and Geoff Thorpe*
13429 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13430 is called via tools/c89.sh because arguments have to be
13431 rearranged (all '-L' options must appear before the first object
13434 *Richard Shapiro <rshapiro@abinitio.com>*
13436 * [In 0.9.6c-engine release:]
13437 Add support for Broadcom crypto accelerator cards, backported
13440 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13442 * [In 0.9.6c-engine release:]
13443 Add support for SureWare crypto accelerator cards from
13444 Baltimore Technologies. (Use engine 'sureware')
13446 *Baltimore Technologies and Mark Cox*
13448 * [In 0.9.6c-engine release:]
13449 Add support for crypto accelerator cards from Accelerated
13450 Encryption Processing, www.aep.ie. (Use engine 'aep')
13452 *AEP Inc. and Mark Cox*
13454 * Add a configuration entry for gcc on UnixWare.
13456 *Gary Benson <gbenson@redhat.com>*
13458 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13459 messages are stored in a single piece (fixed-length part and
13460 variable-length part combined) and fix various bugs found on the way.
13464 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13465 instead. BIO_gethostbyname() does not know what timeouts are
13466 appropriate, so entries would stay in cache even when they have
13468 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13470 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13471 faced with a pathologically small ClientHello fragment that does
13472 not contain client_version: Instead of aborting with an error,
13473 simply choose the highest available protocol version (i.e.,
13474 TLS 1.0 unless it is disabled). In practice, ClientHello
13475 messages are never sent like this, but this change gives us
13476 strictly correct behaviour at least for TLS.
13480 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13481 never resets s->method to s->ctx->method when called from within
13482 one of the SSL handshake functions.
13484 *Bodo Moeller; problem pointed out by Niko Baric*
13486 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13487 (sent using the client's version number) if client_version is
13488 smaller than the protocol version in use. Also change
13489 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13490 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13491 the client will at least see that alert.
13495 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13500 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13501 client receives HelloRequest while in a handshake.
13503 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13505 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13506 should end in 'break', not 'goto end' which circumvents various
13507 cleanups done in state SSL_ST_OK. But session related stuff
13508 must be disabled for SSL_ST_OK in the case that we just sent a
13511 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13512 before just sending a HelloRequest.
13514 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13516 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13517 reveal whether illegal block cipher padding was found or a MAC
13518 verification error occurred. (Neither SSLerr() codes nor alerts
13519 are directly visible to potential attackers, but the information
13520 may leak via logfiles.)
13522 Similar changes are not required for the SSL 2.0 implementation
13523 because the number of padding bytes is sent in clear for SSL 2.0,
13524 and the extra bytes are just ignored. However ssl/s2_pkt.c
13525 failed to verify that the purported number of padding bytes is in
13530 * Add OpenUNIX-8 support including shared libraries
13531 (Boyd Lynn Gerber <gerberb@zenez.com>).
13535 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13536 'wristwatch attack' using huge encoding parameters (cf.
13537 James H. Manger's CRYPTO 2001 paper). Note that the
13538 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13539 encoding parameters and hence was not vulnerable.
13543 * BN_sqr() bug fix.
13545 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13547 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13548 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13549 followed by modular reduction.
13551 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13553 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13554 equivalent based on BN_pseudo_rand() instead of BN_rand().
13558 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13559 This function was broken, as the check for a new client hello message
13560 to handle SGC did not allow these large messages.
13561 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13565 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13569 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13570 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13574 * Rework the configuration and shared library support for Tru64 Unix.
13575 The configuration part makes use of modern compiler features and
13576 still retains old compiler behavior for those that run older versions
13577 of the OS. The shared library support part includes a variant that
13578 uses the RPATH feature, and is available through the special
13579 configuration target "alpha-cc-rpath", which will never be selected
13582 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13584 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13585 with the same message size as in ssl3_get_certificate_request().
13586 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13587 messages might inadvertently be reject as too long.
13589 *Petr Lampa <lampa@fee.vutbr.cz>*
13591 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13595 * Modified SSL library such that the verify_callback that has been set
13596 specifically for an SSL object with SSL_set_verify() is actually being
13597 used. Before the change, a verify_callback set with this function was
13598 ignored and the verify_callback() set in the SSL_CTX at the time of
13599 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13600 to allow the necessary settings.
13604 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13605 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13606 done automatically (in contradiction to the requirements of the C
13607 standard). This made problems when used from OpenSSH.
13611 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13612 dh->length and always used
13614 BN_rand_range(priv_key, dh->p).
13616 BN_rand_range() is not necessary for Diffie-Hellman, and this
13617 specific range makes Diffie-Hellman unnecessarily inefficient if
13618 dh->length (recommended exponent length) is much smaller than the
13619 length of dh->p. We could use BN_rand_range() if the order of
13620 the subgroup was stored in the DH structure, but we only have
13625 BN_rand(priv_key, l, ...)
13627 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13634 RSA_eay_public_encrypt
13635 RSA_eay_private_decrypt
13636 RSA_eay_private_encrypt (signing)
13637 RSA_eay_public_decrypt (signature verification)
13639 (default implementations for RSA_public_encrypt,
13640 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13641 always reject numbers >= n.
13645 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13646 to synchronize access to 'locking_thread'. This is necessary on
13647 systems where access to 'locking_thread' (an 'unsigned long'
13648 variable) is not atomic.
13652 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13653 *before* setting the 'crypto_lock_rand' flag. The previous code had
13654 a race condition if 0 is a valid thread ID.
13656 *Travis Vitek <vitek@roguewave.com>*
13658 * Add support for shared libraries under Irix.
13660 *Albert Chin-A-Young <china@thewrittenword.com>*
13662 * Add configuration option to build on Linux on both big-endian and
13663 little-endian MIPS.
13665 *Ralf Baechle <ralf@uni-koblenz.de>*
13667 * Add the possibility to create shared libraries on HP-UX.
13671 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13673 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13674 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13675 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13676 PRNG state recovery was possible based on the output of
13677 one PRNG request appropriately sized to gain knowledge on
13678 'md' followed by enough consecutive 1-byte PRNG requests
13679 to traverse all of 'state'.
13681 1. When updating 'md_local' (the current thread's copy of 'md')
13682 during PRNG output generation, hash all of the previous
13683 'md_local' value, not just the half used for PRNG output.
13685 2. Make the number of bytes from 'state' included into the hash
13686 independent from the number of PRNG bytes requested.
13688 The first measure alone would be sufficient to avoid
13689 Markku-Juhani's attack. (Actually it had never occurred
13690 to me that the half of 'md_local' used for chaining was the
13691 half from which PRNG output bytes were taken -- I had always
13692 assumed that the secret half would be used.) The second
13693 measure makes sure that additional data from 'state' is never
13694 mixed into 'md_local' in small portions; this heuristically
13695 further strengthens the PRNG.
13699 * Fix crypto/bn/asm/mips3.s.
13703 * When only the key is given to "enc", the IV is undefined. Print out
13704 an error message in this case.
13708 * Handle special case when X509_NAME is empty in X509 printing routines.
13712 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13713 positive and less than q.
13717 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13718 used: it isn't thread safe and the add_lock_callback should handle
13721 *Paul Rose <Paul.Rose@bridge.com>*
13723 * Verify that incoming data obeys the block size in
13724 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13730 *Ulf Möller, Bodo Möller*
13732 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13733 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13734 when fixing the server behaviour for backwards-compatible 'client
13735 hello' messages. (Note that the attack is impractical against
13736 SSL 3.0 and TLS 1.0 anyway because length and version checking
13737 means that the probability of guessing a valid ciphertext is
13738 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13741 Before 0.9.5, the countermeasure (hide the error by generating a
13742 random 'decryption result') did not work properly because
13743 ERR_clear_error() was missing, meaning that SSL_get_error() would
13744 detect the supposedly ignored error.
13746 Both problems are now fixed.
13750 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13751 (previously it was 1024).
13755 * Fix for compatibility mode trust settings: ignore trust settings
13756 unless some valid trust or reject settings are present.
13760 * Fix for blowfish EVP: its a variable length cipher.
13764 * Fix various bugs related to DSA S/MIME verification. Handle missing
13765 parameters in DSA public key structures and return an error in the
13766 DSA routines if parameters are absent.
13770 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13771 in the current directory if neither $RANDFILE nor $HOME was set.
13772 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13773 caused some confusion to Windows users who haven't defined $HOME.
13774 Thus RAND_file_name() is changed again: e_os.h can define a
13775 DEFAULT_HOME, which will be used if $HOME is not set.
13776 For Windows, we use "C:"; on other platforms, we still require
13777 environment variables.
13779 * Move 'if (!initialized) RAND_poll()' into regions protected by
13780 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13781 having multiple threads call RAND_poll() concurrently.
13785 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13786 combination of a flag and a thread ID variable.
13787 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13788 flag), *other* threads can enter ssleay_add_bytes without obeying
13789 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13790 that they do not hold after the first thread unsets add_do_not_lock).
13794 * Change bctest again: '-x' expressions are not available in all
13795 versions of 'test'.
13799 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13801 * Fix a couple of memory leaks in PKCS7_dataDecode()
13803 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13805 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13806 the default extension for executables, if any. Also, make the perl
13807 scripts that use symlink() to test if it really exists and use "cp"
13808 if it doesn't. All this made OpenSSL compilable and installable in
13813 * Fix for asn1_GetSequence() for indefinite length constructed data.
13814 If SEQUENCE is length is indefinite just set c->slen to the total
13815 amount of data available.
13817 *Steve Henson, reported by shige@FreeBSD.org*
13819 *This change does not apply to 0.9.7.*
13821 * Change bctest to avoid here-documents inside command substitution
13822 (workaround for FreeBSD /bin/sh bug).
13823 For compatibility with Ultrix, avoid shell functions (introduced
13824 in the bctest version that searches along $PATH).
13828 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13829 with des_encrypt() defined on some operating systems, like Solaris
13834 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13835 On the Importance of Eliminating Errors in Cryptographic
13836 Computations, J. Cryptology 14 (2001) 2, 101-119,
13837 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13841 * MIPS assembler BIGNUM division bug fix.
13845 * Disabled incorrect Alpha assembler code.
13849 * Fix PKCS#7 decode routines so they correctly update the length
13850 after reading an EOC for the EXPLICIT tag.
13854 *This change does not apply to 0.9.7.*
13856 * Fix bug in PKCS#12 key generation routines. This was triggered
13857 if a 3DES key was generated with a 0 initial byte. Include
13858 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13859 (but broken) behaviour.
13863 * Enhance bctest to search for a working bc along $PATH and print
13866 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13868 * Fix memory leaks in err.c: free err_data string if necessary;
13869 don't write to the wrong index in ERR_set_error_data.
13873 * Implement ssl23_peek (analogous to ssl23_read), which previously
13878 * Replace rdtsc with `_emit` statements for VC++ version 5.
13880 *Jeremy Cooper <jeremy@baymoo.org>*
13882 * Make it possible to reuse SSLv2 sessions.
13886 * In copy_email() check for >= 0 as a return value for
13887 X509_NAME_get_index_by_NID() since 0 is a valid index.
13889 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13891 * Avoid coredump with unsupported or invalid public keys by checking if
13892 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13893 PKCS7_verify() fails with non detached data.
13897 * Don't use getenv in library functions when run as setuid/setgid.
13898 New function OPENSSL_issetugid().
13902 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13903 due to incorrect handling of multi-threading:
13905 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13907 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13909 3. Count how many times MemCheck_off() has been called so that
13910 nested use can be treated correctly. This also avoids
13911 inband-signalling in the previous code (which relied on the
13912 assumption that thread ID 0 is impossible).
13916 * Add "-rand" option also to s_client and s_server.
13920 * Fix CPU detection on Irix 6.x.
13921 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13922 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13924 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13929 *This change does not apply to 0.9.7.*
13931 * Use the cached encoding of an X509_NAME structure rather than
13932 copying it. This is apparently the reason for the libsafe "errors"
13933 but the code is actually correct.
13937 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13938 Bleichenbacher's DSA attack.
13939 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13940 to be set and top=0 forces the highest bit to be set; top=-1 is new
13941 and leaves the highest bit random.
13943 *Ulf Moeller, Bodo Moeller*
13945 * In the `NCONF_...`-based implementations for `CONF_...` queries
13946 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13947 a temporary CONF structure with the data component set to NULL
13948 (which gives segmentation faults in lh_retrieve).
13949 Instead, use NULL for the CONF pointer in CONF_get_string and
13950 CONF_get_number (which may use environment variables) and directly
13951 return NULL from CONF_get_section.
13955 * Fix potential buffer overrun for EBCDIC.
13959 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13960 keyUsage if basicConstraints absent for a CA.
13964 * Make SMIME_write_PKCS7() write mail header values with a format that
13965 is more generally accepted (no spaces before the semicolon), since
13966 some programs can't parse those values properly otherwise. Also make
13967 sure BIO's that break lines after each write do not create invalid
13972 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13973 macros previously used would not encode an empty SEQUENCE OF
13974 and break the signature.
13978 *This change does not apply to 0.9.7.*
13980 * Zero the premaster secret after deriving the master secret in
13985 * Add some EVP_add_digest_alias registrations (as found in
13986 OpenSSL_add_all_digests()) to SSL_library_init()
13987 aka OpenSSL_add_ssl_algorithms(). This provides improved
13988 compatibility with peers using X.509 certificates
13989 with unconventional AlgorithmIdentifier OIDs.
13993 * Fix for Irix with NO_ASM.
13995 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13997 * ./config script fixes.
13999 *Ulf Moeller, Richard Levitte*
14001 * Fix 'openssl passwd -1'.
14005 * Change PKCS12_key_gen_asc() so it can cope with non null
14006 terminated strings whose length is passed in the passlen
14007 parameter, for example from PEM callbacks. This was done
14008 by adding an extra length parameter to asc2uni().
14010 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14012 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14013 call failed, free the DSA structure.
14017 * Fix to uni2asc() to cope with zero length Unicode strings.
14018 These are present in some PKCS#12 files.
14022 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14023 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14024 when writing a 32767 byte record.
14026 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14028 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14029 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14031 (RSA objects have a reference count access to which is protected
14032 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14033 so they are meant to be shared between threads.)
14034 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14035 "Reddie, Steven" <Steven.Reddie@ca.com>*
14037 * Fix a deadlock in CRYPTO_mem_leaks().
14041 * Use better test patterns in bntest.
14045 * rand_win.c fix for Borland C.
14049 * BN_rshift bugfix for n == 0.
14053 * Add a 'bctest' script that checks for some known 'bc' bugs
14054 so that 'make test' does not abort just because 'bc' is broken.
14058 * Store verify_result within SSL_SESSION also for client side to
14059 avoid potential security hole. (Re-used sessions on the client side
14060 always resulted in verify_result==X509_V_OK, not using the original
14061 result of the server certificate verification.)
14065 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14066 SSL3_RT_APPLICATION_DATA, return 0.
