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_POINT_make_affine() and EC_POINTs_make_affine(). These
27 functions are not widely used and now OpenSSL automatically perform this
28 conversion when needed.
32 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
33 EC_KEY_precompute_mult(). These functions are not widely used and
34 applications should instead switch to named curves which OpenSSL has
35 hardcoded lookup tables for.
39 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
40 should instead use the L<EC_POINT_mul(3)> function.
44 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
45 that are not applicable to the new provider model. Applications should
46 instead use EVP_default_properties_is_fips_enabled() and
47 EVP_default_properties_enable_fips().
51 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
52 is set, an unexpected EOF is ignored, it pretends a close notify was received
53 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
57 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
58 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
59 used and applications should instead use the
60 L<EC_POINT_set_affine_coordinates(3)> and
61 L<EC_POINT_get_affine_coordinates(3)> functions.
65 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
66 arrays to be more easily constructed via a series of utility functions.
67 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
68 the various push functions and finally convert to a passable OSSL_PARAM
69 array using OSSL_PARAM_BLD_to_param().
73 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
74 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
75 internal keys, if they correspond to one of those built in types.
79 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
80 contain a provider side internal key.
84 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
85 They are old functions that we don't use, and that you could disable with
86 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
90 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
91 have been converted to Markdown with the goal to produce documents
92 which not only look pretty when viewed online in the browser, but
93 remain well readable inside a plain text editor.
95 To achieve this goal, a 'minimalistic' Markdown style has been applied
96 which avoids formatting elements that interfere too much with the
97 reading flow in the text file. For example, it
99 * avoids [ATX headings][] and uses [setext headings][] instead
100 (which works for `<h1>` and `<h2>` headings only).
101 * avoids [inline links][] and uses [reference links][] instead.
102 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
104 [ATX headings]: https://github.github.com/gfm/#atx-headings
105 [setext headings]: https://github.github.com/gfm/#setext-headings
106 [inline links]: https://github.github.com/gfm/#inline-link
107 [reference links]: https://github.github.com/gfm/#reference-link
108 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
109 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
111 *Matthias St. Pierre*
113 * The test suite is changed to preserve results of each test recipe.
114 A new directory test-runs/ with subdirectories named like the
115 test recipes are created in the build tree for this purpose.
119 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
120 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
121 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
123 *David von Oheimb, Martin Peylo*
125 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
126 The legacy OCSP-focused and only partly documented API is retained.
127 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
131 * All of the low level RSA functions have been deprecated including:
133 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
134 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
135 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
136 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
137 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
138 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
139 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
140 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
141 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
142 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
143 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
144 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
145 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
146 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
147 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
148 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
149 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
150 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
151 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
152 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
153 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
154 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
155 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
156 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
157 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
158 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
159 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
160 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
162 Use of these low level functions has been informally discouraged for a long
163 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
164 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
165 L<EVP_PKEY_decrypt(3)>.
169 * X509 certificates signed using SHA1 are no longer allowed at security
171 In TLS/SSL the default security level is 1. It can be set either
172 using the cipher string with @SECLEVEL, or calling
173 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
174 a call to SSL_CTX_use_certificate() will fail if the security level is not
176 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
177 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
182 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
183 modified to use PKEY APIs. These commands are now in maintenance mode
184 and no new features will be added to them.
188 * The command line utility rsautl has been deprecated.
189 Instead use the pkeyutl program.
193 * The command line utilities genrsa and rsa have been modified to use PKEY
194 APIs These commands are now in maintenance mode and no new features will
199 * All of the low level DH functions have been deprecated including:
201 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
202 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
203 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
204 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
205 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
206 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
207 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
208 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
209 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
210 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
211 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
212 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
213 DH_meth_get_generate_params and DH_meth_set_generate_params.
215 Use of these low level functions has been informally discouraged for a long
216 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
217 and L<EVP_PKEY_derive(3)>.
221 * All of the low level DSA functions have been deprecated including:
223 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
224 DSA_get_default_method, DSA_set_method, DSA_get_method,
225 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
226 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
227 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
228 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
229 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
230 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
231 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
232 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
233 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
234 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
235 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
238 Use of these low level functions has been informally discouraged for a long
239 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
240 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
244 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
245 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
246 This means that applications don't have to look at the curve NID and
247 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
248 However, they still can, that EVP_PKEY_set_alias_type() call acts as
249 a no-op when the EVP_PKEY is already of the given type.
251 Parameter and key generation is also reworked to make it possible
252 to generate EVP_PKEY_SM2 parameters and keys without having to go
253 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
254 However, code that does the latter will still work as before.
258 * Deprecated low level ECDH and ECDSA functions. These include:
260 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
261 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
264 Use of these low level functions has been informally discouraged for a long
265 time. Instead applications should use the EVP_PKEY_derive(3),
266 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
270 * Deprecated the EC_KEY_METHOD functions. These include:
272 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
273 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
274 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
275 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
276 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
277 EC_KEY_METHOD_get_verify.
279 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
283 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
284 and EVP_PKEY_decrypt() instead.
285 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
286 and EVP_PKEY_encrypt() instead.
290 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
291 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
292 a new formulation to include all the things it can be used for,
293 as well as words of caution.
297 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
298 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
302 * All of the low level HMAC functions have been deprecated including:
304 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
305 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
308 Use of these low level functions has been informally discouraged for a long
309 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
310 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
311 and L<EVP_MAC_final(3)>.
315 * Over two thousand fixes were made to the documentation, including:
316 - Common options (such as -rand/-writerand, TLS version control, etc)
317 were refactored and point to newly-enhanced descriptions in openssl.pod.
318 - Added style conformance for all options (with help from Richard Levitte),
319 documented all reported missing options, added a CI build to check
320 that all options are documented and that no unimplemented options
322 - Documented some internals, such as all use of environment variables.
323 - Addressed all internal broken L<> references.
327 * All of the low level CMAC functions have been deprecated including:
329 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
330 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
332 Use of these low level functions has been informally discouraged for a long
333 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
334 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
335 and L<EVP_MAC_final(3)>.
339 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
340 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
343 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
344 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
345 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
346 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
347 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
348 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
349 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
350 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
351 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
352 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
354 Use of these low level functions has been informally discouraged
355 for a long time. Applications should use the EVP_DigestInit_ex(3),
356 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
360 * Corrected the documentation of the return values from the `EVP_DigestSign*`
361 set of functions. The documentation mentioned negative values for some
362 errors, but this was never the case, so the mention of negative values
365 Code that followed the documentation and thereby check with something
366 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
370 * All of the low level cipher functions have been deprecated including:
372 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
373 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
374 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
375 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
376 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
377 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
378 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
379 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
380 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
381 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
382 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
383 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
384 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
385 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
386 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
387 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
388 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
389 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
390 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
391 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
392 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
393 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
394 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
395 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
396 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
397 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
398 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
399 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
400 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
402 Use of these low level functions has been informally discouraged for
403 a long time. Applications should use the high level EVP APIs, e.g.
404 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
405 equivalently named decrypt functions instead.
407 *Matt Caswell and Paul Dale*
409 * Removed include/openssl/opensslconf.h.in and replaced it with
410 include/openssl/configuration.h.in, which differs in not including
411 <openssl/macros.h>. A short header include/openssl/opensslconf.h
412 was added to include both.
414 This allows internal hacks where one might need to modify the set
415 of configured macros, for example this if deprecated symbols are
416 still supposed to be available internally:
418 #include <openssl/configuration.h>
420 #undef OPENSSL_NO_DEPRECATED
421 #define OPENSSL_SUPPRESS_DEPRECATED
423 #include <openssl/macros.h>
425 This should not be used by applications that use the exported
426 symbols, as that will lead to linking errors.
430 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
431 used in exponentiation with 512-bit moduli. No EC algorithms are
432 affected. Analysis suggests that attacks against 2-prime RSA1024,
433 3-prime RSA1536, and DSA1024 as a result of this defect would be very
434 difficult to perform and are not believed likely. Attacks against DH512
435 are considered just feasible. However, for an attack the target would
436 have to re-use the DH512 private key, which is not recommended anyway.
437 Also applications directly using the low level API BN_mod_exp may be
438 affected if they use BN_FLG_CONSTTIME.
443 * Most memory-debug features have been deprecated, and the functionality
444 replaced with no-ops.
448 * Added documentation for the STACK API. OpenSSL only defines the STACK
449 functions where they are used.
453 * Introduced a new method type and API, OSSL_SERIALIZER, to
454 represent generic serializers. An implementation is expected to
455 be able to serialize an object associated with a given name (such
456 as an algorithm name for an asymmetric key) into forms given by
457 implementation properties.
459 Serializers are primarily used from inside libcrypto, through
460 calls to functions like EVP_PKEY_print_private(),
461 PEM_write_bio_PrivateKey() and similar.
463 Serializers are specified in such a way that they can be made to
464 directly handle the provider side portion of an object, if this
465 provider side part comes from the same provider as the serializer
466 itself, but can also be made to handle objects in parametrized
467 form (as an OSSL_PARAM array of data). This allows a provider to
468 offer generic serializers as a service for any other provider.
472 * Added a .pragma directive to the syntax of configuration files, to
473 allow varying behavior in a supported and predictable manner.
474 Currently added pragma:
478 This allows dollar signs to be a keyword character unless it's
479 followed by a opening brace or parenthesis. This is useful for
480 platforms where dollar signs are commonly used in names, such as
481 volume names and system directory names on VMS.
485 * Added functionality to create an EVP_PKEY from user data. This
486 is effectively the same as creating a RSA, DH or DSA object and
487 then assigning them to an EVP_PKEY, but directly using algorithm
488 agnostic EVP functions. A benefit is that this should be future
489 proof for public key algorithms to come.
493 * Change the interpretation of the '--api' configuration option to
494 mean that this is a desired API compatibility level with no
495 further meaning. The previous interpretation, that this would
496 also mean to remove all deprecated symbols up to and including
497 the given version, no requires that 'no-deprecated' is also used
498 in the configuration.
500 When building applications, the desired API compatibility level
501 can be set with the OPENSSL_API_COMPAT macro like before. For
502 API compatibility version below 3.0, the old style numerical
503 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
504 For version 3.0 and on, the value is expected to be the decimal
505 value calculated from the major and minor version like this:
507 MAJOR * 10000 + MINOR * 100
511 -DOPENSSL_API_COMPAT=30000 For 3.0
512 -DOPENSSL_API_COMPAT=30200 For 3.2
514 To hide declarations that are deprecated up to and including the
515 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
516 given when building the application as well.
520 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
521 access to certificate and CRL stores via URIs and OSSL_STORE
524 This adds the following functions:
526 - X509_LOOKUP_store()
527 - X509_STORE_load_file()
528 - X509_STORE_load_path()
529 - X509_STORE_load_store()
530 - SSL_add_store_cert_subjects_to_stack()
531 - SSL_CTX_set_default_verify_store()
532 - SSL_CTX_load_verify_file()
533 - SSL_CTX_load_verify_dir()
534 - SSL_CTX_load_verify_store()
536 Also, the following functions are now deprecated:
538 - X509_STORE_load_locations() (use X509_STORE_load_file(),
539 X509_STORE_load_path() or X509_STORE_load_store() instead)
540 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
541 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
545 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
546 The presence of this system service is determined at run-time.
550 * Added functionality to create an EVP_PKEY context based on data
551 for methods from providers. This takes an algorithm name and a
552 property query string and simply stores them, with the intent
553 that any operation that uses this context will use those strings
554 to fetch the needed methods implicitly, thereby making the port
555 of application written for pre-3.0 OpenSSL easier.
559 * The undocumented function NCONF_WIN32() has been deprecated; for
560 conversion details see the HISTORY section of doc/man5/config.pod
564 * Introduced the new functions EVP_DigestSignInit_ex() and
565 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
566 EVP_DigestVerifyUpdate() have been converted to functions. See the man
567 pages for further details.