14067 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14072 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14073 releases, have been re-implemented by renaming the previous
14074 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14075 and ssl3_read_internal, respectively, and adding 'peek' parameters
14076 to them. The new ssl[23]_{read,peek} functions are calls to
14077 ssl[23]_read_internal with the 'peek' flag set appropriately.
14078 A 'peek' parameter has also been added to ssl3_read_bytes, which
14079 does the actual work for ssl3_read_internal.
14083 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14084 the method-specific "init()" handler. Also clean up ex_data after
14085 calling the method-specific "finish()" handler. Previously, this was
14086 happening the other way round.
14090 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14091 The previous value, 12, was not always sufficient for BN_mod_exp().
14095 * Make sure that shared libraries get the internal name engine with
14096 the full version number and not just 0. This should mark the
14097 shared libraries as not backward compatible. Of course, this should
14098 be changed again when we can guarantee backward binary compatibility.
14102 * Fix typo in get_cert_by_subject() in by_dir.c
14104 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14106 * Rework the system to generate shared libraries:
14108 - Make note of the expected extension for the shared libraries and
14109 if there is a need for symbolic links from for example libcrypto.so.0
14110 to libcrypto.so.0.9.7. There is extended info in Configure for
14113 - Make as few rebuilds of the shared libraries as possible.
14115 - Still avoid linking the OpenSSL programs with the shared libraries.
14117 - When installing, install the shared libraries separately from the
14122 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14124 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14125 and not in SSL_clear because the latter is also used by the
14126 accept/connect functions; previously, the settings made by
14127 SSL_set_read_ahead would be lost during the handshake.
14129 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14131 * Correct util/mkdef.pl to be selective about disabled algorithms.
14132 Previously, it would create entries for disabled algorithms no
14137 * Added several new manual pages for SSL_* function.
14141 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14143 * In ssl23_get_client_hello, generate an error message when faced
14144 with an initial SSL 3.0/TLS record that is too small to contain the
14145 first two bytes of the ClientHello message, i.e. client_version.
14146 (Note that this is a pathologic case that probably has never happened
14147 in real life.) The previous approach was to use the version number
14148 from the record header as a substitute; but our protocol choice
14149 should not depend on that one because it is not authenticated
14150 by the Finished messages.
14154 * More robust randomness gathering functions for Windows.
14156 *Jeffrey Altman <jaltman@columbia.edu>*
14158 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14159 not set then we don't setup the error code for issuer check errors
14160 to avoid possibly overwriting other errors which the callback does
14161 handle. If an application does set the flag then we assume it knows
14162 what it is doing and can handle the new informational codes
14167 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14168 a general "ANY" type, as such it should be able to decode anything
14169 including tagged types. However it didn't check the class so it would
14170 wrongly interpret tagged types in the same way as their universal
14171 counterpart and unknown types were just rejected. Changed so that the
14172 tagged and unknown types are handled in the same way as a SEQUENCE:
14173 that is the encoding is stored intact. There is also a new type
14174 "V_ASN1_OTHER" which is used when the class is not universal, in this
14175 case we have no idea what the actual type is so we just lump them all
14180 * On VMS, stdout may very well lead to a file that is written to
14181 in a record-oriented fashion. That means that every write() will
14182 write a separate record, which will be read separately by the
14183 programs trying to read from it. This can be very confusing.
14185 The solution is to put a BIO filter in the way that will buffer
14186 text until a linefeed is reached, and then write everything a
14187 line at a time, so every record written will be an actual line,
14188 not chunks of lines and not (usually doesn't happen, but I've
14189 seen it once) several lines in one record. BIO_f_linebuffer() is
14192 Currently, it's a VMS-only method, because that's where it has
14193 been tested well enough.
14197 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14198 it can return incorrect results.
14199 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14200 but it was in 0.9.6-beta[12].)
14204 * Disable the check for content being present when verifying detached
14205 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14206 include zero length content when signing messages.
14210 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14211 BIO_ctrl (for BIO pairs).
14215 * Add DSO method for VMS.
14219 * Bug fix: Montgomery multiplication could produce results with the
14224 * Add RPM specification openssl.spec and modify it to build three
14225 packages. The default package contains applications, application
14226 documentation and run-time libraries. The devel package contains
14227 include files, static libraries and function documentation. The
14228 doc package contains the contents of the doc directory. The original
14229 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14233 * Add a large number of documentation files for many SSL routines.
14235 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14237 * Add a configuration entry for Sony News 4.
14239 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14241 * Don't set the two most significant bits to one when generating a
14242 random number < q in the DSA library.
14246 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14247 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14248 the underlying transport is blocking) if a handshake took place.
14249 (The default behaviour is needed by applications such as s_client
14250 and s_server that use select() to determine when to use SSL_read;
14251 but for applications that know in advance when to expect data, it
14252 just makes things more complicated.)
14256 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14261 * Add a few more EBCDIC conditionals that make `req` and `x509`
14262 work better on such systems.
14264 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14266 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14267 Update PKCS12_parse() so it copies the friendlyName and the
14268 keyid to the certificates aux info.
14272 * Fix bug in PKCS7_verify() which caused an infinite loop
14273 if there was more than one signature.
14275 *Sven Uszpelkat <su@celocom.de>*
14277 * Major change in util/mkdef.pl to include extra information
14278 about each symbol, as well as presenting variables as well
14279 as functions. This change means that there's n more need
14280 to rebuild the .num files when some algorithms are excluded.
14284 * Allow the verify time to be set by an application,
14285 rather than always using the current time.
14289 * Phase 2 verify code reorganisation. The certificate
14290 verify code now looks up an issuer certificate by a
14291 number of criteria: subject name, authority key id
14292 and key usage. It also verifies self signed certificates
14293 by the same criteria. The main comparison function is
14294 X509_check_issued() which performs these checks.
14296 Lot of changes were necessary in order to support this
14297 without completely rewriting the lookup code.
14299 Authority and subject key identifier are now cached.
14301 The LHASH 'certs' is X509_STORE has now been replaced
14302 by a STACK_OF(X509_OBJECT). This is mainly because an
14303 LHASH can't store or retrieve multiple objects with
14304 the same hash value.
14306 As a result various functions (which were all internal
14307 use only) have changed to handle the new X509_STORE
14308 structure. This will break anything that messed round
14309 with X509_STORE internally.
14311 The functions X509_STORE_add_cert() now checks for an
14312 exact match, rather than just subject name.
14314 The X509_STORE API doesn't directly support the retrieval
14315 of multiple certificates matching a given criteria, however
14316 this can be worked round by performing a lookup first
14317 (which will fill the cache with candidate certificates)
14318 and then examining the cache for matches. This is probably
14319 the best we can do without throwing out X509_LOOKUP
14320 entirely (maybe later...).
14322 The X509_VERIFY_CTX structure has been enhanced considerably.
14324 All certificate lookup operations now go via a get_issuer()
14325 callback. Although this currently uses an X509_STORE it
14326 can be replaced by custom lookups. This is a simple way
14327 to bypass the X509_STORE hackery necessary to make this
14328 work and makes it possible to use more efficient techniques
14329 in future. A very simple version which uses a simple
14330 STACK for its trusted certificate store is also provided
14331 using X509_STORE_CTX_trusted_stack().
14333 The verify_cb() and verify() callbacks now have equivalents
14334 in the X509_STORE_CTX structure.
14336 X509_STORE_CTX also has a 'flags' field which can be used
14337 to customise the verify behaviour.
14341 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14342 excludes S/MIME capabilities.
14346 * When a certificate request is read in keep a copy of the
14347 original encoding of the signed data and use it when outputting
14348 again. Signatures then use the original encoding rather than
14349 a decoded, encoded version which may cause problems if the
14350 request is improperly encoded.
14354 * For consistency with other BIO_puts implementations, call
14355 buffer_write(b, ...) directly in buffer_puts instead of calling
14358 In BIO_puts, increment b->num_write as in BIO_write.
14360 *Peter.Sylvester@EdelWeb.fr*
14362 * Fix BN_mul_word for the case where the word is 0. (We have to use
14363 BN_zero, we may not return a BIGNUM with an array consisting of
14364 words set to zero.)
14368 * Avoid calling abort() from within the library when problems are
14369 detected, except if preprocessor symbols have been defined
14370 (such as REF_CHECK, BN_DEBUG etc.).
14374 * New openssl application 'rsautl'. This utility can be
14375 used for low level RSA operations. DER public key
14376 BIO/fp routines also added.
14380 * New Configure entry and patches for compiling on QNX 4.
14382 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14384 * A demo state-machine implementation was sponsored by
14385 Nuron (<http://www.nuron.com/>) and is now available in
14386 demos/state_machine.
14390 * New options added to the 'dgst' utility for signature
14391 generation and verification.
14395 * Unrecognized PKCS#7 content types are now handled via a
14396 catch all ASN1_TYPE structure. This allows unsupported
14397 types to be stored as a "blob" and an application can
14398 encode and decode it manually.
14402 * Fix various signed/unsigned issues to make a_strex.c
14403 compile under VC++.
14405 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14407 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14408 length if passed a buffer. ASN1_INTEGER_to_BN failed
14409 if passed a NULL BN and its argument was negative.
14411 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14413 * Modification to PKCS#7 encoding routines to output definite
14414 length encoding. Since currently the whole structures are in
14415 memory there's not real point in using indefinite length
14416 constructed encoding. However if OpenSSL is compiled with
14417 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14421 * Added BIO_vprintf() and BIO_vsnprintf().
14425 * Added more prefixes to parse for in the strings written
14426 through a logging bio, to cover all the levels that are available
14427 through syslog. The prefixes are now:
14429 PANIC, EMERG, EMR => LOG_EMERG
14430 ALERT, ALR => LOG_ALERT
14431 CRIT, CRI => LOG_CRIT
14432 ERROR, ERR => LOG_ERR
14433 WARNING, WARN, WAR => LOG_WARNING
14434 NOTICE, NOTE, NOT => LOG_NOTICE
14435 INFO, INF => LOG_INFO
14436 DEBUG, DBG => LOG_DEBUG
14438 and as before, if none of those prefixes are present at the
14439 beginning of the string, LOG_ERR is chosen.
14441 On Win32, the `LOG_*` levels are mapped according to this:
14443 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14444 LOG_WARNING => EVENTLOG_WARNING_TYPE
14445 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14449 * Made it possible to reconfigure with just the configuration
14450 argument "reconf" or "reconfigure". The command line arguments
14451 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14452 and are retrieved from there when reconfiguring.
14458 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14460 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14464 * The obj_dat.pl script was messing up the sorting of object
14465 names. The reason was that it compared the quoted version
14466 of strings as a result "OCSP" > "OCSP Signing" because
14467 " > SPACE. Changed script to store unquoted versions of
14468 names and add quotes on output. It was also omitting some
14469 names from the lookup table if they were given a default
14470 value (that is if SN is missing it is given the same
14471 value as LN and vice versa), these are now added on the
14472 grounds that if an object has a name we should be able to
14473 look it up. Finally added warning output when duplicate
14474 short or long names are found.
14478 * Changes needed for Tandem NSK.
14480 *Scott Uroff <scott@xypro.com>*
14482 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14483 RSA_padding_check_SSLv23(), special padding was never detected
14484 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14485 version rollback attacks was not effective.
14487 In s23_clnt.c, don't use special rollback-attack detection padding
14488 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14489 client; similarly, in s23_srvr.c, don't do the rollback check if
14490 SSL 2.0 is the only protocol enabled in the server.
14494 * Make it possible to get hexdumps of unprintable data with 'openssl
14495 asn1parse'. By implication, the functions ASN1_parse_dump() and
14496 BIO_dump_indent() are added.
14500 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14501 these print out strings and name structures based on various
14502 flags including RFC2253 support and proper handling of
14503 multibyte characters. Added options to the 'x509' utility
14504 to allow the various flags to be set.
14508 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14509 Also change the functions X509_cmp_current_time() and
14510 X509_gmtime_adj() work with an ASN1_TIME structure,
14511 this will enable certificates using GeneralizedTime in validity
14512 dates to be checked.
14516 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14517 negative public key encodings) on by default,
14518 NO_NEG_PUBKEY_BUG can be set to disable it.
14522 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14523 content octets. An i2c_ASN1_OBJECT is unnecessary because
14524 the encoding can be trivially obtained from the structure.
14528 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14529 not read locks (`CRYPTO_r_[un]lock`).
14533 * A first attempt at creating official support for shared
14534 libraries through configuration. I've kept it so the
14535 default is static libraries only, and the OpenSSL programs
14536 are always statically linked for now, but there are
14537 preparations for dynamic linking in place.
14538 This has been tested on Linux and Tru64.
14542 * Randomness polling function for Win9x, as described in:
14543 Peter Gutmann, Software Generation of Practically Strong
14548 * Fix so PRNG is seeded in req if using an already existing
14553 * New options to smime application. -inform and -outform
14554 allow alternative formats for the S/MIME message including
14555 PEM and DER. The -content option allows the content to be
14556 specified separately. This should allow things like Netscape
14557 form signing output easier to verify.
14561 * Fix the ASN1 encoding of tags using the 'long form'.
14565 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14566 STRING types. These convert content octets to and from the
14567 underlying type. The actual tag and length octets are
14568 already assumed to have been read in and checked. These
14569 are needed because all other string types have virtually
14570 identical handling apart from the tag. By having versions
14571 of the ASN1 functions that just operate on content octets
14572 IMPLICIT tagging can be handled properly. It also allows
14573 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14574 and ASN1_INTEGER are identical apart from the tag.
14578 * Change the handling of OID objects as follows:
14580 - New object identifiers are inserted in objects.txt, following
14581 the syntax given in objects.README.
14582 - objects.pl is used to process obj_mac.num and create a new
14584 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14587 This is currently kind of a hack, and the perl code in objects.pl
14588 isn't very elegant, but it works as I intended. The simplest way
14589 to check that it worked correctly is to look in obj_dat.h and
14590 check the array nid_objs and make sure the objects haven't moved
14591 around (this is important!). Additions are OK, as well as
14592 consistent name changes.
14596 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14600 * Addition of the command line parameter '-rand file' to 'openssl req'.
14601 The given file adds to whatever has already been seeded into the
14602 random pool through the RANDFILE configuration file option or
14603 environment variable, or the default random state file.
14607 * mkstack.pl now sorts each macro group into lexical order.
14608 Previously the output order depended on the order the files
14609 appeared in the directory, resulting in needless rewriting
14614 * Patches to make OpenSSL compile under Win32 again. Mostly
14615 work arounds for the VC++ problem that it treats func() as
14616 func(void). Also stripped out the parts of mkdef.pl that
14617 added extra typesafe functions: these no longer exist.
14621 * Reorganisation of the stack code. The macros are now all
14622 collected in safestack.h . Each macro is defined in terms of
14623 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14624 DEBUG_SAFESTACK is now handled in terms of function casts,
14625 this has the advantage of retaining type safety without the
14626 use of additional functions. If DEBUG_SAFESTACK is not defined
14627 then the non typesafe macros are used instead. Also modified the
14628 mkstack.pl script to handle the new form. Needs testing to see
14629 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14630 the default if no major problems. Similar behaviour for ASN1_SET_OF
14631 and PKCS12_STACK_OF.
14635 * When some versions of IIS use the 'NET' form of private key the
14636 key derivation algorithm is different. Normally MD5(password) is
14637 used as a 128 bit RC4 key. In the modified case
14638 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14639 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14640 as the old Netscape_RSA functions except they have an additional
14641 'sgckey' parameter which uses the modified algorithm. Also added
14642 an -sgckey command line option to the rsa utility. Thanks to
14643 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14644 algorithm to openssl-dev.
14648 * The evp_local.h macros were using 'c.##kname' which resulted in
14649 invalid expansion on some systems (SCO 5.0.5 for example).
14650 Corrected to 'c.kname'.
14652 *Phillip Porch <root@theporch.com>*
14654 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14655 a STACK of email addresses from a certificate or request, these look
14656 in the subject name and the subject alternative name extensions and
14657 omit any duplicate addresses.
14661 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14662 This makes DSA verification about 2 % faster.
14666 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14667 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14668 plus overhead for 1024 bit moduli).
14669 This makes exponentiations about 0.5 % faster for 1024 bit
14670 exponents (as measured by "openssl speed rsa2048").
14674 * Rename memory handling macros to avoid conflicts with other
14676 Malloc => OPENSSL_malloc
14677 Malloc_locked => OPENSSL_malloc_locked
14678 Realloc => OPENSSL_realloc
14679 Free => OPENSSL_free
14683 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14684 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14688 * CygWin32 support.
14690 *John Jarvie <jjarvie@newsguy.com>*
14692 * The type-safe stack code has been rejigged. It is now only compiled
14693 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14694 by default all type-specific stack functions are "#define"d back to
14695 standard stack functions. This results in more streamlined output
14696 but retains the type-safety checking possibilities of the original
14701 * The STACK code has been cleaned up, and certain type declarations
14702 that didn't make a lot of sense have been brought in line. This has
14703 also involved a cleanup of sorts in safestack.h to more correctly
14704 map type-safe stack functions onto their plain stack counterparts.
14705 This work has also resulted in a variety of "const"ifications of
14706 lots of the code, especially `_cmp` operations which should normally
14707 be prototyped with "const" parameters anyway.
14711 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14712 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14713 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14714 where all of 'md' is used each time the PRNG is used, but 'state'
14715 is used only indexed by a cyclic counter. As entropy may not be
14716 well distributed from the beginning, 'md' is important as a
14717 chaining variable. However, the output function chains only half
14718 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14719 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14720 in all of 'state' being rewritten, with the new values depending
14721 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14725 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14726 the handshake is continued after ssl_verify_cert_chain();
14727 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14728 can lead to 'unexplainable' connection aborts later.
14730 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14732 * Major EVP API cipher revision.
14733 Add hooks for extra EVP features. This allows various cipher
14734 parameters to be set in the EVP interface. Support added for variable
14735 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14736 setting of RC2 and RC5 parameters.
14738 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14741 Remove lots of duplicated code from the EVP library. For example *every*
14742 cipher init() function handles the 'iv' in the same way according to the
14743 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14744 for CFB and OFB modes they zero ctx->num.
14746 New functionality allows removal of S/MIME code RC2 hack.
14748 Most of the routines have the same form and so can be declared in terms
14751 By shifting this to the top level EVP_CipherInit() it can be removed from
14752 all individual ciphers. If the cipher wants to handle IVs or keys
14753 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14756 Change lots of functions like EVP_EncryptUpdate() to now return a
14757 value: although software versions of the algorithms cannot fail
14758 any installed hardware versions can.
14762 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14763 this option is set, tolerate broken clients that send the negotiated
14764 protocol version number instead of the requested protocol version
14769 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14770 i.e. non-zero for export ciphersuites, zero otherwise.
14771 Previous versions had this flag inverted, inconsistent with
14772 rsa_tmp_cb (..._TMP_RSA_CB).
14774 *Bodo Moeller; problem reported by Amit Chopra*
14776 * Add missing DSA library text string. Work around for some IIS
14777 key files with invalid SEQUENCE encoding.
14781 * Add a document (doc/standards.txt) that list all kinds of standards
14782 and so on that are implemented in OpenSSL.
14786 * Enhance c_rehash script. Old version would mishandle certificates
14787 with the same subject name hash and wouldn't handle CRLs at all.
14788 Added -fingerprint option to crl utility, to support new c_rehash
14793 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14797 * Fix for SSL server purpose checking. Server checking was
14798 rejecting certificates which had extended key usage present
14799 but no ssl client purpose.
14801 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14803 * Make PKCS#12 code work with no password. The PKCS#12 spec
14804 is a little unclear about how a blank password is handled.
14805 Since the password in encoded as a BMPString with terminating
14806 double NULL a zero length password would end up as just the
14807 double NULL. However no password at all is different and is
14808 handled differently in the PKCS#12 key generation code. NS
14809 treats a blank password as zero length. MSIE treats it as no
14810 password on export: but it will try both on import. We now do
14811 the same: PKCS12_parse() tries zero length and no password if
14812 the password is set to "" or NULL (NULL is now a valid password:
14813 it wasn't before) as does the pkcs12 application.
14817 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14818 perror when PEM_read_bio_X509_REQ fails, the error message must
14819 be obtained from the error queue.
14823 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14824 it in ERR_remove_state if appropriate, and change ERR_get_state
14825 accordingly to avoid race conditions (this is necessary because
14826 thread_hash is no longer constant once set).
14830 * Bugfix for linux-elf makefile.one.
14834 * RSA_get_default_method() will now cause a default
14835 RSA_METHOD to be chosen if one doesn't exist already.
14836 Previously this was only set during a call to RSA_new()
14837 or RSA_new_method(NULL) meaning it was possible for
14838 RSA_get_default_method() to return NULL.
14842 * Added native name translation to the existing DSO code
14843 that will convert (if the flag to do so is set) filenames
14844 that are sufficiently small and have no path information
14845 into a canonical native form. Eg. "blah" converted to
14846 "libblah.so" or "blah.dll" etc.
14850 * New function ERR_error_string_n(e, buf, len) which is like
14851 ERR_error_string(e, buf), but writes at most 'len' bytes
14852 including the 0 terminator. For ERR_error_string_n, 'buf'
14855 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14857 * CONF library reworked to become more general. A new CONF
14858 configuration file reader "class" is implemented as well as a
14859 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14860 old `CONF_*` functions are still there, but are reimplemented to
14861 work in terms of the new functions. Also, a set of functions
14862 to handle the internal storage of the configuration data is
14863 provided to make it easier to write new configuration file
14864 reader "classes" (I can definitely see something reading a
14865 configuration file in XML format, for example), called `_CONF_*`,
14866 or "the configuration storage API"...
14868 The new configuration file reading functions are:
14870 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14871 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14873 NCONF_default, NCONF_WIN32
14875 NCONF_dump_fp, NCONF_dump_bio
14877 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14878 NCONF_new creates a new CONF object. This works in the same way
14879 as other interfaces in OpenSSL, like the BIO interface.
14880 `NCONF_dump_*` dump the internal storage of the configuration file,
14881 which is useful for debugging. All other functions take the same
14882 arguments as the old `CONF_*` functions with the exception of the
14883 first that must be a `CONF *` instead of a `LHASH *`.
14885 To make it easier to use the new classes with the old `CONF_*` functions,
14886 the function CONF_set_default_method is provided.
14890 * Add '-tls1' option to 'openssl ciphers', which was already
14891 mentioned in the documentation but had not been implemented.
14892 (This option is not yet really useful because even the additional
14893 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14897 * Initial DSO code added into libcrypto for letting OpenSSL (and
14898 OpenSSL-based applications) load shared libraries and bind to
14899 them in a portable way.
14901 *Geoff Thorpe, with contributions from Richard Levitte*
14903 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
14905 * Make sure _lrotl and _lrotr are only used with MSVC.
14907 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14908 (the default implementation of RAND_status).
14910 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14911 to '-clrext' (= clear extensions), as intended and documented.
14912 *Bodo Moeller; inconsistency pointed out by Michael Attili
14913 <attili@amaxo.com>*
14915 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14916 was larger than the MD block size.
14918 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14920 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14921 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14922 using the passed key: if the passed key was a private key the result
14923 of X509_print(), for example, would be to print out all the private key
14928 * des_quad_cksum() byte order bug fix.
14929 *Ulf Möller, using the problem description in krb4-0.9.7, where
14930 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14932 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14935 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14937 * For easily testing in shell scripts whether some command
14938 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14939 returns with exit code 0 iff no command of the given name is available.
14940 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14941 the output goes to stdout and nothing is printed to stderr.
14942 Additional arguments are always ignored.
14944 Since for each cipher there is a command of the same name,
14945 the 'no-cipher' compilation switches can be tested this way.
14947 ('openssl no-XXX' is not able to detect pseudo-commands such
14948 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14952 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14956 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14957 is set; it will be thrown away anyway because each handshake creates
14959 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14960 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14961 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14962 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14966 * New s_client option -ign_eof: EOF at stdin is ignored, and
14967 'Q' and 'R' lose their special meanings (quit/renegotiate).
14968 This is part of what -quiet does; unlike -quiet, -ign_eof
14969 does not suppress any output.
14973 * Add compatibility options to the purpose and trust code. The
14974 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14975 accepts a certificate or CA, this was the previous behaviour,
14976 with all the associated security issues.
14978 X509_TRUST_COMPAT is the old trust behaviour: only and
14979 automatically trust self signed roots in certificate store. A
14980 new trust setting X509_TRUST_DEFAULT is used to specify that
14981 a purpose has no associated trust setting and it should instead
14982 use the value in the default purpose.
14986 * Fix the PKCS#8 DSA private key code so it decodes keys again
14987 and fix a memory leak.
14991 * In util/mkerr.pl (which implements 'make errors'), preserve
14992 reason strings from the previous version of the .c file, as
14993 the default to have only downcase letters (and digits) in
14994 automatically generated reasons codes is not always appropriate.
14998 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
14999 using strerror. Previously, ERR_reason_error_string() returned
15000 library names as reason strings for SYSerr; but SYSerr is a special
15001 case where small numbers are errno values, not library numbers.
15005 * Add '-dsaparam' option to 'openssl dhparam' application. This
15006 converts DSA parameters into DH parameters. (When creating parameters,
15007 DSA_generate_parameters is used.)
15011 * Include 'length' (recommended exponent length) in C code generated
15012 by 'openssl dhparam -C'.
15016 * The second argument to set_label in perlasm was already being used
15017 so couldn't be used as a "file scope" flag. Moved to third argument
15022 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15023 instead of RAND_bytes for encryption IVs and salts.
15027 * Include RAND_status() into RAND_METHOD instead of implementing
15028 it only for md_rand.c Otherwise replacing the PRNG by calling
15029 RAND_set_rand_method would be impossible.
15033 * Don't let DSA_generate_key() enter an infinite loop if the random
15034 number generation fails.
15038 * New 'rand' application for creating pseudo-random output.
15042 * Added configuration support for Linux/IA64
15044 *Rolf Haberrecker <rolf@suse.de>*
15046 * Assembler module support for Mingw32.
15050 * Shared library support for HPUX (in shlib/).
15052 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15054 * Shared library support for Solaris gcc.
15056 *Lutz Behnke <behnke@trustcenter.de>*
15058 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15060 * PKCS7_encrypt() was adding text MIME headers twice because they
15061 were added manually and by SMIME_crlf_copy().
15065 * In bntest.c don't call BN_rand with zero bits argument.
15067 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15069 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15070 case was implemented. This caused BN_div_recp() to fail occasionally.
15074 * Add an optional second argument to the set_label() in the perl
15075 assembly language builder. If this argument exists and is set
15076 to 1 it signals that the assembler should use a symbol whose
15077 scope is the entire file, not just the current function. This
15078 is needed with MASM which uses the format label:: for this scope.
15080 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15082 * Change the ASN1 types so they are typedefs by default. Before
15083 almost all types were #define'd to ASN1_STRING which was causing
15084 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15089 * Change names of new functions to the new get1/get0 naming
15090 convention: After 'get1', the caller owns a reference count
15091 and has to call `..._free`; 'get0' returns a pointer to some
15092 data structure without incrementing reference counters.
15093 (Some of the existing 'get' functions increment a reference
15094 counter, some don't.)
15095 Similarly, 'set1' and 'add1' functions increase reference
15096 counters or duplicate objects.
15100 * Allow for the possibility of temp RSA key generation failure:
15101 the code used to assume it always worked and crashed on failure.
15105 * Fix potential buffer overrun problem in BIO_printf().
15106 *Ulf Möller, using public domain code by Patrick Powell; problem
15107 pointed out by David Sacerdote <das33@cornell.edu>*
15109 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15110 RAND_egd() and RAND_status(). In the command line application,
15111 the EGD socket can be specified like a seed file using RANDFILE
15116 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15117 Some CAs (e.g. Verisign) distribute certificates in this form.
15121 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15122 list to exclude them. This means that no special compilation option
15123 is needed to use anonymous DH: it just needs to be included in the
15128 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15129 EVP_MD_type. The old functionality is available in a new macro called
15130 EVP_MD_md(). Change code that uses it and update docs.
15134 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15135 where the `void *` argument is replaced by a function pointer argument.
15136 Previously `void *` was abused to point to functions, which works on
15137 many platforms, but is not correct. As these functions are usually
15138 called by macros defined in OpenSSL header files, most source code
15139 should work without changes.
15143 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15144 sections with information on -D... compiler switches used for
15145 compiling the library so that applications can see them. To enable
15146 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15147 must be defined. E.g.,
15148 #define OPENSSL_ALGORITHM_DEFINES
15149 #include <openssl/opensslconf.h>
15150 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15152 *Richard Levitte, Ulf and Bodo Möller*
15154 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15159 * Change the 'other' type in certificate aux info to a STACK_OF
15160 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15161 the required ASN1 format: arbitrary types determined by an OID.
15165 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15166 argument to 'req'. This is not because the function is newer or
15167 better than others it just uses the work 'NEW' in the certificate
15168 request header lines. Some software needs this.
15172 * Reorganise password command line arguments: now passwords can be
15173 obtained from various sources. Delete the PEM_cb function and make
15174 it the default behaviour: i.e. if the callback is NULL and the
15175 usrdata argument is not NULL interpret it as a null terminated pass
15176 phrase. If usrdata and the callback are NULL then the pass phrase
15177 is prompted for as usual.
15181 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15182 the support is automatically enabled. The resulting binaries will
15183 autodetect the card and use it if present.