571 * Over two thousand fixes were made to the documentation, including:
572 adding missing command flags, better style conformance, documentation
575 *Rich Salz, Richard Levitte*
577 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
578 X25519, X448, Ed25519 and Ed448.
582 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
587 * Deprecated the public definition of ERR_STATE as well as the function
588 ERR_get_state(). This is done in preparation of making ERR_STATE an
593 * Added ERR functionality to give callers access to the stored function
594 names that have replaced the older function code based functions.
596 New functions are ERR_get_error_func(), ERR_peek_error_func(),
597 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
598 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
599 and ERR_peek_last_error_all().
601 These functions have become deprecated: ERR_get_error_line_data(),
602 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
603 ERR_func_error_string().
607 * Extended testing to be verbose for failing tests only. The make variables
608 VERBOSE_FAILURE or VF can be used to enable this:
610 $ make VF=1 test # Unix
611 $ mms /macro=(VF=1) test ! OpenVMS
612 $ nmake VF=1 test # Windows
616 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
617 used even when parsing explicit parameters, when loading a serialized key
618 or calling `EC_GROUP_new_from_ecpkparameters()`/
619 `EC_GROUP_new_from_ecparameters()`.
620 This prevents bypass of security hardening and performance gains,
621 especially for curves with specialized EC_METHODs.
622 By default, if a key encoded with explicit parameters is loaded and later
623 serialized, the output is still encoded with explicit parameters, even if
624 internally a "named" EC_GROUP is used for computation.
628 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
629 this change, EC_GROUP_set_generator would accept order and/or cofactor as
630 NULL. After this change, only the cofactor parameter can be NULL. It also
631 does some minimal sanity checks on the passed order.
636 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
637 An attack is simple, if the first CMS_recipientInfo is valid but the
638 second CMS_recipientInfo is chosen ciphertext. If the second
639 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
640 encryption key will be replaced by garbage, and the message cannot be
641 decoded, but if the RSA decryption fails, the correct encryption key is
642 used and the recipient will not notice the attack.
643 As a work around for this potential attack the length of the decrypted
644 key must be equal to the cipher default key length, in case the
645 certifiate is not given and all recipientInfo are tried out.
646 The old behaviour can be re-enabled in the CMS code by setting the
647 CMS_DEBUG_DECRYPT flag.
651 * Early start up entropy quality from the DEVRANDOM seed source has been
652 improved for older Linux systems. The RAND subsystem will wait for
653 /dev/random to be producing output before seeding from /dev/urandom.
654 The seeded state is stored for future library initialisations using
655 a system global shared memory segment. The shared memory identifier
656 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
657 the desired value. The default identifier is 114.
661 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
662 when primes for RSA keys are computed.
663 Since we previously always generated primes == 2 (mod 3) for RSA keys,
664 the 2-prime and 3-prime RSA modules were easy to distinguish, since
665 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
666 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
667 This avoids possible fingerprinting of newly generated RSA modules.
671 * Correct the extended master secret constant on EBCDIC systems. Without this
672 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
673 negotiate EMS will fail. Unfortunately this also means that TLS connections
674 between EBCDIC systems with this fix, and EBCDIC systems without this
675 fix will fail if they negotiate EMS.
679 * Changed the library initialisation so that the config file is now loaded
680 by default. This was already the case for libssl. It now occurs for both
681 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
682 OPENSSL_init_crypto() to suppress automatic loading of a config file.
686 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
687 where the former acts as a replacement for ERR_put_error(), and the
688 latter replaces the combination ERR_put_error()+ERR_add_error_data().
689 ERR_raise_data() adds more flexibility by taking a format string and
690 an arbitrary number of arguments following it, to be processed with
695 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
696 to check if a named provider is loaded and available. When called, it
697 will also activate all fallback providers if such are still present.
701 * Enforce a minimum DH modulus size of 512 bits.
705 * Changed DH parameters to generate the order q subgroup instead of 2q.
706 Previously generated DH parameters are still accepted by DH_check
707 but DH_generate_key works around that by clearing bit 0 of the
708 private key for those. This avoids leaking bit 0 of the private key.
712 * Significantly reduce secure memory usage by the randomness pools.
716 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
721 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
722 algorithms. An implementation of a key exchange algorithm can be obtained
723 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
724 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
725 the older EVP_PKEY_derive_init() function. See the man pages for the new
726 functions for further details.
730 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
734 * Removed the function names from error messages and deprecated the
737 * Removed NextStep support and the macro OPENSSL_UNISTD
741 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
742 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
743 Also removed "export var as function" capability; we do not export
744 variables, only functions.
748 * RC5_32_set_key has been changed to return an int type, with 0 indicating
749 an error and 1 indicating success. In previous versions of OpenSSL this
750 was a void type. If a key was set longer than the maximum possible this
755 * Support SM2 signing and verification schemes with X509 certificate.
759 * Use SHA256 as the default digest for TS query in the ts app.
763 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
764 This checks that the salt length is at least 128 bits, the derived key
765 length is at least 112 bits, and that the iteration count is at least 1000.
766 For backwards compatibility these checks are disabled by default in the
767 default provider, but are enabled by default in the fips provider.
768 To enable or disable these checks use the control
769 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
773 * Default cipher lists/suites are now available via a function, the
774 #defines are deprecated.
778 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
779 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
780 for Windows Store apps easier. Also, the "no-uplink" option has been added.
784 * Join the directories crypto/x509 and crypto/x509v3
788 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
789 This changes the size when using the genpkey app when no size is given. It
790 fixes an omission in earlier changes that changed all RSA, DSA and DH
791 generation apps to use 2048 bits by default.
795 * Added command 'openssl kdf' that uses the EVP_KDF API.
799 * Added command 'openssl mac' that uses the EVP_MAC API.
803 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
804 as default directories. Also added the command 'openssl info'
805 for scripting purposes.
809 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
810 deprecated. These undocumented functions were never integrated into the EVP
811 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
812 Bi-directional IGE mode. These modes were never formally standardised and
813 usage of these functions is believed to be very small. In particular
814 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
815 is ever used. The security implications are believed to be minimal, but
816 this issue was never fixed for backwards compatibility reasons. New code
817 should not use these modes.
821 * Add prediction resistance to the DRBG reseeding process.
825 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
826 mandated by IEEE Std 1619-2018.
830 * Added newline escaping functionality to a filename when using openssl dgst.
831 This output format is to replicate the output format found in the `*sum`
832 checksum programs. This aims to preserve backward compatibility.
834 *Matt Eaton, Richard Levitte, and Paul Dale*
836 * Removed the heartbeat message in DTLS feature, as it has very
837 little usage and doesn't seem to fulfill a valuable purpose.
838 The configuration option is now deprecated.
842 * Changed the output of 'openssl {digestname} < file' to display the
843 digest name in its output.
847 * Added a new generic trace API which provides support for enabling
848 instrumentation through trace output. This feature is mainly intended
849 as an aid for developers and is disabled by default. To utilize it,
850 OpenSSL needs to be configured with the `enable-trace` option.
852 If the tracing API is enabled, the application can activate trace output
853 by registering BIOs as trace channels for a number of tracing and debugging
856 The 'openssl' application has been expanded to enable any of the types
857 available via environment variables defined by the user, and serves as
858 one possible example on how to use this functionality.
860 *Richard Levitte & Matthias St. Pierre*
862 * Added build tests for C++. These are generated files that only do one
863 thing, to include one public OpenSSL head file each. This tests that
864 the public header files can be usefully included in a C++ application.
866 This test isn't enabled by default. It can be enabled with the option
867 'enable-buildtest-c++'.
871 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
875 * Add KMAC to EVP_MAC.
879 * Added property based algorithm implementation selection framework to
884 * Added SCA hardening for modular field inversion in EC_GROUP through
885 a new dedicated field_inv() pointer in EC_METHOD.
886 This also addresses a leakage affecting conversions from projective
887 to affine coordinates.
889 *Billy Bob Brumley, Nicola Tuveri*
891 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
892 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
893 those algorithms that were already supported through the EVP_PKEY API
894 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
895 and scrypt are now wrappers that call EVP_KDF.
899 * Build devcrypto engine as a dynamic engine.
903 * Add keyed BLAKE2 to EVP_MAC.
907 * Fix a bug in the computation of the endpoint-pair shared secret used
908 by DTLS over SCTP. This breaks interoperability with older versions
909 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
910 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
911 interoperability with such broken implementations. However, enabling
912 this switch breaks interoperability with correct implementations.
914 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
915 re-used X509_PUBKEY object if the second PUBKEY is malformed.
919 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
923 * Change the license to the Apache License v2.0.
927 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
929 - Major releases (indicated by incrementing the MAJOR release number)
930 may introduce incompatible API/ABI changes.
931 - Minor releases (indicated by incrementing the MINOR release number)
932 may introduce new features but retain API/ABI compatibility.
933 - Patch releases (indicated by incrementing the PATCH number)
934 are intended for bug fixes and other improvements of existing
935 features only (like improving performance or adding documentation)
936 and retain API/ABI compatibility.
940 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
944 * Remove the 'dist' target and add a tarball building script. The
945 'dist' target has fallen out of use, and it shouldn't be
946 necessary to configure just to create a source distribution.
950 * Recreate the OS390-Unix config target. It no longer relies on a
951 special script like it did for OpenSSL pre-1.1.0.
955 * Instead of having the source directories listed in Configure, add
956 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
961 * Add GMAC to EVP_MAC.
965 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
969 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
970 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
971 to facilitate the continued use of MACs through raw private keys in
972 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
976 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
977 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
981 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
982 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
983 are retained for backwards compatibility.
987 * AES-XTS mode now enforces that its two keys are different to mitigate
988 the attacked described in "Efficient Instantiations of Tweakable
989 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
990 Details of this attack can be obtained from:
991 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
995 * Rename the object files, i.e. give them other names than in previous
996 versions. Their names now include the name of the final product, as
997 well as its type mnemonic (bin, lib, shlib).
1001 * Added new option for 'openssl list', '-objects', which will display the
1002 list of built in objects, i.e. OIDs with names.
1006 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1007 improves application performance by removing data copies and providing
1008 applications with zero-copy system calls such as sendfile and splice.
1015 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1017 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1019 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1020 while reading in libssl then we would report an error back to the
1021 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1022 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1023 therefore give a hint as to what went wrong.
1027 * Check that ed25519 and ed448 are allowed by the security level. Previously
1028 signature algorithms not using an MD were not being checked that they were
1029 allowed by the security level.
1033 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1034 was not quite right. The behaviour was not consistent between resumption
1035 and normal handshakes, and also not quite consistent with historical
1036 behaviour. The behaviour in various scenarios has been clarified and
1037 it has been updated to make it match historical behaviour as closely as
1042 * *[VMS only]* The header files that the VMS compilers include automatically,
1043 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1044 that the C++ compiler doesn't understand. This is a shortcoming in the
1045 compiler, but can be worked around with `__cplusplus` guards.
1047 C++ applications that use OpenSSL libraries must be compiled using the
1048 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1049 functions. Otherwise, only functions with symbols of less than 31
1050 characters can be used, as the linker will not be able to successfully
1051 resolve symbols with longer names.
1055 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1056 The presence of this system service is determined at run-time.
1060 * Added newline escaping functionality to a filename when using openssl dgst.
1061 This output format is to replicate the output format found in the `*sum`
1062 checksum programs. This aims to preserve backward compatibility.
1064 *Matt Eaton, Richard Levitte, and Paul Dale*
1066 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1071 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1073 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1074 number generator (RNG). This was intended to include protection in the
1075 event of a fork() system call in order to ensure that the parent and child
1076 processes did not share the same RNG state. However this protection was not
1077 being used in the default case.
1079 A partial mitigation for this issue is that the output from a high
1080 precision timer is mixed into the RNG state so the likelihood of a parent
1081 and child process sharing state is significantly reduced.