15185 *Ben Laurie and Compaq Inc.*
15187 * Work around for Netscape hang bug. This sends certificate request
15188 and server done in one record. Since this is perfectly legal in the
15189 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15190 the bugs/SSLv3 entry for more info.
15194 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15198 * Add -rand argument to smime and pkcs12 applications and read/write
15203 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15207 * Add command line password options to the remaining applications.
15211 * Bug fix for BN_div_recp() for numerators with an even number of
15216 * More tests in bntest.c, and changed test_bn output.
15220 * ./config recognizes MacOS X now.
15224 * Bug fix for BN_div() when the first words of num and divisor are
15225 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15229 * Add support for various broken PKCS#8 formats, and command line
15230 options to produce them.
15234 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15235 get temporary BIGNUMs from a BN_CTX.
15239 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15244 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15245 include a #define from the old name to the new. The original intent
15246 was that statically linked binaries could for example just call
15247 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15248 link with digests. This never worked because SSLeay_add_all_digests()
15249 and SSLeay_add_all_ciphers() were in the same source file so calling
15250 one would link with the other. They are now in separate source files.
15254 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15258 * Use a less unusual form of the Miller-Rabin primality test (it used
15259 a binary algorithm for exponentiation integrated into the Miller-Rabin
15260 loop, our standard modexp algorithms are faster).
15264 * Support for the EBCDIC character set completed.
15266 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15268 * Source code cleanups: use const where appropriate, eliminate casts,
15269 use `void *` instead of `char *` in lhash.
15273 * Bugfix: ssl3_send_server_key_exchange was not restartable
15274 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15275 this the server could overwrite ephemeral keys that the client
15280 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15281 using 50 iterations of the Rabin-Miller test.
15283 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15284 iterations of the Rabin-Miller test as required by the appendix
15285 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15286 As BN_is_prime_fasttest includes trial division, DSA parameter
15287 generation becomes much faster.
15289 This implies a change for the callback functions in DSA_is_prime
15290 and DSA_generate_parameters: The callback function is called once
15291 for each positive witness in the Rabin-Miller test, not just
15292 occasionally in the inner loop; and the parameters to the
15293 callback function now provide an iteration count for the outer
15294 loop rather than for the current invocation of the inner loop.
15295 DSA_generate_parameters additionally can call the callback
15296 function with an 'iteration count' of -1, meaning that a
15297 candidate has passed the trial division test (when q is generated
15298 from an application-provided seed, trial division is skipped).
15302 * New function BN_is_prime_fasttest that optionally does trial
15303 division before starting the Rabin-Miller test and has
15304 an additional BN_CTX * argument (whereas BN_is_prime always
15305 has to allocate at least one BN_CTX).
15306 'callback(1, -1, cb_arg)' is called when a number has passed the
15307 trial division stage.
15311 * Fix for bug in CRL encoding. The validity dates weren't being handled
15316 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15320 * New function BN_pseudo_rand().
15324 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15325 bignum version of BN_from_montgomery() with the working code from
15326 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15331 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15332 made it impossible to use the same SSL_SESSION data structure in
15333 SSL2 clients in multiple threads.
15337 * The return value of RAND_load_file() no longer counts bytes obtained
15338 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15339 to seed the PRNG (previously an explicit byte count was required).
15341 *Ulf Möller, Bodo Möller*
15343 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15344 used `char *` instead of `void *` and had casts all over the place.
15348 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15352 * Retain source code compatibility for BN_prime_checks macro:
15353 BN_is_prime(..., BN_prime_checks, ...) now uses
15354 BN_prime_checks_for_size to determine the appropriate number of
15355 Rabin-Miller iterations.
15359 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15360 DH_CHECK_P_NOT_SAFE_PRIME.
15361 (Check if this is true? OpenPGP calls them "strong".)
15365 * Merge the functionality of "dh" and "gendh" programs into a new program
15366 "dhparam". The old programs are retained for now but will handle DH keys
15367 (instead of parameters) in future.
15371 * Make the ciphers, s_server and s_client programs check the return values
15372 when a new cipher list is set.
15376 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15377 ciphers. Before when the 56bit ciphers were enabled the sorting was
15380 The syntax for the cipher sorting has been extended to support sorting by
15381 cipher-strength (using the strength_bits hard coded in the tables).
15382 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15384 Fix a bug in the cipher-command parser: when supplying a cipher command
15385 string with an "undefined" symbol (neither command nor alphanumeric
15386 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15387 an error is flagged.
15389 Due to the strength-sorting extension, the code of the
15390 ssl_create_cipher_list() function was completely rearranged. I hope that
15391 the readability was also increased :-)
15393 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15395 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15396 for the first serial number and places 2 in the serial number file. This
15397 avoids problems when the root CA is created with serial number zero and
15398 the first user certificate has the same issuer name and serial number
15403 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15404 the new code. Add documentation for this stuff.
15408 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15409 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15410 structures and behave in an analogous way to the X509v3 functions:
15411 they shouldn't be called directly but wrapper functions should be used
15414 So we also now have some wrapper functions that call the X509at functions
15415 when passed certificate requests. (TO DO: similar things can be done with
15416 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15417 things. Some of these need some d2i or i2d and print functionality
15418 because they handle more complex structures.)
15422 * Add missing #ifndefs that caused missing symbols when building libssl
15423 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15424 NO_RSA in `ssl/s2*.c`.
15426 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15428 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15429 has a return value which indicates the quality of the random data
15430 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15431 error queue. New function RAND_pseudo_bytes() generates output that is
15432 guaranteed to be unique but not unpredictable. RAND_add is like
15433 RAND_seed, but takes an extra argument for an entropy estimate
15434 (RAND_seed always assumes full entropy).
15438 * Do more iterations of Rabin-Miller probable prime test (specifically,
15439 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15440 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15441 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15442 false-positive rate of at most 2^-80 for random input.
15446 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15450 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15451 in the 0.9.5 release), this returns the chain
15452 from an X509_CTX structure with a dup of the stack and all
15453 the X509 reference counts upped: so the stack will exist
15454 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15457 Also make SSL_SESSION_print() print out the verify return
15462 * Add manpage for the pkcs12 command. Also change the default
15463 behaviour so MAC iteration counts are used unless the new
15464 -nomaciter option is used. This improves file security and
15465 only older versions of MSIE (4.0 for example) need it.
15469 * Honor the no-xxx Configure options when creating .DEF files.
15473 * Add PKCS#10 attributes to field table: challengePassword,
15474 unstructuredName and unstructuredAddress. These are taken from
15475 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15476 international characters are used.
15478 More changes to X509_ATTRIBUTE code: allow the setting of types
15479 based on strings. Remove the 'loc' parameter when adding
15480 attributes because these will be a SET OF encoding which is sorted
15485 * Initial changes to the 'req' utility to allow request generation
15486 automation. This will allow an application to just generate a template
15487 file containing all the field values and have req construct the
15490 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15491 used all over the place including certificate requests and PKCS#7
15492 structures. They are currently handled manually where necessary with
15493 some primitive wrappers for PKCS#7. The new functions behave in a
15494 manner analogous to the X509 extension functions: they allow
15495 attributes to be looked up by NID and added.
15497 Later something similar to the X509V3 code would be desirable to
15498 automatically handle the encoding, decoding and printing of the
15499 more complex types. The string types like challengePassword can
15500 be handled by the string table functions.
15502 Also modified the multi byte string table handling. Now there is
15503 a 'global mask' which masks out certain types. The table itself
15504 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15505 is useful when for example there is only one permissible type
15506 (as in countryName) and using the mask might result in no valid
15511 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15512 SSL_get_peer_finished to allow applications to obtain the latest
15513 Finished messages sent to the peer or expected from the peer,
15514 respectively. (SSL_get_peer_finished is usually the Finished message
15515 actually received from the peer, otherwise the protocol will be aborted.)
15517 As the Finished message are message digests of the complete handshake
15518 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15519 be used for external authentication procedures when the authentication
15520 provided by SSL/TLS is not desired or is not enough.
15524 * Enhanced support for Alpha Linux is added. Now ./config checks if
15525 the host supports BWX extension and if Compaq C is present on the
15526 $PATH. Just exploiting of the BWX extension results in 20-30%
15527 performance kick for some algorithms, e.g. DES and RC4 to mention
15528 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15533 * Add support for MS "fast SGC". This is arguably a violation of the
15534 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15535 weak crypto and after checking the certificate is SGC a second one
15536 with strong crypto. MS SGC stops the first handshake after receiving
15537 the server certificate message and sends a second client hello. Since
15538 a server will typically do all the time consuming operations before
15539 expecting any further messages from the client (server key exchange
15540 is the most expensive) there is little difference between the two.
15542 To get OpenSSL to support MS SGC we have to permit a second client
15543 hello message after we have sent server done. In addition we have to
15544 reset the MAC if we do get this second client hello.
15548 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15549 if a DER encoded private key is RSA or DSA traditional format. Changed
15550 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15551 format DER encoded private key. Newer code should use PKCS#8 format which
15552 has the key type encoded in the ASN1 structure. Added DER private key
15553 support to pkcs8 application.
15557 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15558 ciphersuites has been selected (as required by the SSL 3/TLS 1
15559 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15560 is set, we interpret this as a request to violate the specification
15561 (the worst that can happen is a handshake failure, and 'correct'
15562 behaviour would result in a handshake failure anyway).
15566 * In SSL_CTX_add_session, take into account that there might be multiple
15567 SSL_SESSION structures with the same session ID (e.g. when two threads
15568 concurrently obtain them from an external cache).
15569 The internal cache can handle only one SSL_SESSION with a given ID,
15570 so if there's a conflict, we now throw out the old one to achieve
15575 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15576 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15577 some routines that use cipher OIDs: some ciphers do not have OIDs
15578 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15583 * Simplify the trust setting structure and code. Now we just have
15584 two sequences of OIDs for trusted and rejected settings. These will
15585 typically have values the same as the extended key usage extension
15586 and any application specific purposes.
15588 The trust checking code now has a default behaviour: it will just
15589 check for an object with the same NID as the passed id. Functions can
15590 be provided to override either the default behaviour or the behaviour
15591 for a given id. SSL client, server and email already have functions
15592 in place for compatibility: they check the NID and also return "trusted"
15593 if the certificate is self signed.
15597 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15598 traditional format into an EVP_PKEY structure.
15602 * Add a password callback function PEM_cb() which either prompts for
15603 a password if usr_data is NULL or otherwise assumes it is a null
15604 terminated password. Allow passwords to be passed on command line
15605 environment or config files in a few more utilities.
15609 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15610 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15611 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15612 Update documentation.
15616 * Support for ASN1 "NULL" type. This could be handled before by using
15617 ASN1_TYPE but there wasn't any function that would try to read a NULL
15618 and produce an error if it couldn't. For compatibility we also have
15619 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15620 don't allocate anything because they don't need to.
15624 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15627 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15629 * Rebuild of the memory allocation routines used by OpenSSL code and
15630 possibly others as well. The purpose is to make an interface that
15631 provide hooks so anyone can build a separate set of allocation and
15632 deallocation routines to be used by OpenSSL, for example memory
15633 pool implementations, or something else, which was previously hard
15634 since Malloc(), Realloc() and Free() were defined as macros having
15635 the values malloc, realloc and free, respectively (except for Win32
15636 compilations). The same is provided for memory debugging code.
15637 OpenSSL already comes with functionality to find memory leaks, but
15638 this gives people a chance to debug other memory problems.
15640 With these changes, a new set of functions and macros have appeared:
15642 CRYPTO_set_mem_debug_functions() [F]
15643 CRYPTO_get_mem_debug_functions() [F]
15644 CRYPTO_dbg_set_options() [F]
15645 CRYPTO_dbg_get_options() [F]
15646 CRYPTO_malloc_debug_init() [M]
15648 The memory debug functions are NULL by default, unless the library
15649 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15650 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15651 gives the standard debugging functions that come with OpenSSL) or
15652 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15653 provided by the library user) must be used. When the standard
15654 debugging functions are used, CRYPTO_dbg_set_options can be used to
15655 request additional information:
15656 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15657 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15659 Also, things like CRYPTO_set_mem_functions will always give the
15660 expected result (the new set of functions is used for allocation
15661 and deallocation) at all times, regardless of platform and compiler
15664 To finish it up, some functions that were never use in any other
15665 way than through macros have a new API and new semantic:
15667 CRYPTO_dbg_malloc()
15668 CRYPTO_dbg_realloc()
15671 All macros of value have retained their old syntax.
15673 *Richard Levitte and Bodo Moeller*
15675 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15676 ordering of SMIMECapabilities wasn't in "strength order" and there
15677 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15682 * Some ASN1 types with illegal zero length encoding (INTEGER,
15683 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15685 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15687 * Merge in my S/MIME library for OpenSSL. This provides a simple
15688 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15689 functionality to handle multipart/signed properly) and a utility
15690 called 'smime' to call all this stuff. This is based on code I
15691 originally wrote for Celo who have kindly allowed it to be
15692 included in OpenSSL.
15696 * Add variants des_set_key_checked and des_set_key_unchecked of
15697 des_set_key (aka des_key_sched). Global variable des_check_key
15698 decides which of these is called by des_set_key; this way
15699 des_check_key behaves as it always did, but applications and
15700 the library itself, which was buggy for des_check_key == 1,
15701 have a cleaner way to pick the version they need.
15705 * New function PKCS12_newpass() which changes the password of a
15710 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15711 dynamic mix. In both cases the ids can be used as an index into the
15712 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15713 functions so they accept a list of the field values and the
15714 application doesn't need to directly manipulate the X509_TRUST
15719 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15724 * Modify the way the V3 extension code looks up extensions. This now
15725 works in a similar way to the object code: we have some "standard"
15726 extensions in a static table which is searched with OBJ_bsearch()
15727 and the application can add dynamic ones if needed. The file
15728 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15729 updated whenever a new extension is added to the core code and kept
15730 in ext_nid order. There is a simple program 'tabtest.c' which checks
15731 this. New extensions are not added too often so this file can readily
15732 be maintained manually.
15734 There are two big advantages in doing things this way. The extensions
15735 can be looked up immediately and no longer need to be "added" using
15736 X509V3_add_standard_extensions(): this function now does nothing.
15737 Side note: I get *lots* of email saying the extension code doesn't
15738 work because people forget to call this function.
15739 Also no dynamic allocation is done unless new extensions are added:
15740 so if we don't add custom extensions there is no need to call
15741 X509V3_EXT_cleanup().
15745 * Modify enc utility's salting as follows: make salting the default. Add a
15746 magic header, so unsalted files fail gracefully instead of just decrypting
15747 to garbage. This is because not salting is a big security hole, so people
15748 should be discouraged from doing it.
15752 * Fixes and enhancements to the 'x509' utility. It allowed a message
15753 digest to be passed on the command line but it only used this
15754 parameter when signing a certificate. Modified so all relevant
15755 operations are affected by the digest parameter including the
15756 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15757 DSA key was used because it didn't fix the digest.
15761 * Initial certificate chain verify code. Currently tests the untrusted
15762 certificates for consistency with the verify purpose (which is set
15763 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15765 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15766 this is because it will reject chains with invalid extensions whereas
15767 every previous version of OpenSSL and SSLeay made no checks at all.
15769 Trust code: checks the root CA for the relevant trust settings. Trust
15770 settings have an initial value consistent with the verify purpose: e.g.
15771 if the verify purpose is for SSL client use it expects the CA to be
15772 trusted for SSL client use. However the default value can be changed to
15773 permit custom trust settings: one example of this would be to only trust
15774 certificates from a specific "secure" set of CAs.
15776 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15777 which should be used for version portability: especially since the
15778 verify structure is likely to change more often now.
15780 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15781 to set them. If not set then assume SSL clients will verify SSL servers
15784 Two new options to the verify program: -untrusted allows a set of
15785 untrusted certificates to be passed in and -purpose which sets the
15786 intended purpose of the certificate. If a purpose is set then the
15787 new chain verify code is used to check extension consistency.
15791 * Support for the authority information access extension.
15795 * Modify RSA and DSA PEM read routines to transparently handle
15796 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15797 public keys in a format compatible with certificate
15798 SubjectPublicKeyInfo structures. Unfortunately there were already
15799 functions called *_PublicKey_* which used various odd formats so
15800 these are retained for compatibility: however the DSA variants were
15801 never in a public release so they have been deleted. Changed dsa/rsa
15802 utilities to handle the new format: note no releases ever handled public
15803 keys so we should be OK.
15805 The primary motivation for this change is to avoid the same fiasco
15806 that dogs private keys: there are several incompatible private key
15807 formats some of which are standard and some OpenSSL specific and
15808 require various evil hacks to allow partial transparent handling and
15809 even then it doesn't work with DER formats. Given the option anything
15810 other than PKCS#8 should be dumped: but the other formats have to
15811 stay in the name of compatibility.
15813 With public keys and the benefit of hindsight one standard format
15814 is used which works with EVP_PKEY, RSA or DSA structures: though
15815 it clearly returns an error if you try to read the wrong kind of key.
15817 Added a -pubkey option to the 'x509' utility to output the public key.
15818 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15819 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15820 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15821 that do the same as the `EVP_PKEY_assign_*()` except they up the
15822 reference count of the added key (they don't "swallow" the
15827 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15828 CRLs would fail if the file contained no certificates or no CRLs:
15829 added a new function to read in both types and return the number
15830 read: this means that if none are read it will be an error. The
15831 DER versions of the certificate and CRL reader would always fail
15832 because it isn't possible to mix certificates and CRLs in DER format
15833 without choking one or the other routine. Changed this to just read
15834 a certificate: this is the best we can do. Also modified the code
15835 in apps/verify.c to take notice of return codes: it was previously
15836 attempting to read in certificates from NULL pointers and ignoring
15837 any errors: this is one reason why the cert and CRL reader seemed
15838 to work. It doesn't check return codes from the default certificate
15839 routines: these may well fail if the certificates aren't installed.
15843 * Code to support otherName option in GeneralName.
15847 * First update to verify code. Change the verify utility
15848 so it warns if it is passed a self signed certificate:
15849 for consistency with the normal behaviour. X509_verify
15850 has been modified to it will now verify a self signed
15851 certificate if *exactly* the same certificate appears
15852 in the store: it was previously impossible to trust a
15853 single self signed certificate. This means that:
15854 openssl verify ss.pem
15855 now gives a warning about a self signed certificate but
15856 openssl verify -CAfile ss.pem ss.pem
15861 * For servers, store verify_result in SSL_SESSION data structure
15862 (and add it to external session representation).
15863 This is needed when client certificate verifications fails,
15864 but an application-provided verification callback (set by
15865 SSL_CTX_set_cert_verify_callback) allows accepting the session
15866 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15867 but returns 1): When the session is reused, we have to set
15868 ssl->verify_result to the appropriate error code to avoid
15871 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15873 * Fix a bug in the new PKCS#7 code: it didn't consider the
15874 case in PKCS7_dataInit() where the signed PKCS7 structure
15875 didn't contain any existing data because it was being created.
15877 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15879 * Add a salt to the key derivation routines in enc.c. This
15880 forms the first 8 bytes of the encrypted file. Also add a
15881 -S option to allow a salt to be input on the command line.
15885 * New function X509_cmp(). Oddly enough there wasn't a function
15886 to compare two certificates. We do this by working out the SHA1
15887 hash and comparing that. X509_cmp() will be needed by the trust
15892 * SSL_get1_session() is like SSL_get_session(), but increments
15893 the reference count in the SSL_SESSION returned.
15895 *Geoff Thorpe <geoff@eu.c2.net>*
15897 * Fix for 'req': it was adding a null to request attributes.
15898 Also change the X509_LOOKUP and X509_INFO code to handle
15899 certificate auxiliary information.
15903 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15908 * Add the possibility to add extra information to the memory leak
15909 detecting output, to form tracebacks, showing from where each
15910 allocation was originated: CRYPTO_push_info("constant string") adds
15911 the string plus current file name and line number to a per-thread
15912 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15913 is like calling CYRPTO_pop_info() until the stack is empty.
15914 Also updated memory leak detection code to be multi-thread-safe.
15918 * Add options -text and -noout to pkcs7 utility and delete the
15919 encryption options which never did anything. Update docs.
15923 * Add options to some of the utilities to allow the pass phrase
15924 to be included on either the command line (not recommended on
15925 OSes like Unix) or read from the environment. Update the
15926 manpages and fix a few bugs.
15930 * Add a few manpages for some of the openssl commands.
15934 * Fix the -revoke option in ca. It was freeing up memory twice,
15935 leaking and not finding already revoked certificates.
15939 * Extensive changes to support certificate auxiliary information.
15940 This involves the use of X509_CERT_AUX structure and X509_AUX
15941 functions. An X509_AUX function such as PEM_read_X509_AUX()
15942 can still read in a certificate file in the usual way but it
15943 will also read in any additional "auxiliary information". By
15944 doing things this way a fair degree of compatibility can be
15945 retained: existing certificates can have this information added
15946 using the new 'x509' options.
15948 Current auxiliary information includes an "alias" and some trust
15949 settings. The trust settings will ultimately be used in enhanced
15950 certificate chain verification routines: currently a certificate
15951 can only be trusted if it is self signed and then it is trusted
15956 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
15957 The problem was that one of the replacement routines had not been working
15958 since SSLeay releases. For now the offending routine has been replaced
15959 with non-optimised assembler. Even so, this now gives around 95%
15960 performance improvement for 1024 bit RSA signs.
15964 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15965 handling. Most clients have the effective key size in bits equal to
15966 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15967 A few however don't do this and instead use the size of the decrypted key
15968 to determine the RC2 key length and the AlgorithmIdentifier to determine
15969 the effective key length. In this case the effective key length can still
15970 be 40 bits but the key length can be 168 bits for example. This is fixed
15971 by manually forcing an RC2 key into the EVP_PKEY structure because the
15972 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15973 the key length and effective key length are equal.
15977 * Add a bunch of functions that should simplify the creation of
15978 X509_NAME structures. Now you should be able to do:
15979 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15980 and have it automatically work out the correct field type and fill in
15981 the structures. The more adventurous can try:
15982 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15983 and it will (hopefully) work out the correct multibyte encoding.
15987 * Change the 'req' utility to use the new field handling and multibyte
15988 copy routines. Before the DN field creation was handled in an ad hoc
15989 way in req, ca, and x509 which was rather broken and didn't support
15990 BMPStrings or UTF8Strings. Since some software doesn't implement
15991 BMPStrings or UTF8Strings yet, they can be enabled using the config file
15992 using the dirstring_type option. See the new comment in the default
15993 openssl.cnf for more info.
15997 * Make crypto/rand/md_rand.c more robust:
15998 - Assure unique random numbers after fork().
15999 - Make sure that concurrent threads access the global counter and
16000 md serializably so that we never lose entropy in them
16001 or use exactly the same state in multiple threads.
16002 Access to the large state is not always serializable because
16003 the additional locking could be a performance killer, and
16004 md should be large enough anyway.
16008 * New file apps/app_rand.c with commonly needed functionality
16009 for handling the random seed file.
16011 Use the random seed file in some applications that previously did not:
16013 dsaparam -genkey (which also ignored its '-rand' option),
16016 x509 (when signing).
16017 Except on systems with /dev/urandom, it is crucial to have a random
16018 seed file at least for key creation, DSA signing, and for DH exchanges;
16019 for RSA signatures we could do without one.
16021 gendh and gendsa (unlike genrsa) used to read only the first byte
16022 of each file listed in the '-rand' option. The function as previously
16023 found in genrsa is now in app_rand.c and is used by all programs
16024 that support '-rand'.
16028 * In RAND_write_file, use mode 0600 for creating files;
16029 don't just chmod when it may be too late.
16033 * Report an error from X509_STORE_load_locations
16034 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16038 * New function ASN1_mbstring_copy() this copies a string in either
16039 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16040 into an ASN1_STRING type. A mask of permissible types is passed
16041 and it chooses the "minimal" type to use or an error if not type
16046 * Add function equivalents to the various macros in asn1.h. The old
16047 macros are retained with an `M_` prefix. Code inside the library can
16048 use the `M_` macros. External code (including the openssl utility)
16049 should *NOT* in order to be "shared library friendly".
16053 * Add various functions that can check a certificate's extensions
16054 to see if it usable for various purposes such as SSL client,
16055 server or S/MIME and CAs of these types. This is currently
16056 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16057 verification. Also added a -purpose flag to x509 utility to
16058 print out all the purposes.
16062 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16067 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16068 for, obtain and decode and extension and obtain its critical flag.
16069 This allows all the necessary extension code to be handled in a
16070 single function call.
16074 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16075 platforms. See crypto/rc4/rc4_enc.c for further details.
16079 * New -noout option to asn1parse. This causes no output to be produced
16080 its main use is when combined with -strparse and -out to extract data
16081 from a file (which may not be in ASN.1 format).
16085 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16086 when producing the local key id.
16088 *Richard Levitte <levitte@stacken.kth.se>*
16090 * New option -dhparam in s_server. This allows a DH parameter file to be
16091 stated explicitly. If it is not stated then it tries the first server
16092 certificate file. The previous behaviour hard coded the filename
16097 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16098 a public key to be input or output. For example:
16099 openssl rsa -in key.pem -pubout -out pubkey.pem
16100 Also added necessary DSA public key functions to handle this.
16104 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16105 in the message. This was handled by allowing
16106 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16108 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16110 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16111 to the end of the strings whereas this didn't. This would cause problems
16112 if strings read with d2i_ASN1_bytes() were later modified.
16114 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16116 * Fix for base64 decode bug. When a base64 bio reads only one line of
16117 data and it contains EOF it will end up returning an error. This is
16118 caused by input 46 bytes long. The cause is due to the way base64
16119 BIOs find the start of base64 encoded data. They do this by trying a
16120 trial decode on each line until they find one that works. When they
16121 do a flag is set and it starts again knowing it can pass all the
16122 data directly through the decoder. Unfortunately it doesn't reset
16123 the context it uses. This means that if EOF is reached an attempt
16124 is made to pass two EOFs through the context and this causes the
16125 resulting error. This can also cause other problems as well. As is
16126 usual with these problems it takes *ages* to find and the fix is
16127 trivial: move one line.
16129 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16131 * Ugly workaround to get s_client and s_server working under Windows. The
16132 old code wouldn't work because it needed to select() on sockets and the
16133 tty (for keypresses and to see if data could be written). Win32 only
16134 supports select() on sockets so we select() with a 1s timeout on the
16135 sockets and then see if any characters are waiting to be read, if none
16136 are present then we retry, we also assume we can always write data to
16137 the tty. This isn't nice because the code then blocks until we've
16138 received a complete line of data and it is effectively polling the
16139 keyboard at 1s intervals: however it's quite a bit better than not
16140 working at all :-) A dedicated Windows application might handle this
16141 with an event loop for example.
16145 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16146 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16147 will be called when RSA_sign() and RSA_verify() are used. This is useful
16148 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16149 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16150 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16151 This necessitated the support of an extra signature type NID_md5_sha1
16152 for SSL signatures and modifications to the SSL library to use it instead
16153 of calling RSA_public_decrypt() and RSA_private_encrypt().
16157 * Add new -verify -CAfile and -CApath options to the crl program, these
16158 will lookup a CRL issuers certificate and verify the signature in a
16159 similar way to the verify program. Tidy up the crl program so it
16160 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16161 less strict. It will now permit CRL extensions even if it is not
16162 a V2 CRL: this will allow it to tolerate some broken CRLs.
16166 * Initialize all non-automatic variables each time one of the openssl
16167 sub-programs is started (this is necessary as they may be started
16168 multiple times from the "OpenSSL>" prompt).
16170 *Lennart Bang, Bodo Moeller*
16172 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16173 removing all other RSA functionality (this is what NO_RSA does). This
16174 is so (for example) those in the US can disable those operations covered
16175 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16180 * Non-copying interface to BIO pairs.
16181 (still largely untested)
16185 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16186 ASCII string. This was handled independently in various places before.
16190 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16191 UTF8 strings a character at a time.
16195 * Use client_version from client hello to select the protocol
16196 (s23_srvr.c) and for RSA client key exchange verification
16197 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16201 * Add various utility functions to handle SPKACs, these were previously
16202 handled by poking round in the structure internals. Added new function
16203 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16204 print, verify and generate SPKACs. Based on an original idea from
16205 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16209 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16213 * Allow the config file extension section to be overwritten on the
16214 command line. Based on an original idea from Massimiliano Pala
16215 <madwolf@comune.modena.it>. The new option is called -extensions
16216 and can be applied to ca, req and x509. Also -reqexts to override
16217 the request extensions in req and -crlexts to override the crl extensions
16222 * Add new feature to the SPKAC handling in ca. Now you can include
16223 the same field multiple times by preceding it by "XXXX." for example:
16226 this is the same syntax as used in the req config file.
16230 * Allow certificate extensions to be added to certificate requests. These
16231 are specified in a 'req_extensions' option of the req section of the
16232 config file. They can be printed out with the -text option to req but
16233 are otherwise ignored at present.
16237 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16238 data read consists of only the final block it would not decrypted because
16239 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16240 A misplaced 'break' also meant the decrypted final block might not be
16241 copied until the next read.
16245 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16246 a few extra parameters to the DH structure: these will be useful if
16247 for example we want the value of 'q' or implement X9.42 DH.
16251 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16252 provides hooks that allow the default DSA functions or functions on a
16253 "per key" basis to be replaced. This allows hardware acceleration and
16254 hardware key storage to be handled without major modification to the
16255 library. Also added low level modexp hooks and CRYPTO_EX structure and
16256 associated functions.