1083 If an application already calls OPENSSL_init_crypto() explicitly using
1084 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1087 *Matthias St. Pierre*
1089 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1090 used even when parsing explicit parameters, when loading a serialized key
1091 or calling `EC_GROUP_new_from_ecpkparameters()`/
1092 `EC_GROUP_new_from_ecparameters()`.
1093 This prevents bypass of security hardening and performance gains,
1094 especially for curves with specialized EC_METHODs.
1095 By default, if a key encoded with explicit parameters is loaded and later
1096 serialized, the output is still encoded with explicit parameters, even if
1097 internally a "named" EC_GROUP is used for computation.
1101 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1102 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1103 NULL. After this change, only the cofactor parameter can be NULL. It also
1104 does some minimal sanity checks on the passed order.
1109 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1110 An attack is simple, if the first CMS_recipientInfo is valid but the
1111 second CMS_recipientInfo is chosen ciphertext. If the second
1112 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1113 encryption key will be replaced by garbage, and the message cannot be
1114 decoded, but if the RSA decryption fails, the correct encryption key is
1115 used and the recipient will not notice the attack.
1116 As a work around for this potential attack the length of the decrypted
1117 key must be equal to the cipher default key length, in case the
1118 certifiate is not given and all recipientInfo are tried out.
1119 The old behaviour can be re-enabled in the CMS code by setting the
1120 CMS_DEBUG_DECRYPT flag.
1125 * Early start up entropy quality from the DEVRANDOM seed source has been
1126 improved for older Linux systems. The RAND subsystem will wait for
1127 /dev/random to be producing output before seeding from /dev/urandom.
1128 The seeded state is stored for future library initialisations using
1129 a system global shared memory segment. The shared memory identifier
1130 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1131 the desired value. The default identifier is 114.
1135 * Correct the extended master secret constant on EBCDIC systems. Without this
1136 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1137 negotiate EMS will fail. Unfortunately this also means that TLS connections
1138 between EBCDIC systems with this fix, and EBCDIC systems without this
1139 fix will fail if they negotiate EMS.
1143 * Use Windows installation paths in the mingw builds
1145 Mingw isn't a POSIX environment per se, which means that Windows
1146 paths should be used for installation.
1151 * Changed DH_check to accept parameters with order q and 2q subgroups.
1152 With order 2q subgroups the bit 0 of the private key is not secret
1153 but DH_generate_key works around that by clearing bit 0 of the
1154 private key for those. This avoids leaking bit 0 of the private key.
1158 * Significantly reduce secure memory usage by the randomness pools.
1162 * Revert the DEVRANDOM_WAIT feature for Linux systems
1164 The DEVRANDOM_WAIT feature added a select() call to wait for the
1165 /dev/random device to become readable before reading from the
1166 /dev/urandom device.
1168 It turned out that this change had negative side effects on
1169 performance which were not acceptable. After some discussion it
1170 was decided to revert this feature and leave it up to the OS
1171 resp. the platform maintainer to ensure a proper initialization
1172 during early boot time.
1174 *Matthias St. Pierre*
1176 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1178 * Add build tests for C++. These are generated files that only do one
1179 thing, to include one public OpenSSL head file each. This tests that
1180 the public header files can be usefully included in a C++ application.
1182 This test isn't enabled by default. It can be enabled with the option
1183 'enable-buildtest-c++'.
1187 * Enable SHA3 pre-hashing for ECDSA and DSA.
1191 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1192 This changes the size when using the genpkey app when no size is given. It
1193 fixes an omission in earlier changes that changed all RSA, DSA and DH
1194 generation apps to use 2048 bits by default.
1198 * Reorganize the manual pages to consistently have RETURN VALUES,
1199 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1200 util/fix-doc-nits accordingly.
1202 *Paul Yang, Joshua Lock*
1204 * Add the missing accessor EVP_PKEY_get0_engine()
1208 * Have apps like 's_client' and 's_server' output the signature scheme
1209 along with other cipher suite parameters when debugging.
1213 * Make OPENSSL_config() error agnostic again.
1217 * Do the error handling in RSA decryption constant time.
1221 * Prevent over long nonces in ChaCha20-Poly1305.
1223 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1224 for every encryption operation. RFC 7539 specifies that the nonce value
1225 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1226 and front pads the nonce with 0 bytes if it is less than 12
1227 bytes. However it also incorrectly allows a nonce to be set of up to 16
1228 bytes. In this case only the last 12 bytes are significant and any
1229 additional leading bytes are ignored.
1231 It is a requirement of using this cipher that nonce values are
1232 unique. Messages encrypted using a reused nonce value are susceptible to
1233 serious confidentiality and integrity attacks. If an application changes
1234 the default nonce length to be longer than 12 bytes and then makes a
1235 change to the leading bytes of the nonce expecting the new value to be a
1236 new unique nonce then such an application could inadvertently encrypt
1237 messages with a reused nonce.
1239 Additionally the ignored bytes in a long nonce are not covered by the
1240 integrity guarantee of this cipher. Any application that relies on the
1241 integrity of these ignored leading bytes of a long nonce may be further
1242 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1243 is safe because no such use sets such a long nonce value. However user
1244 applications that use this cipher directly and set a non-default nonce
1245 length to be longer than 12 bytes may be vulnerable.
1247 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1253 * Add DEVRANDOM_WAIT feature for Linux systems
1255 On older Linux systems where the getrandom() system call is not available,
1256 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1257 Contrary to getrandom(), the /dev/urandom device will not block during
1258 early boot when the kernel CSPRNG has not been seeded yet.
1260 To mitigate this known weakness, use select() to wait for /dev/random to
1261 become readable before reading from /dev/urandom.
1263 * Ensure that SM2 only uses SM3 as digest algorithm
1267 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1269 * Change the info callback signals for the start and end of a post-handshake
1270 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1271 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1272 confused by this and assume that a TLSv1.2 renegotiation has started. This
1273 can break KeyUpdate handling. Instead we no longer signal the start and end
1274 of a post handshake message exchange (although the messages themselves are
1275 still signalled). This could break some applications that were expecting
1276 the old signals. However without this KeyUpdate is not usable for many
1281 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1283 * Timing vulnerability in DSA signature generation
1285 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1286 timing side channel attack. An attacker could use variations in the signing
1287 algorithm to recover the private key.
1289 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1294 * Timing vulnerability in ECDSA signature generation
1296 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1297 timing side channel attack. An attacker could use variations in the signing
1298 algorithm to recover the private key.
1300 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1305 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1306 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1307 of two gigabytes and the error handling improved.
1309 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1310 categorized as a normal bug, not a security issue, because the DRBG reseeds
1311 automatically and is fully functional even without additional randomness
1312 provided by the application.
1314 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1316 * Add a new ClientHello callback. Provides a callback interface that gives
1317 the application the ability to adjust the nascent SSL object at the
1318 earliest stage of ClientHello processing, immediately after extensions have
1319 been collected but before they have been processed. In particular, this
1320 callback can adjust the supported TLS versions in response to the contents
1325 * Add SM2 base algorithm support.
1329 * s390x assembly pack: add (improved) hardware-support for the following
1330 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1331 aes-cfb/cfb8, aes-ecb.
1335 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1336 parameter is no longer accepted, as it leads to a corrupt table. NULL
1337 pem_str is reserved for alias entries only.
1341 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1342 step for prime curves. The new implementation is based on formulae from
1343 differential addition-and-doubling in homogeneous projective coordinates
1344 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1345 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1346 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1347 to work in projective coordinates.
1349 *Billy Bob Brumley, Nicola Tuveri*
1351 * Change generating and checking of primes so that the error rate of not
1352 being prime depends on the intended use based on the size of the input.
1353 For larger primes this will result in more rounds of Miller-Rabin.
1354 The maximal error rate for primes with more than 1080 bits is lowered
1357 *Kurt Roeckx, Annie Yousar*
1359 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1363 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1364 moving between systems, and to avoid confusion when a Windows build is
1365 done with mingw vs with MSVC. For POSIX installs, there's still a
1366 symlink or copy named 'tsget' to avoid that confusion as well.
1370 * Revert blinding in ECDSA sign and instead make problematic addition
1371 length-invariant. Switch even to fixed-length Montgomery multiplication.
1375 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1376 step for binary curves. The new implementation is based on formulae from
1377 differential addition-and-doubling in mixed Lopez-Dahab projective
1378 coordinates, modified to independently blind the operands.
1380 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1382 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1383 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1384 EC_METHODs to implement their own specialized "ladder step", to take
1385 advantage of more favorable coordinate systems or more efficient
1386 differential addition-and-doubling algorithms.
1388 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1390 * Modified the random device based seed sources to keep the relevant
1391 file descriptors open rather than reopening them on each access.
1392 This allows such sources to operate in a chroot() jail without
1393 the associated device nodes being available. This behaviour can be
1394 controlled using RAND_keep_random_devices_open().
1398 * Numerous side-channel attack mitigations have been applied. This may have
1399 performance impacts for some algorithms for the benefit of improved
1400 security. Specific changes are noted in this change log by their respective
1405 * AIX shared library support overhaul. Switch to AIX "natural" way of
1406 handling shared libraries, which means collecting shared objects of
1407 different versions and bitnesses in one common archive. This allows to
1408 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1409 doesn't affect the way 3rd party applications are linked, only how
1410 multi-version installation is managed.
1414 * Make ec_group_do_inverse_ord() more robust and available to other
1415 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1416 mitigations are applied to the fallback BN_mod_inverse().
1417 When using this function rather than BN_mod_inverse() directly, new
1418 EC cryptosystem implementations are then safer-by-default.
1422 * Add coordinate blinding for EC_POINT and implement projective
1423 coordinate blinding for generic prime curves as a countermeasure to
1424 chosen point SCA attacks.
1426 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1428 * Add blinding to ECDSA and DSA signatures to protect against side channel
1429 attacks discovered by Keegan Ryan (NCC Group).
1433 * Enforce checking in the pkeyutl command line app to ensure that the input
1434 length does not exceed the maximum supported digest length when performing
1435 a sign, verify or verifyrecover operation.
1439 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1440 I/O in combination with something like select() or poll() will hang. This
1441 can be turned off again using SSL_CTX_clear_mode().
1442 Many applications do not properly handle non-application data records, and
1443 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1444 around the problems in those applications, but can also break some.
1445 It's recommended to read the manpages about SSL_read(), SSL_write(),
1446 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1447 SSL_CTX_set_read_ahead() again.
1451 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1452 now allow empty (zero character) pass phrases.
1456 * Apply blinding to binary field modular inversion and remove patent
1457 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1461 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1462 binary and prime elliptic curves.
1466 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1467 constant time fixed point multiplication.
1471 * Revise elliptic curve scalar multiplication with timing attack
1472 defenses: ec_wNAF_mul redirects to a constant time implementation
1473 when computing fixed point and variable point multiplication (which
1474 in OpenSSL are mostly used with secret scalars in keygen, sign,
1475 ECDH derive operations).
1476 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1479 * Updated CONTRIBUTING
1483 * Updated DRBG / RAND to request nonce and additional low entropy
1484 randomness from the system.
1486 *Matthias St. Pierre*
1488 * Updated 'openssl rehash' to use OpenSSL consistent default.
1492 * Moved the load of the ssl_conf module to libcrypto, which helps
1493 loading engines that libssl uses before libssl is initialised.
1497 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1501 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1503 *Ingo Schwarze, Rich Salz*
1505 * Added output of accepting IP address and port for 'openssl s_server'
1509 * Added a new API for TLSv1.3 ciphersuites:
1510 SSL_CTX_set_ciphersuites()
1511 SSL_set_ciphersuites()
1515 * Memory allocation failures consistently add an error to the error
1520 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1521 in libcrypto when run as setuid/setgid.
1525 * Load any config file by default when libssl is used.