16260 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16261 as "read only": it can't be written to and the buffer it points to will
16262 not be freed. Reading from a read only BIO is much more efficient than
16263 a normal memory BIO. This was added because there are several times when
16264 an area of memory needs to be read from a BIO. The previous method was
16265 to create a memory BIO and write the data to it, this results in two
16266 copies of the data and an O(n^2) reading algorithm. There is a new
16267 function BIO_new_mem_buf() which creates a read only memory BIO from
16268 an area of memory. Also modified the PKCS#7 routines to use read only
16273 * Bugfix: ssl23_get_client_hello did not work properly when called in
16274 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16275 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16276 but a retry condition occurred while trying to read the rest.
16280 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16281 NID_pkcs7_encrypted by default: this was wrong since this should almost
16282 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16283 the encrypted data type: this is a more sensible place to put it and it
16284 allows the PKCS#12 code to be tidied up that duplicated this
16289 * Changed obj_dat.pl script so it takes its input and output files on
16290 the command line. This should avoid shell escape redirection problems
16295 * Initial support for certificate extension requests, these are included
16296 in things like Xenroll certificate requests. Included functions to allow
16297 extensions to be obtained and added.
16301 * -crlf option to s_client and s_server for sending newlines as
16302 CRLF (as required by many protocols).
16306 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16308 * Install libRSAglue.a when OpenSSL is built with RSAref.
16310 *Ralf S. Engelschall*
16312 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16314 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16316 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16321 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16322 DH parameters/keys (q is lost during that conversion, but the resulting
16323 DH parameters contain its length).
16325 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16326 much faster than DH_generate_parameters (which creates parameters
16327 where `p = 2*q + 1`), and also the smaller q makes DH computations
16328 much more efficient (160-bit exponentiation instead of 1024-bit
16329 exponentiation); so this provides a convenient way to support DHE
16330 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16331 utter importance to use
16332 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16334 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16335 when such DH parameters are used, because otherwise small subgroup
16336 attacks may become possible!
16340 * Avoid memory leak in i2d_DHparams.
16344 * Allow the -k option to be used more than once in the enc program:
16345 this allows the same encrypted message to be read by multiple recipients.
16349 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16350 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16351 it will always use the numerical form of the OID, even if it has a short
16356 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16357 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16358 otherwise bn_mod_exp was called. In the case of hardware keys for example
16359 no private key components need be present and it might store extra data
16360 in the RSA structure, which cannot be accessed from bn_mod_exp.
16361 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16362 private key operations.
16366 * Added support for SPARC Linux.
16370 * pem_password_cb function type incompatibly changed from
16371 typedef int pem_password_cb(char *buf, int size, int rwflag);
16373 ....(char *buf, int size, int rwflag, void *userdata);
16374 so that applications can pass data to their callbacks:
16375 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16376 additional void * argument, which is just handed through whenever
16377 the password callback is called.
16379 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16381 New function SSL_CTX_set_default_passwd_cb_userdata.
16383 Compatibility note: As many C implementations push function arguments
16384 onto the stack in reverse order, the new library version is likely to
16385 interoperate with programs that have been compiled with the old
16386 pem_password_cb definition (PEM_whatever takes some data that
16387 happens to be on the stack as its last argument, and the callback
16388 just ignores this garbage); but there is no guarantee whatsoever that
16391 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16392 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16393 problems not only on Windows, but also on some Unix platforms.
16394 To avoid problematic command lines, these definitions are now in an
16395 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16396 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16400 * MIPS III/IV assembler module is reimplemented.
16404 * More DES library cleanups: remove references to srand/rand and
16405 delete an unused file.
16409 * Add support for the free Netwide assembler (NASM) under Win32,
16410 since not many people have MASM (ml) and it can be hard to obtain.
16411 This is currently experimental but it seems to work OK and pass all
16412 the tests. Check out INSTALL.W32 for info.
16416 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16417 without temporary keys kept an extra copy of the server key,
16418 and connections with temporary keys did not free everything in case
16423 * New function RSA_check_key and new openssl rsa option -check
16424 for verifying the consistency of RSA keys.
16426 *Ulf Moeller, Bodo Moeller*
16428 * Various changes to make Win32 compile work:
16429 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16430 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16431 comparison" warnings.
16432 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16436 * Add a debugging option to PKCS#5 v2 key generation function: when
16437 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16438 derived keys are printed to stderr.
16442 * Copy the flags in ASN1_STRING_dup().
16444 *Roman E. Pavlov <pre@mo.msk.ru>*
16446 * The x509 application mishandled signing requests containing DSA
16447 keys when the signing key was also DSA and the parameters didn't match.
16449 It was supposed to omit the parameters when they matched the signing key:
16450 the verifying software was then supposed to automatically use the CA's
16451 parameters if they were absent from the end user certificate.
16453 Omitting parameters is no longer recommended. The test was also
16454 the wrong way round! This was probably due to unusual behaviour in
16455 EVP_cmp_parameters() which returns 1 if the parameters match.
16456 This meant that parameters were omitted when they *didn't* match and
16457 the certificate was useless. Certificates signed with 'ca' didn't have
16460 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16462 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16463 The interface is as follows:
16464 Applications can use
16465 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16466 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16467 "off" is now the default.
16468 The library internally uses
16469 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16470 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16471 to disable memory-checking temporarily.
16473 Some inconsistent states that previously were possible (and were
16474 even the default) are now avoided.
16476 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16477 with each memory chunk allocated; this is occasionally more helpful
16478 than just having a counter.
16480 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16482 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16487 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16488 which largely parallels "options", but is for changing API behaviour,
16489 whereas "options" are about protocol behaviour.
16490 Initial "mode" flags are:
16492 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16493 a single record has been written.
16494 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16495 retries use the same buffer location.
16496 (But all of the contents must be
16501 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16504 * Fix problems with no-hmac etc.
16506 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16508 * New functions RSA_get_default_method(), RSA_set_method() and
16509 RSA_get_method(). These allows replacement of RSA_METHODs without having
16510 to mess around with the internals of an RSA structure.
16514 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16515 Also really enable memory leak checks in openssl.c and in some
16518 *Chad C. Mulligan, Bodo Moeller*
16520 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16521 up the length of negative integers. This has now been simplified to just
16522 store the length when it is first determined and use it later, rather
16523 than trying to keep track of where data is copied and updating it to
16525 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16527 * Add a new function PKCS7_signatureVerify. This allows the verification
16528 of a PKCS#7 signature but with the signing certificate passed to the
16529 function itself. This contrasts with PKCS7_dataVerify which assumes the
16530 certificate is present in the PKCS#7 structure. This isn't always the
16531 case: certificates can be omitted from a PKCS#7 structure and be
16532 distributed by "out of band" means (such as a certificate database).
16536 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16537 function prototypes in pem.h, also change util/mkdef.pl to add the
16538 necessary function names.
16542 * mk1mf.pl (used by Windows builds) did not properly read the
16543 options set by Configure in the top level Makefile, and Configure
16544 was not even able to write more than one option correctly.
16545 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16549 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16550 file to be loaded from a BIO or FILE pointer. The BIO version will
16551 for example allow memory BIOs to contain config info.
16555 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16556 Whoever hopes to achieve shared-library compatibility across versions
16557 must use this, not the compile-time macro.
16558 (Exercise 0.9.4: Which is the minimum library version required by
16560 Note: All this applies only to multi-threaded programs, others don't
16565 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16566 through a BIO pair triggered the default case, i.e.
16567 SSLerr(...,SSL_R_UNKNOWN_STATE).
16571 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16572 can use the SSL library even if none of the specific BIOs is
16577 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16578 for the encoded length.
16580 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16582 * Add initial documentation of the X509V3 functions.
16586 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16587 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16588 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16589 secure PKCS#8 private key format with a high iteration count.
16593 * Fix determination of Perl interpreter: A perl or perl5
16594 *directory* in $PATH was also accepted as the interpreter.
16596 *Ralf S. Engelschall*
16598 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16599 wrong with it but it was very old and did things like calling
16600 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16601 unusual formatting.
16605 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16606 to use the new extension code.
16610 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16611 with macros. This should make it easier to change their form, add extra
16612 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16617 * Add to configuration table a new entry that can specify an alternative
16618 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16619 according to Mark Crispin <MRC@Panda.COM>.
16623 * DES CBC did not update the IV. Weird.
16627 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16628 Changing the behaviour of the former might break existing programs --
16629 where IV updating is needed, des_ncbc_encrypt can be used.
16632 * When bntest is run from "make test" it drives bc to check its
16633 calculations, as well as internally checking them. If an internal check
16634 fails, it needs to cause bc to give a non-zero result or make test carries
16635 on without noticing the failure. Fixed.
16639 * DES library cleanups.
16643 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16644 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16645 ciphers. NOTE: although the key derivation function has been verified
16646 against some published test vectors it has not been extensively tested
16647 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16652 * Instead of "mkdir -p", which is not fully portable, use new
16653 Perl script "util/mkdir-p.pl".
16657 * Rewrite the way password based encryption (PBE) is handled. It used to
16658 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16659 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16660 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16661 the 'parameter' field of the AlgorithmIdentifier is passed to the
16662 underlying key generation function so it must do its own ASN1 parsing.
16663 This has also changed the EVP_PBE_CipherInit() function which now has a
16664 'parameter' argument instead of literal salt and iteration count values
16665 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16669 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16670 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16671 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16672 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16673 value was just used as a "magic string" and not used directly its
16674 value doesn't matter.
16678 * Introduce some semblance of const correctness to BN. Shame C doesn't
16683 * "linux-sparc64" configuration (ultrapenguin).
16685 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16686 "linux-sparc" configuration.
16688 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16690 * config now generates no-xxx options for missing ciphers.
16694 * Support the EBCDIC character set (work in progress).
16695 File ebcdic.c not yet included because it has a different license.
16697 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16699 * Support BS2000/OSD-POSIX.
16701 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16703 * Make callbacks for key generation use `void *` instead of `char *`.
16707 * Make S/MIME samples compile (not yet tested).
16711 * Additional typesafe stacks.
16715 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16719 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16721 * New configuration variant "sco5-gcc".
16723 * Updated some demos.
16725 *Sean O Riordain, Wade Scholine*
16727 * Add missing BIO_free at exit of pkcs12 application.
16731 * Fix memory leak in conf.c.
16735 * Updates for Win32 to assembler version of MD5.
16739 * Set #! path to perl in apps/der_chop to where we found it
16740 instead of using a fixed path.
16744 * SHA library changes for irix64-mips4-cc.
16748 * Improvements for VMS support.
16752 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16754 * Bignum library bug fix. IRIX 6 passes "make test" now!
16755 This also avoids the problems with SC4.2 and unpatched SC5.
16757 *Andy Polyakov <appro@fy.chalmers.se>*
16759 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16760 These are required because of the typesafe stack would otherwise break
16761 existing code. If old code used a structure member which used to be STACK
16762 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16763 sk_num or sk_value it would produce an error because the num, data members
16764 are not present in STACK_OF. Now it just produces a warning. sk_set
16765 replaces the old method of assigning a value to sk_value
16766 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16767 that does this will no longer work (and should use sk_set instead) but
16768 this could be regarded as a "questionable" behaviour anyway.
16772 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16773 correctly handle encrypted S/MIME data.
16777 * Change type of various DES function arguments from des_cblock
16778 (which means, in function argument declarations, pointer to char)
16779 to des_cblock * (meaning pointer to array with 8 char elements),
16780 which allows the compiler to do more typechecking; it was like
16781 that back in SSLeay, but with lots of ugly casts.
16783 Introduce new type const_des_cblock.
16787 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16788 problems: find RecipientInfo structure that matches recipient certificate
16789 and initialise the ASN1 structures properly based on passed cipher.
16793 * Belatedly make the BN tests actually check the results.
16797 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16798 to and from BNs: it was completely broken. New compilation option
16799 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16800 key elements as negative integers.
16804 * Reorganize and speed up MD5.
16806 *Andy Polyakov <appro@fy.chalmers.se>*
16810 *Richard Levitte <richard@levitte.org>*
16812 * New option -out to asn1parse to allow the parsed structure to be
16813 output to a file. This is most useful when combined with the -strparse
16814 option to examine the output of things like OCTET STRINGS.
16818 * Make SSL library a little more fool-proof by not requiring any longer
16819 that `SSL_set_{accept,connect}_state` be called before
16820 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16821 in many applications because usually everything *appeared* to work as
16822 intended anyway -- now it really works as intended).
16826 * Move openssl.cnf out of lib/.
16830 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16831 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16832 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16834 *Ralf S. Engelschall*
16836 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16837 handle PKCS#7 enveloped data properly.
16839 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16841 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16842 copying pointers. The cert_st handling is changed by this in
16843 various ways (and thus what used to be known as ctx->default_cert
16844 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16845 any longer when s->cert does not give us what we need).
16846 ssl_cert_instantiate becomes obsolete by this change.
16847 As soon as we've got the new code right (possibly it already is?),
16848 we have solved a couple of bugs of the earlier code where s->cert
16849 was used as if it could not have been shared with other SSL structures.
16851 Note that using the SSL API in certain dirty ways now will result
16852 in different behaviour than observed with earlier library versions:
16853 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16854 does not influence s as it used to.
16856 In order to clean up things more thoroughly, inside SSL_SESSION
16857 we don't use CERT any longer, but a new structure SESS_CERT
16858 that holds per-session data (if available); currently, this is
16859 the peer's certificate chain and, for clients, the server's certificate
16860 and temporary key. CERT holds only those values that can have
16861 meaningful defaults in an SSL_CTX.
16865 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16866 from the internal representation. Various PKCS#7 fixes: remove some
16867 evil casts and set the enc_dig_alg field properly based on the signing
16872 * Allow PKCS#12 password to be set from the command line or the
16873 environment. Let 'ca' get its config file name from the environment
16874 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16879 * Allow certificate policies extension to use an IA5STRING for the
16880 organization field. This is contrary to the PKIX definition but
16881 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16886 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16887 without disallowing inline assembler and the like for non-pedantic builds.
16891 * Support Borland C++ builder.
16893 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16899 * SHA-1 cleanups and performance enhancements.
16901 *Andy Polyakov <appro@fy.chalmers.se>*
16903 * Sparc v8plus assembler for the bignum library.
16905 *Andy Polyakov <appro@fy.chalmers.se>*
16907 * Accept any -xxx and +xxx compiler options in Configure.
16911 * Update HPUX configuration.
16915 * Add missing `sk_<type>_unshift()` function to safestack.h
16917 *Ralf S. Engelschall*
16919 * New function SSL_CTX_use_certificate_chain_file that sets the
16920 "extra_cert"s in addition to the certificate. (This makes sense
16921 only for "PEM" format files, as chains as a whole are not
16926 * Support verify_depth from the SSL API.