1529 * Added new public header file <openssl/rand_drbg.h> and documentation
1530 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1532 *Matthias St. Pierre*
1534 * QNX support removed (cannot find contributors to get their approval
1535 for the license change).
1539 * TLSv1.3 replay protection for early data has been implemented. See the
1540 SSL_read_early_data() man page for further details.
1544 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1545 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1546 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1547 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1548 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1549 configuration has been separated out. See the ciphers man page or the
1550 SSL_CTX_set_ciphersuites() man page for more information.
1554 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1555 in responder mode now supports the new "-multi" option, which
1556 spawns the specified number of child processes to handle OCSP
1557 requests. The "-timeout" option now also limits the OCSP
1558 responder's patience to wait to receive the full client request
1559 on a newly accepted connection. Child processes are respawned
1560 as needed, and the CA index file is automatically reloaded
1561 when changed. This makes it possible to run the "ocsp" responder
1562 as a long-running service, making the OpenSSL CA somewhat more
1563 feature-complete. In this mode, most diagnostic messages logged
1564 after entering the event loop are logged via syslog(3) rather than
1569 * Added support for X448 and Ed448. Heavily based on original work by
1574 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1575 objects loaded. This adds the functions OSSL_STORE_expect() and
1576 OSSL_STORE_find() as well as needed tools to construct searches and
1577 get the search data out of them.
1581 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1582 version of OpenSSL should review their configuration settings to ensure
1583 that they are still appropriate for TLSv1.3. For further information see:
1584 <https://wiki.openssl.org/index.php/TLS1.3>
1588 * Grand redesign of the OpenSSL random generator
1590 The default RAND method now utilizes an AES-CTR DRBG according to
1591 NIST standard SP 800-90Ar1. The new random generator is essentially
1592 a port of the default random generator from the OpenSSL FIPS 2.0
1593 object module. It is a hybrid deterministic random bit generator
1594 using an AES-CTR bit stream and which seeds and reseeds itself
1595 automatically using trusted system entropy sources.
1597 Some of its new features are:
1598 - Support for multiple DRBG instances with seed chaining.
1599 - The default RAND method makes use of a DRBG.
1600 - There is a public and private DRBG instance.
1601 - The DRBG instances are fork-safe.
1602 - Keep all global DRBG instances on the secure heap if it is enabled.
1603 - The public and private DRBG instance are per thread for lock free
1606 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1608 * Changed Configure so it only says what it does and doesn't dump
1609 so much data. Instead, ./configdata.pm should be used as a script
1610 to display all sorts of configuration data.
1614 * Added processing of "make variables" to Configure.
1618 * Added SHA512/224 and SHA512/256 algorithm support.
1622 * The last traces of Netware support, first removed in 1.1.0, have
1627 * Get rid of Makefile.shared, and in the process, make the processing
1628 of certain files (rc.obj, or the .def/.map/.opt files produced from
1629 the ordinal files) more visible and hopefully easier to trace and
1630 debug (or make silent).
1634 * Make it possible to have environment variable assignments as
1635 arguments to config / Configure.
1639 * Add multi-prime RSA (RFC 8017) support.
1643 * Add SM3 implemented according to GB/T 32905-2016
1644 * Jack Lloyd <jack.lloyd@ribose.com>,
1645 Ronald Tse <ronald.tse@ribose.com>,
1646 Erick Borsboom <erick.borsboom@ribose.com> *
1648 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1649 as documented in RFC6066.
1650 Based on a patch from Tomasz Moń
1652 *Filipe Raimundo da Silva*
1654 * Add SM4 implemented according to GB/T 32907-2016.
1655 * Jack Lloyd <jack.lloyd@ribose.com>,
1656 Ronald Tse <ronald.tse@ribose.com>,
1657 Erick Borsboom <erick.borsboom@ribose.com> *
1659 * Reimplement -newreq-nodes and ERR_error_string_n; the
1660 original author does not agree with the license change.
1664 * Add ARIA AEAD TLS support.
1668 * Some macro definitions to support VS6 have been removed. Visual
1669 Studio 6 has not worked since 1.1.0
1673 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1674 without clearing the errors.
1678 * Add "atfork" functions. If building on a system that without
1679 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1680 requirements. The RAND facility now uses/requires this.
1688 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1689 not possible to disable entirely. However, it's still possible to
1690 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1693 To disable, configure with 'no-ui-console'. 'no-ui' is still
1694 possible to use as an alias. Check at compile time with the
1695 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1696 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1700 * Add a STORE module, which implements a uniform and URI based reader of
1701 stores that can contain keys, certificates, CRLs and numerous other
1702 objects. The main API is loosely based on a few stdio functions,
1703 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1704 OSSL_STORE_error and OSSL_STORE_close.
1705 The implementation uses backends called "loaders" to implement arbitrary
1706 URI schemes. There is one built in "loader" for the 'file' scheme.
1710 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1711 then adjusted to work on FreeBSD 8.4 as well.
1712 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1713 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1717 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1718 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1719 error code calls like this:
1721 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1723 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1724 that can be encoded in C. For the foreseeable future, this will only
1727 *Richard Levitte and Tim Hudson*
1729 * Removed BSD cryptodev engine.
1733 * Add a build target 'build_all_generated', to build all generated files
1734 and only that. This can be used to prepare everything that requires
1735 things like perl for a system that lacks perl and then move everything
1736 to that system and do the rest of the build there.
1740 * In the UI interface, make it possible to duplicate the user data. This
1741 can be used by engines that need to retain the data for a longer time
1742 than just the call where this user data is passed.
1746 * Ignore the '-named_curve auto' value for compatibility of applications
1749 *Tomas Mraz <tmraz@fedoraproject.org>*
1751 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1752 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1753 alerts across multiple records (some of which could be empty). In practice
1754 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1755 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1756 support this at all. Supporting it adds significant complexity to the
1757 record layer, and its removal is unlikely to cause interoperability
1762 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1763 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1764 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1769 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1770 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1772 *Richard Levitte, Andy Polyakov*
1774 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1779 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1780 platform rather than 'mingw'.
1784 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1785 success if they are asked to add an object which already exists
1786 in the store. This change cascades to other functions which load
1787 certificates and CRLs.
1791 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1792 facilitate stack unwinding even from assembly subroutines.
1796 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1797 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1801 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1802 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1803 which is the minimum version we support.
1807 * Certificate time validation (X509_cmp_time) enforces stricter
1808 compliance with RFC 5280. Fractional seconds and timezone offsets
1809 are no longer allowed.
1813 * Add support for ARIA
1817 * s_client will now send the Server Name Indication (SNI) extension by
1818 default unless the new "-noservername" option is used. The server name is
1819 based on the host provided to the "-connect" option unless overridden by
1820 using "-servername".
1824 * Add support for SipHash
1828 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1829 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1830 prevent issues where no progress is being made and the peer continually
1831 sends unrecognised record types, using up resources processing them.
1835 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1836 using the algorithm defined in
1837 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1841 * Heartbeat support has been removed; the ABI is changed for now.
1843 *Richard Levitte, Rich Salz*
1845 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1849 * The RSA "null" method, which was partially supported to avoid patent
1850 issues, has been replaced to always returns NULL.
1857 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1859 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1860 used even when parsing explicit parameters, when loading a serialized key
1861 or calling `EC_GROUP_new_from_ecpkparameters()`/
1862 `EC_GROUP_new_from_ecparameters()`.
1863 This prevents bypass of security hardening and performance gains,
1864 especially for curves with specialized EC_METHODs.
1865 By default, if a key encoded with explicit parameters is loaded and later
1866 serialized, the output is still encoded with explicit parameters, even if
1867 internally a "named" EC_GROUP is used for computation.
1871 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1872 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1873 NULL. After this change, only the cofactor parameter can be NULL. It also
1874 does some minimal sanity checks on the passed order.
1879 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1880 An attack is simple, if the first CMS_recipientInfo is valid but the
1881 second CMS_recipientInfo is chosen ciphertext. If the second
1882 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1883 encryption key will be replaced by garbage, and the message cannot be
1884 decoded, but if the RSA decryption fails, the correct encryption key is
1885 used and the recipient will not notice the attack.
1886 As a work around for this potential attack the length of the decrypted
1887 key must be equal to the cipher default key length, in case the
1888 certifiate is not given and all recipientInfo are tried out.
1889 The old behaviour can be re-enabled in the CMS code by setting the
1890 CMS_DEBUG_DECRYPT flag.
1895 * Use Windows installation paths in the mingw builds
1897 Mingw isn't a POSIX environment per se, which means that Windows
1898 paths should be used for installation.
1903 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
1905 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1906 This changes the size when using the genpkey app when no size is given. It
1907 fixes an omission in earlier changes that changed all RSA, DSA and DH
1908 generation apps to use 2048 bits by default.
1912 * Prevent over long nonces in ChaCha20-Poly1305.
1914 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1915 for every encryption operation. RFC 7539 specifies that the nonce value
1916 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1917 and front pads the nonce with 0 bytes if it is less than 12
1918 bytes. However it also incorrectly allows a nonce to be set of up to 16
1919 bytes. In this case only the last 12 bytes are significant and any
1920 additional leading bytes are ignored.
1922 It is a requirement of using this cipher that nonce values are
1923 unique. Messages encrypted using a reused nonce value are susceptible to
1924 serious confidentiality and integrity attacks. If an application changes
1925 the default nonce length to be longer than 12 bytes and then makes a
1926 change to the leading bytes of the nonce expecting the new value to be a
1927 new unique nonce then such an application could inadvertently encrypt
1928 messages with a reused nonce.
1930 Additionally the ignored bytes in a long nonce are not covered by the
1931 integrity guarantee of this cipher. Any application that relies on the
1932 integrity of these ignored leading bytes of a long nonce may be further
1933 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1934 is safe because no such use sets such a long nonce value. However user
1935 applications that use this cipher directly and set a non-default nonce
1936 length to be longer than 12 bytes may be vulnerable.
1938 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1944 * Added SCA hardening for modular field inversion in EC_GROUP through
1945 a new dedicated field_inv() pointer in EC_METHOD.
1946 This also addresses a leakage affecting conversions from projective
1947 to affine coordinates.
1949 *Billy Bob Brumley, Nicola Tuveri*
1951 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1952 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1956 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1960 * Remove the 'dist' target and add a tarball building script. The
1961 'dist' target has fallen out of use, and it shouldn't be
1962 necessary to configure just to create a source distribution.
1966 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
1968 * Timing vulnerability in DSA signature generation
1970 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1971 timing side channel attack. An attacker could use variations in the signing
1972 algorithm to recover the private key.
1974 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1979 * Timing vulnerability in ECDSA signature generation
1981 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1982 timing side channel attack. An attacker could use variations in the signing
1983 algorithm to recover the private key.
1985 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1990 * Add coordinate blinding for EC_POINT and implement projective
1991 coordinate blinding for generic prime curves as a countermeasure to
1992 chosen point SCA attacks.
1994 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1996 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
1998 * Client DoS due to large DH parameter
2000 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2001 malicious server can send a very large prime value to the client. This will
2002 cause the client to spend an unreasonably long period of time generating a
2003 key for this prime resulting in a hang until the client has finished. This
2004 could be exploited in a Denial Of Service attack.
2006 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2011 * Cache timing vulnerability in RSA Key Generation
2013 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2014 a cache timing side channel attack. An attacker with sufficient access to
2015 mount cache timing attacks during the RSA key generation process could
2016 recover the private key.
2018 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2019 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2024 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2025 parameter is no longer accepted, as it leads to a corrupt table. NULL
2026 pem_str is reserved for alias entries only.
2030 * Revert blinding in ECDSA sign and instead make problematic addition
2031 length-invariant. Switch even to fixed-length Montgomery multiplication.
2035 * Change generating and checking of primes so that the error rate of not
2036 being prime depends on the intended use based on the size of the input.
2037 For larger primes this will result in more rounds of Miller-Rabin.
2038 The maximal error rate for primes with more than 1080 bits is lowered
2041 *Kurt Roeckx, Annie Yousar*
2043 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2047 * Add blinding to ECDSA and DSA signatures to protect against side channel
2048 attacks discovered by Keegan Ryan (NCC Group).
2052 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2053 now allow empty (zero character) pass phrases.
2057 * Certificate time validation (X509_cmp_time) enforces stricter
2058 compliance with RFC 5280. Fractional seconds and timezone offsets
2059 are no longer allowed.
2063 * Fixed a text canonicalisation bug in CMS
2065 Where a CMS detached signature is used with text content the text goes
2066 through a canonicalisation process first prior to signing or verifying a
2067 signature. This process strips trailing space at the end of lines, converts
2068 line terminators to CRLF and removes additional trailing line terminators
2069 at the end of a file. A bug in the canonicalisation process meant that
2070 some characters, such as form-feed, were incorrectly treated as whitespace
2071 and removed. This is contrary to the specification (RFC5485). This fix
2072 could mean that detached text data signed with an earlier version of
2073 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2074 signed with a fixed OpenSSL may fail to verify with an earlier version of
2075 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2076 and use the "-binary" flag (for the "cms" command line application) or set
2077 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2081 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2083 * Constructed ASN.1 types with a recursive definition could exceed the stack
2085 Constructed ASN.1 types with a recursive definition (such as can be found
2086 in PKCS7) could eventually exceed the stack given malicious input with
2087 excessive recursion. This could result in a Denial Of Service attack. There
2088 are no such structures used within SSL/TLS that come from untrusted sources
2089 so this is considered safe.
2091 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2097 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2099 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2100 effectively reduced to only comparing the least significant bit of each
2101 byte. This allows an attacker to forge messages that would be considered as
2102 authenticated in an amount of tries lower than that guaranteed by the
2103 security claims of the scheme. The module can only be compiled by the
2104 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2106 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2112 * Add a build target 'build_all_generated', to build all generated files
2113 and only that. This can be used to prepare everything that requires
2114 things like perl for a system that lacks perl and then move everything
2115 to that system and do the rest of the build there.
2119 * Backport SSL_OP_NO_RENGOTIATION
2121 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2122 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2123 changes this is no longer possible in 1.1.0. Therefore the new
2124 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2125 1.1.0 to provide equivalent functionality.
2127 Note that if an application built against 1.1.0h headers (or above) is run
2128 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2129 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2133 * Removed the OS390-Unix config target. It relied on a script that doesn't
2138 * rsaz_1024_mul_avx2 overflow bug on x86_64
2140 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2141 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2142 Analysis suggests that attacks against RSA and DSA as a result of this
2143 defect would be very difficult to perform and are not believed likely.
2144 Attacks against DH1024 are considered just feasible, because most of the
2145 work necessary to deduce information about a private key may be performed
2146 offline. The amount of resources required for such an attack would be
2147 significant. However, for an attack on TLS to be meaningful, the server
2148 would have to share the DH1024 private key among multiple clients, which is
2149 no longer an option since CVE-2016-0701.
2151 This only affects processors that support the AVX2 but not ADX extensions
2152 like Intel Haswell (4th generation).
2154 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2155 was originally found via the OSS-Fuzz project.
2160 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2162 * bn_sqrx8x_internal carry bug on x86_64
2164 There is a carry propagating bug in the x86_64 Montgomery squaring
2165 procedure. No EC algorithms are affected. Analysis suggests that attacks
2166 against RSA and DSA as a result of this defect would be very difficult to
2167 perform and are not believed likely. Attacks against DH are considered just
2168 feasible (although very difficult) because most of the work necessary to
2169 deduce information about a private key may be performed offline. The amount
2170 of resources required for such an attack would be very significant and
2171 likely only accessible to a limited number of attackers. An attacker would
2172 additionally need online access to an unpatched system using the target
2173 private key in a scenario with persistent DH parameters and a private
2174 key that is shared between multiple clients.
2176 This only affects processors that support the BMI1, BMI2 and ADX extensions
2177 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2179 This issue was reported to OpenSSL by the OSS-Fuzz project.
2184 * Malformed X.509 IPAddressFamily could cause OOB read
2186 If an X.509 certificate has a malformed IPAddressFamily extension,
2187 OpenSSL could do a one-byte buffer overread. The most likely result
2188 would be an erroneous display of the certificate in text format.
2190 This issue was reported to OpenSSL by the OSS-Fuzz project.
2195 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2197 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2198 platform rather than 'mingw'.
2202 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2203 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2204 which is the minimum version we support.
2208 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2210 * Encrypt-Then-Mac renegotiation crash
2212 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2213 negotiated where it was not in the original handshake (or vice-versa) then
2214 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2215 and servers are affected.
2217 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2222 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2224 * Truncated packet could crash via OOB read
2226 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2227 cipher is being used, then a truncated packet can cause that host to
2228 perform an out-of-bounds read, usually resulting in a crash.
2230 This issue was reported to OpenSSL by Robert Święcki of Google.
2235 * Bad (EC)DHE parameters cause a client crash
2237 If a malicious server supplies bad parameters for a DHE or ECDHE key
2238 exchange then this can result in the client attempting to dereference a
2239 NULL pointer leading to a client crash. This could be exploited in a Denial
2242 This issue was reported to OpenSSL by Guido Vranken.
2247 * BN_mod_exp may produce incorrect results on x86_64
2249 There is a carry propagating bug in the x86_64 Montgomery squaring
2250 procedure. No EC algorithms are affected. Analysis suggests that attacks
2251 against RSA and DSA as a result of this defect would be very difficult to
2252 perform and are not believed likely. Attacks against DH are considered just
2253 feasible (although very difficult) because most of the work necessary to
2254 deduce information about a private key may be performed offline. The amount
2255 of resources required for such an attack would be very significant and
2256 likely only accessible to a limited number of attackers. An attacker would
2257 additionally need online access to an unpatched system using the target
2258 private key in a scenario with persistent DH parameters and a private
2259 key that is shared between multiple clients. For example this can occur by
2260 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2261 similar to CVE-2015-3193 but must be treated as a separate problem.
2263 This issue was reported to OpenSSL by the OSS-Fuzz project.
2268 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2270 * ChaCha20/Poly1305 heap-buffer-overflow
2272 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2273 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2274 crash. This issue is not considered to be exploitable beyond a DoS.
2276 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2281 * CMS Null dereference
2283 Applications parsing invalid CMS structures can crash with a NULL pointer
2284 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2285 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2286 structure callback if an attempt is made to free certain invalid encodings.
2287 Only CHOICE structures using a callback which do not handle NULL value are
2290 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2295 * Montgomery multiplication may produce incorrect results
2297 There is a carry propagating bug in the Broadwell-specific Montgomery
2298 multiplication procedure that handles input lengths divisible by, but
2299 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2300 and DH private keys are impossible. This is because the subroutine in
2301 question is not used in operations with the private key itself and an input
2302 of the attacker's direct choice. Otherwise the bug can manifest itself as
2303 transient authentication and key negotiation failures or reproducible
2304 erroneous outcome of public-key operations with specially crafted input.
2305 Among EC algorithms only Brainpool P-512 curves are affected and one
2306 presumably can attack ECDH key negotiation. Impact was not analyzed in
2307 detail, because pre-requisites for attack are considered unlikely. Namely
2308 multiple clients have to choose the curve in question and the server has to
2309 share the private key among them, neither of which is default behaviour.
2310 Even then only clients that chose the curve will be affected.
2312 This issue was publicly reported as transient failures and was not
2313 initially recognized as a security issue. Thanks to Richard Morgan for
2314 providing reproducible case.
2319 * Removed automatic addition of RPATH in shared libraries and executables,
2320 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2324 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2326 * Fix Use After Free for large message sizes
2328 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2329 message larger than approx 16k is received then the underlying buffer to
2330 store the incoming message is reallocated and moved. Unfortunately a
2331 dangling pointer to the old location is left which results in an attempt to
2332 write to the previously freed location. This is likely to result in a
2333 crash, however it could potentially lead to execution of arbitrary code.
2335 This issue only affects OpenSSL 1.1.0a.
2337 This issue was reported to OpenSSL by Robert Święcki.
2342 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2344 * OCSP Status Request extension unbounded memory growth
2346 A malicious client can send an excessively large OCSP Status Request
2347 extension. If that client continually requests renegotiation, sending a
2348 large OCSP Status Request extension each time, then there will be unbounded
2349 memory growth on the server. This will eventually lead to a Denial Of
2350 Service attack through memory exhaustion. Servers with a default
2351 configuration are vulnerable even if they do not support OCSP. Builds using
2352 the "no-ocsp" build time option are not affected.
2354 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2359 * SSL_peek() hang on empty record
2361 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2362 sends an empty record. This could be exploited by a malicious peer in a
2363 Denial Of Service attack.
2365 This issue was reported to OpenSSL by Alex Gaynor.
2370 * Excessive allocation of memory in tls_get_message_header() and
2371 dtls1_preprocess_fragment()
2373 A (D)TLS message includes 3 bytes for its length in the header for the
2374 message. This would allow for messages up to 16Mb in length. Messages of
2375 this length are excessive and OpenSSL includes a check to ensure that a
2376 peer is sending reasonably sized messages in order to avoid too much memory
2377 being consumed to service a connection. A flaw in the logic of version
2378 1.1.0 means that memory for the message is allocated too early, prior to
2379 the excessive message length check. Due to way memory is allocated in
2380 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2381 to service a connection. This could lead to a Denial of Service through
2382 memory exhaustion. However, the excessive message length check still takes
2383 place, and this would cause the connection to immediately fail. Assuming
2384 that the application calls SSL_free() on the failed connection in a timely
2385 manner then the 21Mb of allocated memory will then be immediately freed
2386 again. Therefore the excessive memory allocation will be transitory in
2387 nature. This then means that there is only a security impact if:
2389 1) The application does not call SSL_free() in a timely manner in the event
2390 that the connection fails
2392 2) The application is working in a constrained environment where there is
2393 very little free memory
2395 3) The attacker initiates multiple connection attempts such that there are
2396 multiple connections in a state where memory has been allocated for the
2397 connection; SSL_free() has not yet been called; and there is insufficient
2398 memory to service the multiple requests.
2400 Except in the instance of (1) above any Denial Of Service is likely to be
2401 transitory because as soon as the connection fails the memory is
2402 subsequently freed again in the SSL_free() call. However there is an
2403 increased risk during this period of application crashes due to the lack of
2404 memory - which would then mean a more serious Denial of Service.
2406 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2407 (CVE-2016-6307 and CVE-2016-6308)
2411 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2412 had to be removed. Primary reason is that vendor assembler can't
2413 assemble our modules with -KPIC flag. As result it, assembly
2414 support, was not even available as option. But its lack means
2415 lack of side-channel resistant code, which is incompatible with
2416 security by todays standards. Fortunately gcc is readily available
2417 prepackaged option, which we firmly point at...
2421 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2423 * Windows command-line tool supports UTF-8 opt-in option for arguments
2424 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2425 (to any value) allows Windows user to access PKCS#12 file generated
2426 with Windows CryptoAPI and protected with non-ASCII password, as well
2427 as files generated under UTF-8 locale on Linux also protected with
2432 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2433 have been disabled by default and removed from DEFAULT, just like RC4.
2434 See the RC4 item below to re-enable both.
2438 * The method for finding the storage location for the Windows RAND seed file
2439 has changed. First we check %RANDFILE%. If that is not set then we check
2440 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2441 all else fails we fall back to C:\.
2445 * The EVP_EncryptUpdate() function has had its return type changed from void
2446 to int. A return of 0 indicates and error while a return of 1 indicates
2451 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2452 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2453 off the constant time implementation for RSA, DSA and DH have been made
2454 no-ops and deprecated.