16927 x509_vfy.c had what can be considered an off-by-one-error:
16928 Its depth (which was not part of the external interface)
16929 was actually counting the number of certificates in a chain;
16930 now it really counts the depth.
16934 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16935 instead of X509err, which often resulted in confusing error
16936 messages since the error codes are not globally unique
16937 (e.g. an alleged error in ssl3_accept when a certificate
16938 didn't match the private key).
16940 * New function SSL_CTX_set_session_id_context that allows to set a default
16941 value (so that you don't need SSL_set_session_id_context for each
16942 connection using the SSL_CTX).
16946 * OAEP decoding bug fix.
16950 * Support INSTALL_PREFIX for package builders, as proposed by
16955 * New Configure options "threads" and "no-threads". For systems
16956 where the proper compiler options are known (currently Solaris
16957 and Linux), "threads" is the default.
16961 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16965 * Install various scripts to $(OPENSSLDIR)/misc, not to
16966 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16967 such as /usr/local/bin.
16971 * "make linux-shared" to build shared libraries.
16973 *Niels Poppe <niels@netbox.org>*
16975 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
16979 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16980 extension adding in x509 utility.
16984 * Remove NOPROTO sections and error code comments.
16988 * Partial rewrite of the DEF file generator to now parse the ANSI
16993 * New Configure options --prefix=DIR and --openssldir=DIR.
16997 * Complete rewrite of the error code script(s). It is all now handled
16998 by one script at the top level which handles error code gathering,
16999 header rewriting and C source file generation. It should be much better
17000 than the old method: it now uses a modified version of Ulf's parser to
17001 read the ANSI prototypes in all header files (thus the old K&R definitions
17002 aren't needed for error creation any more) and do a better job of
17003 translating function codes into names. The old 'ASN1 error code embedded
17004 in a comment' is no longer necessary and it doesn't use .err files which
17005 have now been deleted. Also the error code call doesn't have to appear all
17006 on one line (which resulted in some large lines...).
17010 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17014 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17015 0 (which usually indicates a closed connection), but continue reading.
17019 * Fix some race conditions.
17023 * Add support for CRL distribution points extension. Add Certificate
17024 Policies and CRL distribution points documentation.
17028 * Move the autogenerated header file parts to crypto/opensslconf.h.
17032 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17033 8 of keying material. Merlin has also confirmed interop with this fix
17034 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17036 *Merlin Hughes <merlin@baltimore.ie>*
17038 * Fix lots of warnings.
17040 *Richard Levitte <levitte@stacken.kth.se>*
17042 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17043 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17045 *Richard Levitte <levitte@stacken.kth.se>*
17047 * Fix problems with sizeof(long) == 8.
17049 *Andy Polyakov <appro@fy.chalmers.se>*
17051 * Change functions to ANSI C.
17055 * Fix typos in error codes.
17057 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17059 * Remove defunct assembler files from Configure.
17063 * SPARC v8 assembler BIGNUM implementation.
17065 *Andy Polyakov <appro@fy.chalmers.se>*
17067 * Support for Certificate Policies extension: both print and set.
17068 Various additions to support the r2i method this uses.
17072 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17073 return a const string when you are expecting an allocated buffer.
17077 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17078 types DirectoryString and DisplayText.
17082 * Add code to allow r2i extensions to access the configuration database,
17083 add an LHASH database driver and add several ctx helper functions.
17087 * Fix an evil bug in bn_expand2() which caused various BN functions to
17088 fail when they extended the size of a BIGNUM.
17092 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17093 support typesafe stack.
17097 * Fix typo in SSL_[gs]et_options().
17099 *Nils Frostberg <nils@medcom.se>*
17101 * Delete various functions and files that belonged to the (now obsolete)
17102 old X509V3 handling code.
17106 * New Configure option "rsaref".
17110 * Don't auto-generate pem.h.
17114 * Introduce type-safe ASN.1 SETs.
17118 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17120 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17122 * Introduce type-safe STACKs. This will almost certainly break lots of code
17123 that links with OpenSSL (well at least cause lots of warnings), but fear
17124 not: the conversion is trivial, and it eliminates loads of evil casts. A
17125 few STACKed things have been converted already. Feel free to convert more.
17126 In the fullness of time, I'll do away with the STACK type altogether.
17130 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17131 specified in `<certfile>` by updating the entry in the index.txt file.
17132 This way one no longer has to edit the index.txt file manually for
17133 revoking a certificate. The -revoke option does the gory details now.
17135 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17137 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17138 `-text` option at all and this way the `-noout -text` combination was
17139 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17141 *Ralf S. Engelschall*
17143 * Make sure a corresponding plain text error message exists for the
17144 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17145 verify callback function determined that a certificate was revoked.
17147 *Ralf S. Engelschall*
17149 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17150 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17151 all available ciphers including rc5, which was forgotten until now.
17152 In order to let the testing shell script know which algorithms
17153 are available, a new (up to now undocumented) command
17154 `openssl list-cipher-commands` is used.
17158 * Bugfix: s_client occasionally would sleep in select() when
17159 it should have checked SSL_pending() first.
17163 * New functions DSA_do_sign and DSA_do_verify to provide access to
17164 the raw DSA values prior to ASN.1 encoding.
17168 * Tweaks to Configure
17170 *Niels Poppe <niels@netbox.org>*
17172 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17177 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17181 * New config option to avoid instructions that are illegal on the 80386.
17182 The default code is faster, but requires at least a 486.
17186 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17187 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17188 same as SSL2_VERSION anyway.
17192 * New "-showcerts" option for s_client.
17196 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17197 application. Various cleanups and fixes.
17201 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17202 modify error routines to work internally. Add error codes and PBE init
17203 to library startup routines.
17207 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17208 packing functions to asn1 and evp. Changed function names and error
17209 codes along the way.
17213 * PKCS12 integration: and so it begins... First of several patches to
17214 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17215 objects to objects.h
17219 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17220 and display support for Thawte strong extranet extension.
17224 * Add LinuxPPC support.
17226 *Jeff Dubrule <igor@pobox.org>*
17228 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17229 bn_div_words in alpha.s.
17231 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17233 * Make sure the RSA OAEP test is skipped under -DRSAref because
17234 OAEP isn't supported when OpenSSL is built with RSAref.
17236 *Ulf Moeller <ulf@fitug.de>*
17238 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17239 so they no longer are missing under -DNOPROTO.
17241 *Soren S. Jorvang <soren@t.dk>*
17243 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17245 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17246 doesn't work when the session is reused. Coming soon!
17250 * Fix a security hole, that allows sessions to be reused in the wrong
17251 context thus bypassing client cert protection! All software that uses
17252 client certs and session caches in multiple contexts NEEDS PATCHING to
17253 allow session reuse! A fuller solution is in the works.
17255 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17257 * Some more source tree cleanups (removed obsolete files
17258 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17259 permission on "config" script to be executable) and a fix for the INSTALL
17262 *Ulf Moeller <ulf@fitug.de>*
17264 * Remove some legacy and erroneous uses of malloc, free instead of
17267 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17269 * Make rsa_oaep_test return non-zero on error.
17271 *Ulf Moeller <ulf@fitug.de>*
17273 * Add support for native Solaris shared libraries. Configure
17274 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17275 if someone would make that last step automatic.
17277 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17279 * ctx_size was not built with the right compiler during "make links". Fixed.
17283 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17284 except NULL ciphers". This means the default cipher list will no longer
17285 enable NULL ciphers. They need to be specifically enabled e.g. with
17286 the string "DEFAULT:eNULL".
17290 * Fix to RSA private encryption routines: if p < q then it would
17291 occasionally produce an invalid result. This will only happen with
17292 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17296 * Be less restrictive and allow also `perl util/perlpath.pl
17297 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17298 because this way one can also use an interpreter named `perl5' (which is
17299 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17300 installed as `perl').
17302 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17304 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17306 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17308 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17309 advapi32.lib to Win32 build and change the pem test comparison
17310 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17311 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17312 and crypto/des/ede_cbcm_enc.c.
17316 * DES quad checksum was broken on big-endian architectures. Fixed.
17320 * Comment out two functions in bio.h that aren't implemented. Fix up the
17321 Win32 test batch file so it (might) work again. The Win32 test batch file
17322 is horrible: I feel ill....
17326 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17327 in e_os.h. Audit of header files to check ANSI and non ANSI
17328 sections: 10 functions were absent from non ANSI section and not exported
17329 from Windows DLLs. Fixed up libeay.num for new functions.
17333 * Make `openssl version' output lines consistent.
17335 *Ralf S. Engelschall*
17337 * Fix Win32 symbol export lists for BIO functions: Added
17338 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17339 to ms/libeay{16,32}.def.
17341 *Ralf S. Engelschall*
17343 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17344 fine under Unix and passes some trivial tests I've now added. But the
17345 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17346 added to make sure no one expects that this stuff really works in the
17347 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17348 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17351 *Ralf S. Engelschall*
17353 * Fix the generation of two part addresses in perl.
17355 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17357 * Add config entry for Linux on MIPS.
17359 *John Tobey <jtobey@channel1.com>*
17361 * Make links whenever Configure is run, unless we are on Windoze.
17365 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17366 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17371 * Add a useful kludge to allow package maintainers to specify compiler and
17372 other platforms details on the command line without having to patch the
17373 Configure script every time: One now can use
17374 `perl Configure <id>:<details>`,
17375 i.e. platform ids are allowed to have details appended
17376 to them (separated by colons). This is treated as there would be a static
17377 pre-configured entry in Configure's %table under key `<id>` with value
17378 `<details>` and `perl Configure <id>` is called. So, when you want to
17379 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17380 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17381 now, which overrides the FreeBSD-elf entry on-the-fly.
17383 *Ralf S. Engelschall*
17385 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17389 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17390 on the `perl Configure ...' command line. This way one can compile
17391 OpenSSL libraries with Position Independent Code (PIC) which is needed
17392 for linking it into DSOs.
17394 *Ralf S. Engelschall*
17396 * Remarkably, export ciphers were totally broken and no-one had noticed!
17401 * Cleaned up the LICENSE document: The official contact for any license
17402 questions now is the OpenSSL core team under openssl-core@openssl.org.
17403 And add a paragraph about the dual-license situation to make sure people
17404 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17405 to the OpenSSL toolkit.
17407 *Ralf S. Engelschall*
17409 * General source tree makefile cleanups: Made `making xxx in yyy...'
17410 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17411 Additionally cleaned up the `make links' target: Remove unnecessary
17412 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17413 to speed processing and no longer clutter the display with confusing
17414 stuff. Instead only the actually done links are displayed.
17416 *Ralf S. Engelschall*
17418 * Permit null encryption ciphersuites, used for authentication only. It used
17419 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17420 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17425 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17426 signed attributes when verifying signatures (this would break them),
17427 the detached data encoding was wrong and public keys obtained using
17428 X509_get_pubkey() weren't freed.
17432 * Add text documentation for the BUFFER functions. Also added a work around
17433 to a Win95 console bug. This was triggered by the password read stuff: the
17434 last character typed gets carried over to the next fread(). If you were
17435 generating a new cert request using 'req' for example then the last
17436 character of the passphrase would be CR which would then enter the first
17441 * Added the new 'Includes OpenSSL Cryptography Software' button as
17442 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17443 button and can be used by applications based on OpenSSL to show the
17444 relationship to the OpenSSL project.
17446 *Ralf S. Engelschall*
17448 * Remove confusing variables in function signatures in files
17449 ssl/ssl_lib.c and ssl/ssl.h.
17451 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17453 * Don't install bss_file.c under PREFIX/include/
17455 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17457 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17458 functions that return function pointers and has support for NT specific
17459 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17460 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17461 unsigned to signed types: this was killing the Win32 compile.
17465 * Add new certificate file to stack functions,
17466 SSL_add_dir_cert_subjects_to_stack() and
17467 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17468 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17469 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17470 This means that Apache-SSL and similar packages don't have to mess around
17471 to add as many CAs as they want to the preferred list.
17475 * Experiment with doxygen documentation. Currently only partially applied to
17477 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17478 openssl.doxy as the configuration file.
17482 * Get rid of remaining C++-style comments which strict C compilers hate.
17484 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17486 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17487 compiled in by default: it has problems with large keys.
17491 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17492 DH private keys and/or callback functions which directly correspond to
17493 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17494 is needed for applications which have to configure certificates on a
17495 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17497 For the RSA certificate situation is makes no difference, but
17498 for the DSA certificate situation this fixes the "no shared cipher"
17499 problem where the OpenSSL cipher selection procedure failed because the
17500 temporary keys were not overtaken from the context and the API provided
17501 no way to reconfigure them.
17502 The new functions now let applications reconfigure the stuff and they
17503 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17504 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17505 non-public-API function ssl_cert_instantiate() is used as a helper
17506 function and also to reduce code redundancy inside ssl_rsa.c.
17508 *Ralf S. Engelschall*
17510 * Move s_server -dcert and -dkey options out of the undocumented feature
17511 area because they are useful for the DSA situation and should be
17512 recognized by the users.
17514 *Ralf S. Engelschall*
17516 * Fix the cipher decision scheme for export ciphers: the export bits are
17517 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17518 SSL_EXP_MASK. So, the original variable has to be used instead of the
17519 already masked variable.
17521 *Richard Levitte <levitte@stacken.kth.se>*
17523 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17525 *Richard Levitte <levitte@stacken.kth.se>*
17527 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17528 from `int` to `unsigned int` because it is a length and initialized by
17529 EVP_DigestFinal() which expects an `unsigned int *`.
17531 *Richard Levitte <levitte@stacken.kth.se>*
17533 * Don't hard-code path to Perl interpreter on shebang line of Configure
17534 script. Instead use the usual Shell->Perl transition trick.
17536 *Ralf S. Engelschall*
17538 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17539 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17540 -noout -modulus' as it's already the case for `openssl rsa -noout
17541 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17542 currently the public key is printed (a decision which was already done by
17543 `openssl dsa -modulus' in the past) which serves a similar purpose.
17544 Additionally the NO_RSA no longer completely removes the whole -modulus
17545 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17548 *Ralf S. Engelschall*
17550 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17551 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17553 *Arne Ansper <arne@ats.cyber.ee>*
17555 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17556 to be added. Now both 'req' and 'ca' can use new objects defined in the
17561 * Add cool BIO that does syslog (or event log on NT).
17563 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17565 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17566 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17567 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17568 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17572 * Add preliminary config info for new extension code.
17576 * Make RSA_NO_PADDING really use no padding.
17578 *Ulf Moeller <ulf@fitug.de>*
17580 * Generate errors when private/public key check is done.
17584 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17585 for some CRL extensions and new objects added.
17589 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17590 key usage extension and fuller support for authority key id.
17594 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17595 padding method for RSA, which is recommended for new applications in PKCS
17596 #1 v2.0 (RFC 2437, October 1998).
17597 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17598 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17599 against Bleichbacher's attack on RSA.