2458 * Windows RAND implementation was simplified to only get entropy by
2459 calling CryptGenRandom(). Various other RAND-related tickets
2462 *Joseph Wylie Yandle, Rich Salz*
2464 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2465 and `OPENSSL_LH_`, respectively. The old names are available
2466 with API compatibility. They new names are now completely documented.
2470 * Unify TYPE_up_ref(obj) methods signature.
2471 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2472 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2473 int (instead of void) like all others TYPE_up_ref() methods.
2474 So now these methods also check the return value of CRYPTO_atomic_add(),
2475 and the validity of object reference counter.
2477 *fdasilvayy@gmail.com*
2479 * With Windows Visual Studio builds, the .pdb files are installed
2480 alongside the installed libraries and executables. For a static
2481 library installation, ossl_static.pdb is the associate compiler
2482 generated .pdb file to be used when linking programs.
2486 * Remove openssl.spec. Packaging files belong with the packagers.
2490 * Automatic Darwin/OSX configuration has had a refresh, it will now
2491 recognise x86_64 architectures automatically. You can still decide
2492 to build for a different bitness with the environment variable
2493 KERNEL_BITS (can be 32 or 64), for example:
2495 KERNEL_BITS=32 ./config
2499 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2500 256 bit AES and HMAC with SHA256.
2504 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2508 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2512 * To enable users to have their own config files and build file templates,
2513 Configure looks in the directory indicated by the environment variable
2514 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2515 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2516 name and is used as is.
2520 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2521 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2522 X509_CERT_FILE_CTX was removed.
2526 * "shared" builds are now the default. To create only static libraries use
2527 the "no-shared" Configure option.
2531 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2532 All of these option have not worked for some while and are fundamental
2537 * Make various cleanup routines no-ops and mark them as deprecated. Most
2538 global cleanup functions are no longer required because they are handled
2539 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2540 Explicitly de-initing can cause problems (e.g. where a library that uses
2541 OpenSSL de-inits, but an application is still using it). The affected
2542 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2543 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2544 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2545 COMP_zlib_cleanup().
2549 * --strict-warnings no longer enables runtime debugging options
2550 such as REF_DEBUG. Instead, debug options are automatically
2551 enabled with '--debug' builds.
2553 *Andy Polyakov, Emilia Käsper*
2555 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2556 have been moved out of the public header files. New functions for managing
2557 these have been added.
2561 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2562 objects have been moved out of the public header files. New
2563 functions for managing these have been added.
2567 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2568 have been moved out of the public header files. New functions for managing
2569 these have been added.
2573 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2574 moved out of the public header files. New functions for managing these
2579 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2583 * Removed the mk1mf build scripts.
2587 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2588 it is always safe to #include a header now.
2592 * Removed the aged BC-32 config and all its supporting scripts
2596 * Removed support for Ultrix, Netware, and OS/2.
2600 * Add support for HKDF.
2602 *Alessandro Ghedini*
2604 * Add support for blake2b and blake2s
2608 * Added support for "pipelining". Ciphers that have the
2609 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2610 encryptions/decryptions simultaneously. There are currently no built-in
2611 ciphers with this property but the expectation is that engines will be able
2612 to offer it to significantly improve throughput. Support has been extended
2613 into libssl so that multiple records for a single connection can be
2614 processed in one go (for >=TLS 1.1).
2618 * Added the AFALG engine. This is an async capable engine which is able to
2619 offload work to the Linux kernel. In this initial version it only supports
2620 AES128-CBC. The kernel must be version 4.1.0 or greater.
2624 * OpenSSL now uses a new threading API. It is no longer necessary to
2625 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2626 are two supported threading models: pthreads and windows threads. It is
2627 also possible to configure OpenSSL at compile time for "no-threads". The
2628 old threading API should no longer be used. The functions have been
2629 replaced with "no-op" compatibility macros.
2631 *Alessandro Ghedini, Matt Caswell*
2633 * Modify behavior of ALPN to invoke callback after SNI/servername
2634 callback, such that updates to the SSL_CTX affect ALPN.
2638 * Add SSL_CIPHER queries for authentication and key-exchange.
2642 * Changes to the DEFAULT cipherlist:
2643 - Prefer (EC)DHE handshakes over plain RSA.
2644 - Prefer AEAD ciphers over legacy ciphers.
2645 - Prefer ECDSA over RSA when both certificates are available.
2646 - Prefer TLSv1.2 ciphers/PRF.
2647 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2652 * Change the ECC default curve list to be this, in order: x25519,
2653 secp256r1, secp521r1, secp384r1.
2657 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2658 disabled by default. They can be re-enabled using the
2659 enable-weak-ssl-ciphers option to Configure.
2663 * If the server has ALPN configured, but supports no protocols that the
2664 client advertises, send a fatal "no_application_protocol" alert.
2665 This behaviour is SHALL in RFC 7301, though it isn't universally
2666 implemented by other servers.
2670 * Add X25519 support.
2671 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2672 for public and private key encoding using the format documented in
2673 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2674 key generation and key derivation.
2676 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2681 * Deprecate SRP_VBASE_get_by_user.
2682 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2683 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2684 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2685 seed, even if the seed is configured.
2687 Users should use SRP_VBASE_get1_by_user instead. Note that in
2688 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2689 also that even though configuring the SRP seed attempts to hide
2690 invalid usernames by continuing the handshake with fake
2691 credentials, this behaviour is not constant time and no strong
2692 guarantees are made that the handshake is indistinguishable from
2693 that of a valid user.
2697 * Configuration change; it's now possible to build dynamic engines
2698 without having to build shared libraries and vice versa. This
2699 only applies to the engines in engines/, those in crypto/engine/
2700 will always be built into libcrypto (i.e. "static").
2702 Building dynamic engines is enabled by default; to disable, use
2703 the configuration option "disable-dynamic-engine".
2705 The only requirements for building dynamic engines are the
2706 presence of the DSO module and building with position independent
2707 code, so they will also automatically be disabled if configuring
2708 with "disable-dso" or "disable-pic".
2710 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2711 are also taken away from openssl/opensslconf.h, as they are
2716 * Configuration change; if there is a known flag to compile
2717 position independent code, it will always be applied on the
2718 libcrypto and libssl object files, and never on the application
2719 object files. This means other libraries that use routines from
2720 libcrypto / libssl can be made into shared libraries regardless
2721 of how OpenSSL was configured.
2723 If this isn't desirable, the configuration options "disable-pic"
2724 or "no-pic" can be used to disable the use of PIC. This will
2725 also disable building shared libraries and dynamic engines.
2729 * Removed JPAKE code. It was experimental and has no wide use.
2733 * The INSTALL_PREFIX Makefile variable has been renamed to
2734 DESTDIR. That makes for less confusion on what this variable
2735 is for. Also, the configuration option --install_prefix is
2740 * Heartbeat for TLS has been removed and is disabled by default
2741 for DTLS; configure with enable-heartbeats. Code that uses the
2742 old #define's might need to be updated.
2744 *Emilia Käsper, Rich Salz*
2746 * Rename REF_CHECK to REF_DEBUG.
2750 * New "unified" build system
2752 The "unified" build system is aimed to be a common system for all
2753 platforms we support. With it comes new support for VMS.
2755 This system builds supports building in a different directory tree
2756 than the source tree. It produces one Makefile (for unix family
2757 or lookalikes), or one descrip.mms (for VMS).
2759 The source of information to make the Makefile / descrip.mms is
2760 small files called 'build.info', holding the necessary
2761 information for each directory with source to compile, and a
2762 template in Configurations, like unix-Makefile.tmpl or
2765 With this change, the library names were also renamed on Windows
2766 and on VMS. They now have names that are closer to the standard
2767 on Unix, and include the major version number, and in certain
2768 cases, the architecture they are built for. See "Notes on shared
2769 libraries" in INSTALL.
2771 We rely heavily on the perl module Text::Template.
2775 * Added support for auto-initialisation and de-initialisation of the library.
2776 OpenSSL no longer requires explicit init or deinit routines to be called,
2777 except in certain circumstances. See the OPENSSL_init_crypto() and
2778 OPENSSL_init_ssl() man pages for further information.
2782 * The arguments to the DTLSv1_listen function have changed. Specifically the
2783 "peer" argument is now expected to be a BIO_ADDR object.
2785 * Rewrite of BIO networking library. The BIO library lacked consistent
2786 support of IPv6, and adding it required some more extensive
2787 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2788 which hold all types of addresses and chains of address information.
2789 It also introduces a new API, with functions like BIO_socket,
2790 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2791 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2792 have been adapted accordingly.
2796 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2801 * CRIME protection: disable compression by default, even if OpenSSL is
2802 compiled with zlib enabled. Applications can still enable compression
2803 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2804 using the SSL_CONF library to configure compression.
2808 * The signature of the session callback configured with
2809 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2810 was explicitly marked as `const unsigned char*` instead of
2815 * Always DPURIFY. Remove the use of uninitialized memory in the
2816 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2820 * Removed many obsolete configuration items, including
2821 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2822 MD2_CHAR, MD2_INT, MD2_LONG
2824 IDEA_SHORT, IDEA_LONG
2825 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2827 *Rich Salz, with advice from Andy Polyakov*
2829 * Many BN internals have been moved to an internal header file.
2831 *Rich Salz with help from Andy Polyakov*
2833 * Configuration and writing out the results from it has changed.
2834 Files such as Makefile include/openssl/opensslconf.h and are now
2835 produced through general templates, such as Makefile.in and
2836 crypto/opensslconf.h.in and some help from the perl module
2839 Also, the center of configuration information is no longer
2840 Makefile. Instead, Configure produces a perl module in
2841 configdata.pm which holds most of the config data (in the hash
2842 table %config), the target data that comes from the target
2843 configuration in one of the `Configurations/*.conf~ files (in
2848 * To clarify their intended purposes, the Configure options
2849 --prefix and --openssldir change their semantics, and become more
2850 straightforward and less interdependent.
2852 --prefix shall be used exclusively to give the location INSTALLTOP
2853 where programs, scripts, libraries, include files and manuals are
2854 going to be installed. The default is now /usr/local.
2856 --openssldir shall be used exclusively to give the default
2857 location OPENSSLDIR where certificates, private keys, CRLs are
2858 managed. This is also where the default openssl.cnf gets
2860 If the directory given with this option is a relative path, the
2861 values of both the --prefix value and the --openssldir value will
2862 be combined to become OPENSSLDIR.
2863 The default for --openssldir is INSTALLTOP/ssl.
2865 Anyone who uses --openssldir to specify where OpenSSL is to be
2866 installed MUST change to use --prefix instead.
2870 * The GOST engine was out of date and therefore it has been removed. An up
2871 to date GOST engine is now being maintained in an external repository.
2872 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2873 support for GOST ciphersuites (these are only activated if a GOST engine
2878 * EGD is no longer supported by default; use enable-egd when
2881 *Ben Kaduk and Rich Salz*
2883 * The distribution now has Makefile.in files, which are used to
2884 create Makefile's when Configure is run. *Configure must be run
2885 before trying to build now.*
2889 * The return value for SSL_CIPHER_description() for error conditions
2894 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2896 Obtaining and performing DNSSEC validation of TLSA records is
2897 the application's responsibility. The application provides
2898 the TLSA records of its choice to OpenSSL, and these are then
2899 used to authenticate the peer.
2901 The TLSA records need not even come from DNS. They can, for
2902 example, be used to implement local end-entity certificate or
2903 trust-anchor "pinning", where the "pin" data takes the form
2904 of TLSA records, which can augment or replace verification
2905 based on the usual WebPKI public certification authorities.
2909 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2910 continues to support deprecated interfaces in default builds.
2911 However, applications are strongly advised to compile their
2912 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2913 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2914 or the 1.1.0 releases.
2916 In environments in which all applications have been ported to
2917 not use any deprecated interfaces OpenSSL's Configure script
2918 should be used with the --api=1.1.0 option to entirely remove
2919 support for the deprecated features from the library and
2920 unconditionally disable them in the installed headers.