17600 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17603 * Updates to the new SSL compression code
17605 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17607 * Fix so that the version number in the master secret, when passed
17608 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17609 (because the server will not accept higher), that the version number
17610 is 0x03,0x01, not 0x03,0x00
17612 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17614 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17615 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17616 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17620 * Support for RAW extensions where an arbitrary extension can be
17621 created by including its DER encoding. See apps/openssl.cnf for
17626 * Make sure latest Perl versions don't interpret some generated C array
17627 code as Perl array code in the crypto/err/err_genc.pl script.
17629 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17631 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17632 not many people have the assembler. Various Win32 compilation fixes and
17633 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17634 build instructions.
17638 * Modify configure script 'Configure' to automatically create crypto/date.h
17639 file under Win32 and also build pem.h from pem.org. New script
17640 util/mkfiles.pl to create the MINFO file on environments that can't do a
17641 'make files': perl util/mkfiles.pl >MINFO should work.
17645 * Major rework of DES function declarations, in the pursuit of correctness
17646 and purity. As a result, many evil casts evaporated, and some weirdness,
17647 too. You may find this causes warnings in your code. Zapping your evil
17648 casts will probably fix them. Mostly.
17652 * Fix for a typo in asn1.h. Bug fix to object creation script
17653 obj_dat.pl. It considered a zero in an object definition to mean
17654 "end of object": none of the objects in objects.h have any zeros
17655 so it wasn't spotted.
17657 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17659 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17660 Masking (CBCM). In the absence of test vectors, the best I have been able
17661 to do is check that the decrypt undoes the encrypt, so far. Send me test
17662 vectors if you have them.
17666 * Correct calculation of key length for export ciphers (too much space was
17667 allocated for null ciphers). This has not been tested!
17671 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17672 message is now correct (it understands "crypto" and "ssl" on its
17673 command line). There is also now an "update" option. This will update
17674 the util/ssleay.num and util/libeay.num files with any new functions.
17676 perl util/mkdef.pl crypto ssl update
17677 it will update them.
17681 * Overhauled the Perl interface:
17682 - ported BN stuff to OpenSSL's different BN library
17683 - made the perl/ source tree CVS-aware
17684 - renamed the package from SSLeay to OpenSSL (the files still contain
17685 their history because I've copied them in the repository)
17686 - removed obsolete files (the test scripts will be replaced
17687 by better Test::Harness variants in the future)
17689 *Ralf S. Engelschall*
17691 * First cut for a very conservative source tree cleanup:
17692 1. merge various obsolete readme texts into doc/ssleay.txt
17693 where we collect the old documents and readme texts.
17694 2. remove the first part of files where I'm already sure that we no
17695 longer need them because of three reasons: either they are just temporary
17696 files which were left by Eric or they are preserved original files where
17697 I've verified that the diff is also available in the CVS via "cvs diff
17698 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17699 the crypto/md/ stuff).
17701 *Ralf S. Engelschall*
17703 * More extension code. Incomplete support for subject and issuer alt
17704 name, issuer and authority key id. Change the i2v function parameters
17705 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17706 what that's for :-) Fix to ASN1 macro which messed up
17707 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17711 * Preliminary support for ENUMERATED type. This is largely copied from the
17716 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17718 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17720 * Make sure `make rehash` target really finds the `openssl` program.
17722 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17724 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17725 like to hear about it if this slows down other processors.
17729 * Add CygWin32 platform information to Configure script.
17731 *Alan Batie <batie@aahz.jf.intel.com>*
17733 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17735 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17737 * New program nseq to manipulate netscape certificate sequences
17741 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17746 * Fixes to BN code. Previously the default was to define BN_RECURSION
17747 but the BN code had some problems that would cause failures when
17748 doing certificate verification and some other functions.
17750 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17752 * Add ASN1 and PEM code to support netscape certificate sequences.
17756 * Add ASN1 and PEM code to support netscape certificate sequences.
17760 * Add several PKIX and private extended key usage OIDs.
17764 * Modify the 'ca' program to handle the new extension code. Modify
17765 openssl.cnf for new extension format, add comments.
17769 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17770 and add a sample to openssl.cnf so req -x509 now adds appropriate
17775 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17776 error code, add initial support to X509_print() and x509 application.
17780 * Takes a deep breath and start adding X509 V3 extension support code. Add
17781 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17782 stuff is currently isolated and isn't even compiled yet.
17786 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17787 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17788 Removed the versions check from X509 routines when loading extensions:
17789 this allows certain broken certificates that don't set the version
17790 properly to be processed.
17794 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17795 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17796 can still be regenerated with "make depend".
17800 * Spelling mistake in C version of CAST-128.
17802 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17804 * Changes to the error generation code. The perl script err-code.pl
17805 now reads in the old error codes and retains the old numbers, only
17806 adding new ones if necessary. It also only changes the .err files if new
17807 codes are added. The makefiles have been modified to only insert errors
17808 when needed (to avoid needlessly modifying header files). This is done
17809 by only inserting errors if the .err file is newer than the auto generated
17810 C file. To rebuild all the error codes from scratch (the old behaviour)
17811 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17812 or delete all the .err files.
17816 * CAST-128 was incorrectly implemented for short keys. The C version has
17817 been fixed, but is untested. The assembler versions are also fixed, but
17818 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17819 to regenerate it if needed.
17820 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17821 Hagino <itojun@kame.net>*
17823 * File was opened incorrectly in randfile.c.
17825 *Ulf Möller <ulf@fitug.de>*
17827 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17828 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17829 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17830 al: it's just almost always a UTCTime. Note this patch adds new error
17831 codes so do a "make errors" if there are problems.
17835 * Correct Linux 1 recognition in config.
17837 *Ulf Möller <ulf@fitug.de>*
17839 * Remove pointless MD5 hash when using DSA keys in ca.
17841 *Anonymous <nobody@replay.com>*
17843 * Generate an error if given an empty string as a cert directory. Also
17844 generate an error if handed NULL (previously returned 0 to indicate an
17845 error, but didn't set one).
17847 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17849 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17853 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17854 parameters. This was causing a warning which killed off the Win32 compile.
17858 * Remove C++ style comments from crypto/bn/bn_local.h.
17860 *Neil Costigan <neil.costigan@celocom.com>*
17862 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17863 based on a text string, looking up short and long names and finally
17864 "dot" format. The "dot" format stuff didn't work. Added new function
17865 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17866 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17867 OID is not part of the table.
17871 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17872 X509_LOOKUP_by_alias().
17876 * Sort openssl functions by name.
17880 * Get the gendsa program working (hopefully) and add it to app list. Remove
17881 encryption from sample DSA keys (in case anyone is interested the password
17886 * Make *all* `*_free` functions accept a NULL pointer.
17888 *Frans Heymans <fheymans@isaserver.be>*
17890 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17893 *Anonymous <nobody@replay.com>*
17895 * s_server should send the CAfile as acceptable CAs, not its own cert.
17897 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17899 * Don't blow it for numeric -newkey arguments to apps/req.
17901 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17903 * Temp key "for export" tests were wrong in s3_srvr.c.
17905 *Anonymous <nobody@replay.com>*
17907 * Add prototype for temp key callback functions
17908 SSL_CTX_set_tmp_{rsa,dh}_callback().
17912 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17913 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17917 * X509_name_add_entry() freed the wrong thing after an error.
17919 *Arne Ansper <arne@ats.cyber.ee>*
17921 * rsa_eay.c would attempt to free a NULL context.
17923 *Arne Ansper <arne@ats.cyber.ee>*
17925 * BIO_s_socket() had a broken should_retry() on Windoze.
17927 *Arne Ansper <arne@ats.cyber.ee>*
17929 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17931 *Arne Ansper <arne@ats.cyber.ee>*
17933 * Make sure the already existing X509_STORE->depth variable is initialized
17934 in X509_STORE_new(), but document the fact that this variable is still
17935 unused in the certificate verification process.
17937 *Ralf S. Engelschall*
17939 * Fix the various library and apps files to free up pkeys obtained from
17940 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17944 * Fix reference counting in X509_PUBKEY_get(). This makes
17945 demos/maurice/example2.c work, amongst others, probably.
17947 *Steve Henson and Ben Laurie*
17949 * First cut of a cleanup for apps/. First the `ssleay` program is now named
17950 `openssl` and second, the shortcut symlinks for the `openssl <command>`
17951 are no longer created. This way we have a single and consistent command
17952 line interface `openssl <command>`, similar to `cvs <command>`.
17954 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17956 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17957 BIT STRING wrapper always have zero unused bits.
17961 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17965 * Make the top-level INSTALL documentation easier to understand.
17969 * Makefiles updated to exit if an error occurs in a sub-directory
17970 make (including if user presses ^C) [Paul Sutton]
17972 * Make Montgomery context stuff explicit in RSA data structure.
17976 * Fix build order of pem and err to allow for generated pem.h.
17980 * Fix renumbering bug in X509_NAME_delete_entry().
17984 * Enhanced the err-ins.pl script so it makes the error library number
17985 global and can add a library name. This is needed for external ASN1 and
17986 other error libraries.
17990 * Fixed sk_insert which never worked properly.
17994 * Fix ASN1 macros so they can handle indefinite length constructed
17995 EXPLICIT tags. Some non standard certificates use these: they can now
18000 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18001 into a single doc/ssleay.txt bundle. This way the information is still
18002 preserved but no longer messes up this directory. Now it's new room for
18003 the new set of documentation files.
18005 *Ralf S. Engelschall*
18007 * SETs were incorrectly DER encoded. This was a major pain, because they
18008 shared code with SEQUENCEs, which aren't coded the same. This means that
18009 almost everything to do with SETs or SEQUENCEs has either changed name or
18010 number of arguments.
18012 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18014 * Fix test data to work with the above.
18018 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18019 was already fixed by Eric for 0.9.1 it seems.
18021 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18023 * Autodetect FreeBSD3.
18027 * Fix various bugs in Configure. This affects the following platforms:
18031 unixware-2.0-pentium
18036 * Eliminate generated files from CVS. Reorder tests to regenerate files
18037 before they are needed.
18041 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18045 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18047 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18048 changed SSLeay to OpenSSL in version strings.
18050 *Ralf S. Engelschall*
18052 * Some fixups to the top-level documents.
18056 * Fixed the nasty bug where rsaref.h was not found under compile-time
18057 because the symlink to include/ was missing.
18059 *Ralf S. Engelschall*
18061 * Incorporated the popular no-RSA/DSA-only patches
18062 which allow to compile a RSA-free SSLeay.
18064 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18066 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18067 when "ssleay" is still not found.
18069 *Ralf S. Engelschall*
18071 * Added more platforms to Configure: Cray T3E, HPUX 11,
18073 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18075 * Updated the README file.
18077 *Ralf S. Engelschall*
18079 * Added various .cvsignore files in the CVS repository subdirs
18080 to make a "cvs update" really silent.
18082 *Ralf S. Engelschall*
18084 * Recompiled the error-definition header files and added
18085 missing symbols to the Win32 linker tables.
18087 *Ralf S. Engelschall*
18089 * Cleaned up the top-level documents;
18090 o new files: CHANGES and LICENSE
18091 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18092 o merged COPYRIGHT into LICENSE
18093 o removed obsolete TODO file
18094 o renamed MICROSOFT to INSTALL.W32
18096 *Ralf S. Engelschall*
18098 * Removed dummy files from the 0.9.1b source tree:
18099 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18100 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18101 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18102 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18103 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18105 *Ralf S. Engelschall*
18107 * Added various platform portability fixes.
18111 * The Genesis of the OpenSSL rpject:
18112 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18113 Young and Tim J. Hudson created while they were working for C2Net until
18116 *The OpenSSL Project*
18118 ### Changes between 0.9.0b and 0.9.1b [not released]
18120 * Updated a few CA certificates under certs/
18124 * Changed some BIGNUM api stuff.
18128 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18129 DGUX x86, Linux Alpha, etc.
18133 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18134 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18139 * Add -strparse option to asn1pars program which parses nested
18142 *Dr Stephen Henson <shenson@bigfoot.com>*
18144 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18148 * DSA fix for "ca" program.
18152 * Added "-genkey" option to "dsaparam" program.
18156 * Added RIPE MD160 (rmd160) message digest.
18160 * Added -a (all) option to "ssleay version" command.
18164 * Added PLATFORM define which is the id given to Configure.
18168 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18172 * Extended the ASN.1 parser routines.
18176 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18180 * Added a BN_CTX to the BN library.
18184 * Fixed the weak key values in DES library
18188 * Changed API in EVP library for cipher aliases.
18192 * Added support for RC2/64bit cipher.
18196 * Converted the lhash library to the crypto/mem.c functions.
18200 * Added more recognized ASN.1 object ids.
18204 * Added more RSA padding checks for SSL/TLS.
18208 * Added BIO proxy/filter functionality.
18212 * Added extra_certs to SSL_CTX which can be used
18213 send extra CA certificates to the client in the CA cert chain sending
18214 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18218 * Now Fortezza is denied in the authentication phase because
18219 this is key exchange mechanism is not supported by SSLeay at all.
18223 * Additional PKCS1 checks.
18227 * Support the string "TLSv1" for all TLS v1 ciphers.
18231 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18232 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18236 * Fixed a few memory leaks.
18240 * Fixed various code and comment typos.
18244 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18245 bytes sent in the client random.
18247 *Edward Bishop <ebishop@spyglass.com>*
18251 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18252 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18253 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18254 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18255 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18256 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18257 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18258 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18259 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18260 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18261 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18262 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18263 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18264 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18265 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18266 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18267 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18268 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18269 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18270 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18271 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18272 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18273 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18274 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18275 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18276 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18277 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18278 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18279 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18280 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18281 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18282 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18283 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18284 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18285 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18286 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18287 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18288 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18289 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18290 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18291 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18292 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18293 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18294 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18295 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18296 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18297 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18298 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18299 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18300 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18301 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18302 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18303 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18304 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18305 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18306 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18307 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18308 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18309 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18310 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18311 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18312 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18313 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18314 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18315 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18316 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18317 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18318 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18319 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18320 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18321 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18322 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18323 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18324 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18325 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18326 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18327 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18328 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18329 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18330 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18331 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18332 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18333 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18334 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18335 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18336 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18337 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18338 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18339 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18340 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18341 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18342 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18343 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18344 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18345 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18346 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18347 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18348 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18349 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18350 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18351 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18352 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18353 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18354 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18355 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18356 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18357 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18358 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18359 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18360 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18361 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18362 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18363 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18364 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18365 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18366 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18367 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18368 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18369 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18370 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18371 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18372 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18373 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18374 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18375 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18376 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18377 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18378 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18379 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18380 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18381 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18382 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18383 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18384 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18385 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18386 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18387 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18388 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18389 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18390 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18391 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18392 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18393 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18394 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18395 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18396 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18397 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18398 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18399 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18400 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18401 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18402 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18403 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18404 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18405 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18406 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18407 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18408 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18409 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18410 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655