2921 Essentially the same effect can be achieved with the "no-deprecated"
2922 argument to Configure, except that this will always restrict
2923 the build to just the latest API, rather than a fixed API
2926 As applications are ported to future revisions of the API,
2927 they should update their compile-time OPENSSL_API_COMPAT define
2928 accordingly, but in most cases should be able to continue to
2929 compile with later releases.
2931 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2932 0x10000000L and 0x00908000L, respectively. However those
2933 versions did not support the OPENSSL_API_COMPAT feature, and
2934 so applications are not typically tested for explicit support
2935 of just the undeprecated features of either release.
2939 * Add support for setting the minimum and maximum supported protocol.
2940 It can bet set via the SSL_set_min_proto_version() and
2941 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2942 MaxProtocol. It's recommended to use the new APIs to disable
2943 protocols instead of disabling individual protocols using
2944 SSL_set_options() or SSL_CONF's Protocol. This change also
2945 removes support for disabling TLS 1.2 in the OpenSSL TLS
2946 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2950 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2954 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2955 and integrates ECDSA and ECDH functionality into EC. Implementations can
2956 now redirect key generation and no longer need to convert to or from
2959 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2960 include the ec.h header file instead.
2964 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2965 ciphers who are no longer supported and drops support the ephemeral RSA key
2966 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2970 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2971 opaque. For HMAC_CTX, the following constructors and destructors
2974 HMAC_CTX *HMAC_CTX_new(void);
2975 void HMAC_CTX_free(HMAC_CTX *ctx);
2977 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2978 destroy such methods has been added. See EVP_MD_meth_new(3) and
2979 EVP_CIPHER_meth_new(3) for documentation.
2982 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2983 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2984 EVP_MD_CTX_reset() should be called instead to reinitialise
2985 an already created structure.
2986 2) For consistency with the majority of our object creators and
2987 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2988 EVP_MD_CTX_(new|free). The old names are retained as macros
2989 for deprecated builds.
2993 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2994 cryptographic operations to be performed asynchronously as long as an
2995 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2996 further details. Libssl has also had this capability integrated with the
2997 introduction of the new mode SSL_MODE_ASYNC and associated error
2998 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2999 pages. This work was developed in partnership with Intel Corp.
3003 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3004 always enabled now. If you want to disable the support you should
3005 exclude it using the list of supported ciphers. This also means that the
3006 "-no_ecdhe" option has been removed from s_server.
3010 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3011 SSL_{CTX_}set1_curves() which can set a list.
3015 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3016 curve you want to support using SSL_{CTX_}set1_curves().
3020 * State machine rewrite. The state machine code has been significantly
3021 refactored in order to remove much duplication of code and solve issues
3022 with the old code (see ssl/statem/README for further details). This change
3023 does have some associated API changes. Notably the SSL_state() function
3024 has been removed and replaced by SSL_get_state which now returns an
3025 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
3026 altogether. The previous handshake states defined in ssl.h and ssl3.h have
3031 * All instances of the string "ssleay" in the public API were replaced
3032 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3033 Some error codes related to internal RSA_eay API's were renamed.
3037 * The demo files in crypto/threads were moved to demo/threads.
3041 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3044 *Matt Caswell, Rich Salz*
3046 * New ASN.1 embed macro.
3048 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3049 structure is not allocated: it is part of the parent. That is instead of
3057 This reduces memory fragmentation and make it impossible to accidentally
3058 set a mandatory field to NULL.
3060 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3061 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3062 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3067 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3071 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3072 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3073 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3074 DES and RC4 ciphersuites.
3078 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3079 This changes the decoding behaviour for some invalid messages,
3080 though the change is mostly in the more lenient direction, and
3081 legacy behaviour is preserved as much as possible.
3085 * Fix no-stdio build.
3086 * David Woodhouse <David.Woodhouse@intel.com> and also
3087 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3089 * New testing framework
3090 The testing framework has been largely rewritten and is now using
3091 perl and the perl modules Test::Harness and an extended variant of
3092 Test::More called OpenSSL::Test to do its work. All test scripts in
3093 test/ have been rewritten into test recipes, and all direct calls to
3094 executables in test/Makefile have become individual recipes using the
3095 simplified testing OpenSSL::Test::Simple.
3097 For documentation on our testing modules, do:
3099 perldoc test/testlib/OpenSSL/Test/Simple.pm
3100 perldoc test/testlib/OpenSSL/Test.pm
3104 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3105 are used; the latter aborts on memory leaks (usually checked on exit).
3106 Some undocumented "set malloc, etc., hooks" functions were removed
3107 and others were changed. All are now documented.
3111 * In DSA_generate_parameters_ex, if the provided seed is too short,
3114 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3116 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3117 from RFC4279, RFC4785, RFC5487, RFC5489.
3119 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3120 original RSA_PSK patch.
3124 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3125 era flag was never set throughout the codebase (only read). Also removed
3126 SSL3_FLAGS_POP_BUFFER which was only used if
3127 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3131 * Changed the default name options in the "ca", "crl", "req" and "x509"
3132 to be "oneline" instead of "compat".
3136 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3137 not aware of clients that still exhibit this bug, and the workaround
3138 hasn't been working properly for a while.
3142 * The return type of BIO_number_read() and BIO_number_written() as well as
3143 the corresponding num_read and num_write members in the BIO structure has
3144 changed from unsigned long to uint64_t. On platforms where an unsigned
3145 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3150 * Given the pervasive nature of TLS extensions it is inadvisable to run
3151 OpenSSL without support for them. It also means that maintaining
3152 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3153 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3157 * Removed support for the two export grade static DH ciphersuites
3158 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3159 were newly added (along with a number of other static DH ciphersuites) to
3160 1.0.2. However the two export ones have *never* worked since they were
3161 introduced. It seems strange in any case to be adding new export
3162 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3166 * Version negotiation has been rewritten. In particular SSLv23_method(),
3167 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3168 and turned into macros which simply call the new preferred function names
3169 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3170 should use the new names instead. Also as part of this change the ssl23.h
3171 header file has been removed.
3175 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3176 code and the associated standard is no longer considered fit-for-purpose.
3180 * RT2547 was closed. When generating a private key, try to make the
3181 output file readable only by the owner. This behavior change might
3182 be noticeable when interacting with other software.
3184 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3189 * Added HTTP GET support to the ocsp command.
3193 * Changed default digest for the dgst and enc commands from MD5 to
3198 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3202 * Added support for TLS extended master secret from
3203 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3204 initial patch which was a great help during development.
3208 * All libssl internal structures have been removed from the public header
3209 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3210 now redundant). Users should not attempt to access internal structures
3211 directly. Instead they should use the provided API functions.
3215 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3216 Access to deprecated functions can be re-enabled by running config with
3217 "enable-deprecated". In addition applications wishing to use deprecated
3218 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3219 will, by default, disable some transitive includes that previously existed
3220 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3224 * Added support for OCB mode. OpenSSL has been granted a patent license
3225 compatible with the OpenSSL license for use of OCB. Details are available
3226 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3227 for OCB can be removed by calling config with no-ocb.
3231 * SSLv2 support has been removed. It still supports receiving a SSLv2
3232 compatible client hello.
3236 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3237 done while fixing the error code for the key-too-small case.
3239 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3241 * CA.sh has been removed; use CA.pl instead.
3245 * Removed old DES API.
3249 * Remove various unsupported platforms:
3255 Sinix/ReliantUNIX RM400
3260 16-bit platforms such as WIN16
3264 * Clean up OPENSSL_NO_xxx #define's
3265 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3266 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3267 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3268 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3269 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3270 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3271 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3272 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3273 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3274 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3278 * Cleaned up dead code
3279 Remove all but one '#ifdef undef' which is to be looked at.
3283 * Clean up calling of xxx_free routines.
3284 Just like free(), fix most of the xxx_free routines to accept
3285 NULL. Remove the non-null checks from callers. Save much code.
3289 * Add secure heap for storage of private keys (when possible).
3290 Add BIO_s_secmem(), CBIGNUM, etc.
3291 Contributed by Akamai Technologies under our Corporate CLA.
3295 * Experimental support for a new, fast, unbiased prime candidate generator,
3296 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3298 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3300 * New output format NSS in the sess_id command line tool. This allows
3301 exporting the session id and the master key in NSS keylog format.
3303 *Martin Kaiser <martin@kaiser.cx>*
3305 * Harmonize version and its documentation. -f flag is used to display
3308 *mancha <mancha1@zoho.com>*
3310 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3311 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3313 *mancha <mancha1@zoho.com>*
3315 * Fix some double frees. These are not thought to be exploitable.
3317 *mancha <mancha1@zoho.com>*
3319 * A missing bounds check in the handling of the TLS heartbeat extension
3320 can be used to reveal up to 64k of memory to a connected client or
3323 Thanks for Neel Mehta of Google Security for discovering this bug and to
3324 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3325 preparing the fix [CVE-2014-0160][]
3327 *Adam Langley, Bodo Moeller*
3329 * Fix for the attack described in the paper "Recovering OpenSSL
3330 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3331 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3332 <http://eprint.iacr.org/2014/140>
3334 Thanks to Yuval Yarom and Naomi Benger for discovering this
3335 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3337 *Yuval Yarom and Naomi Benger*
3339 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3340 this fixes a limitation in previous versions of OpenSSL.
3344 * Experimental encrypt-then-mac support.
3346 Experimental support for encrypt then mac from
3347 draft-gutmann-tls-encrypt-then-mac-02.txt
3349 To enable it set the appropriate extension number (0x42 for the test
3350 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3352 For non-compliant peers (i.e. just about everything) this should have no
3355 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3359 * Add EVP support for key wrapping algorithms, to avoid problems with
3360 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3361 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3362 algorithms and include tests cases.
3366 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3371 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3372 MGF1 digest and OAEP label.
3376 * Make openssl verify return errors.
3378 *Chris Palmer <palmer@google.com> and Ben Laurie*
3380 * New function ASN1_TIME_diff to calculate the difference between two
3381 ASN1_TIME structures or one structure and the current time.
3385 * Update fips_test_suite to support multiple command line options. New
3386 test to induce all self test errors in sequence and check expected
3391 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3392 sign or verify all in one operation.
3396 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3397 test programs and fips_test_suite. Includes functionality to parse
3398 the minimal script output of fipsalgest.pl directly.
3402 * Add authorisation parameter to FIPS_module_mode_set().
3406 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3410 * Use separate DRBG fields for internal and external flags. New function
3411 FIPS_drbg_health_check() to perform on demand health checking. Add
3412 generation tests to fips_test_suite with reduced health check interval to
3413 demonstrate periodic health checking. Add "nodh" option to
3414 fips_test_suite to skip very slow DH test.
3418 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3423 * More extensive health check for DRBG checking many more failure modes.
3424 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3425 combination: call this in fips_test_suite.
3429 * Add support for canonical generation of DSA parameter 'g'. See
3432 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3433 POST to handle HMAC cases.
3437 * Add functions FIPS_module_version() and FIPS_module_version_text()
3438 to return numerical and string versions of the FIPS module number.
3442 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3443 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3444 outside the validated module in the FIPS capable OpenSSL.
3448 * Minor change to DRBG entropy callback semantics. In some cases
3449 there is no multiple of the block length between min_len and
3450 max_len. Allow the callback to return more than max_len bytes
3451 of entropy but discard any extra: it is the callback's responsibility
3452 to ensure that the extra data discarded does not impact the
3453 requested amount of entropy.
3457 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3458 information in FIPS186-3, SP800-57 and SP800-131A.
3462 * CCM support via EVP. Interface is very similar to GCM case except we
3463 must supply all data in one chunk (i.e. no update, final) and the
3464 message length must be supplied if AAD is used. Add algorithm test
3469 * Initial version of POST overhaul. Add POST callback to allow the status
3470 of POST to be monitored and/or failures induced. Modify fips_test_suite
3471 to use callback. Always run all selftests even if one fails.
3475 * XTS support including algorithm test driver in the fips_gcmtest program.
3476 Note: this does increase the maximum key length from 32 to 64 bytes but
3477 there should be no binary compatibility issues as existing applications
3478 will never use XTS mode.
3482 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3483 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3484 performs algorithm blocking for unapproved PRNG types. Also do not
3485 set PRNG type in FIPS_mode_set(): leave this to the application.
3486 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3487 the standard OpenSSL PRNG: set additional data to a date time vector.
3491 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3492 This shouldn't present any incompatibility problems because applications
3493 shouldn't be using these directly and any that are will need to rethink
3494 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3498 * Extensive self tests and health checking required by SP800-90 DRBG.
3499 Remove strength parameter from FIPS_drbg_instantiate and always
3500 instantiate at maximum supported strength.
3504 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3508 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3512 * New function DH_compute_key_padded() to compute a DH key and pad with
3513 leading zeroes if needed: this complies with SP800-56A et al.
3517 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3518 anything, incomplete, subject to change and largely untested at present.
3522 * Modify fipscanisteronly build option to only build the necessary object
3523 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3527 * Add experimental option FIPSSYMS to give all symbols in
3528 fipscanister.o and FIPS or fips prefix. This will avoid
3529 conflicts with future versions of OpenSSL. Add perl script
3530 util/fipsas.pl to preprocess assembly language source files
3531 and rename any affected symbols.
3535 * Add selftest checks and algorithm block of non-fips algorithms in
3536 FIPS mode. Remove DES2 from selftests.
3540 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3541 return internal method without any ENGINE dependencies. Add new
3542 tiny fips sign and verify functions.
3546 * New build option no-ec2m to disable characteristic 2 code.
3550 * New build option "fipscanisteronly". This only builds fipscanister.o
3551 and (currently) associated fips utilities. Uses the file Makefile.fips
3552 instead of Makefile.org as the prototype.
3556 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3557 Update fips_gcmtest to use IV generator.
3561 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3562 setting output buffer to NULL. The `*Final` function must be
3563 called although it will not retrieve any additional data. The tag
3564 can be set or retrieved with a ctrl. The IV length is by default 12
3565 bytes (96 bits) but can be set to an alternative value. If the IV
3566 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3571 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3572 underlying do_cipher function handles all cipher semantics itself
3573 including padding and finalisation. This is useful if (for example)
3574 an ENGINE cipher handles block padding itself. The behaviour of
3575 do_cipher is subtly changed if this flag is set: the return value
3576 is the number of characters written to the output buffer (zero is
3577 no longer an error code) or a negative error code. Also if the
3578 input buffer is NULL and length 0 finalisation should be performed.
3582 * If a candidate issuer certificate is already part of the constructed
3583 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3587 * Improve forward-security support: add functions
3589 void SSL_CTX_set_not_resumable_session_callback(
3590 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3591 void SSL_set_not_resumable_session_callback(
3592 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3594 for use by SSL/TLS servers; the callback function will be called whenever a
3595 new session is created, and gets to decide whether the session may be
3596 cached to make it resumable (return 0) or not (return 1). (As by the
3597 SSL/TLS protocol specifications, the session_id sent by the server will be
3598 empty to indicate that the session is not resumable; also, the server will
3599 not generate RFC 4507 (RFC 5077) session tickets.)
3601 A simple reasonable callback implementation is to return is_forward_secure.
3602 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3603 by the SSL/TLS server library, indicating whether it can provide forward
3606 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3608 * New -verify_name option in command line utilities to set verification
3613 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3614 Add CMAC pkey methods.
3618 * Experimental renegotiation in s_server -www mode. If the client
3619 browses /reneg connection is renegotiated. If /renegcert it is
3620 renegotiated requesting a certificate.
3624 * Add an "external" session cache for debugging purposes to s_server. This
3625 should help trace issues which normally are only apparent in deployed
3626 multi-process servers.
3630 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3631 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3632 BIO_set_cipher() and some obscure PEM functions were changed so they
3633 can now return an error. The RAND changes required a change to the
3634 RAND_METHOD structure.
3638 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3639 a gcc attribute to warn if the result of a function is ignored. This
3640 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3641 whose return value is often ignored.
3645 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3646 These allow SCTs (signed certificate timestamps) to be requested and
3647 validated when establishing a connection.
3649 *Rob Percival <robpercival@google.com>*
3654 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3656 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3657 used even when parsing explicit parameters, when loading a serialized key
3658 or calling `EC_GROUP_new_from_ecpkparameters()`/
3659 `EC_GROUP_new_from_ecparameters()`.
3660 This prevents bypass of security hardening and performance gains,
3661 especially for curves with specialized EC_METHODs.
3662 By default, if a key encoded with explicit parameters is loaded and later
3663 serialized, the output is still encoded with explicit parameters, even if
3664 internally a "named" EC_GROUP is used for computation.
3668 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3669 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3670 NULL. After this change, only the cofactor parameter can be NULL. It also
3671 does some minimal sanity checks on the passed order.
3676 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3677 An attack is simple, if the first CMS_recipientInfo is valid but the
3678 second CMS_recipientInfo is chosen ciphertext. If the second
3679 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3680 encryption key will be replaced by garbage, and the message cannot be
3681 decoded, but if the RSA decryption fails, the correct encryption key is
3682 used and the recipient will not notice the attack.
3683 As a work around for this potential attack the length of the decrypted
3684 key must be equal to the cipher default key length, in case the
3685 certifiate is not given and all recipientInfo are tried out.
3686 The old behaviour can be re-enabled in the CMS code by setting the
3687 CMS_DEBUG_DECRYPT flag.
3692 * Document issue with installation paths in diverse Windows builds
3694 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3695 binaries and run-time config file.
3700 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3702 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3703 This changes the size when using the genpkey app when no size is given. It
3704 fixes an omission in earlier changes that changed all RSA, DSA and DH
3705 generation apps to use 2048 bits by default.
3709 * Add FIPS support for Android Arm 64-bit
3711 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3712 Module in Version 2.0.10. For some reason, the corresponding target
3713 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3714 built with FIPS support on Android Arm 64-bit. This omission has been
3717 *Matthias St. Pierre*
3719 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3721 * 0-byte record padding oracle
3723 If an application encounters a fatal protocol error and then calls
3724 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3725 then OpenSSL can respond differently to the calling application if a 0 byte
3726 record is received with invalid padding compared to if a 0 byte record is
3727 received with an invalid MAC. If the application then behaves differently
3728 based on that in a way that is detectable to the remote peer, then this
3729 amounts to a padding oracle that could be used to decrypt data.
3731 In order for this to be exploitable "non-stitched" ciphersuites must be in
3732 use. Stitched ciphersuites are optimised implementations of certain
3733 commonly used ciphersuites. Also the application must call SSL_shutdown()
3734 twice even if a protocol error has occurred (applications should not do
3735 this but some do anyway).
3737 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3738 Aviram, with additional investigation by Steven Collison and Andrew
3739 Hourselt. It was reported to OpenSSL on 10th December 2018.
3744 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3748 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3750 * Microarchitecture timing vulnerability in ECC scalar multiplication
3752 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3753 shown to be vulnerable to a microarchitecture timing side channel attack.
3754 An attacker with sufficient access to mount local timing attacks during
3755 ECDSA signature generation could recover the private key.
3757 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3758 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3764 * Timing vulnerability in DSA signature generation
3766 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3767 timing side channel attack. An attacker could use variations in the signing
3768 algorithm to recover the private key.
3770 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3775 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3776 Module, accidentally introduced while backporting security fixes from the
3777 development branch and hindering the use of ECC in FIPS mode.
3781 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3783 * Client DoS due to large DH parameter
3785 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3786 malicious server can send a very large prime value to the client. This will
3787 cause the client to spend an unreasonably long period of time generating a
3788 key for this prime resulting in a hang until the client has finished. This
3789 could be exploited in a Denial Of Service attack.
3791 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3796 * Cache timing vulnerability in RSA Key Generation
3798 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3799 a cache timing side channel attack. An attacker with sufficient access to
3800 mount cache timing attacks during the RSA key generation process could
3801 recover the private key.
3803 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3804 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3809 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3810 parameter is no longer accepted, as it leads to a corrupt table. NULL
3811 pem_str is reserved for alias entries only.
3815 * Revert blinding in ECDSA sign and instead make problematic addition
3816 length-invariant. Switch even to fixed-length Montgomery multiplication.
3820 * Change generating and checking of primes so that the error rate of not
3821 being prime depends on the intended use based on the size of the input.
3822 For larger primes this will result in more rounds of Miller-Rabin.
3823 The maximal error rate for primes with more than 1080 bits is lowered
3826 *Kurt Roeckx, Annie Yousar*
3828 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3832 * Add blinding to ECDSA and DSA signatures to protect against side channel
3833 attacks discovered by Keegan Ryan (NCC Group).
3837 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3838 now allow empty (zero character) pass phrases.
3842 * Certificate time validation (X509_cmp_time) enforces stricter
3843 compliance with RFC 5280. Fractional seconds and timezone offsets
3844 are no longer allowed.
3848 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3850 * Constructed ASN.1 types with a recursive definition could exceed the stack
3852 Constructed ASN.1 types with a recursive definition (such as can be found
3853 in PKCS7) could eventually exceed the stack given malicious input with
3854 excessive recursion. This could result in a Denial Of Service attack. There
3855 are no such structures used within SSL/TLS that come from untrusted sources
3856 so this is considered safe.
3858 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3864 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3866 * Read/write after SSL object in error state
3868 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3869 mechanism. The intent was that if a fatal error occurred during a handshake
3870 then OpenSSL would move into the error state and would immediately fail if
3871 you attempted to continue the handshake. This works as designed for the
3872 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3873 SSL_connect()), however due to a bug it does not work correctly if
3874 SSL_read() or SSL_write() is called directly. In that scenario, if the
3875 handshake fails then a fatal error will be returned in the initial function
3876 call. If SSL_read()/SSL_write() is subsequently called by the application
3877 for the same SSL object then it will succeed and the data is passed without
3878 being decrypted/encrypted directly from the SSL/TLS record layer.
3880 In order to exploit this issue an application bug would have to be present
3881 that resulted in a call to SSL_read()/SSL_write() being issued after having
3882 already received a fatal error.
3884 This issue was reported to OpenSSL by David Benjamin (Google).
3889 * rsaz_1024_mul_avx2 overflow bug on x86_64
3891 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3892 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3893 Analysis suggests that attacks against RSA and DSA as a result of this
3894 defect would be very difficult to perform and are not believed likely.
3895 Attacks against DH1024 are considered just feasible, because most of the
3896 work necessary to deduce information about a private key may be performed
3897 offline. The amount of resources required for such an attack would be
3898 significant. However, for an attack on TLS to be meaningful, the server
3899 would have to share the DH1024 private key among multiple clients, which is
3900 no longer an option since CVE-2016-0701.
3902 This only affects processors that support the AVX2 but not ADX extensions
3903 like Intel Haswell (4th generation).
3905 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3906 was originally found via the OSS-Fuzz project.
3911 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
3913 * bn_sqrx8x_internal carry bug on x86_64
3915 There is a carry propagating bug in the x86_64 Montgomery squaring
3916 procedure. No EC algorithms are affected. Analysis suggests that attacks
3917 against RSA and DSA as a result of this defect would be very difficult to
3918 perform and are not believed likely. Attacks against DH are considered just
3919 feasible (although very difficult) because most of the work necessary to
3920 deduce information about a private key may be performed offline. The amount
3921 of resources required for such an attack would be very significant and
3922 likely only accessible to a limited number of attackers. An attacker would
3923 additionally need online access to an unpatched system using the target
3924 private key in a scenario with persistent DH parameters and a private
3925 key that is shared between multiple clients.
3927 This only affects processors that support the BMI1, BMI2 and ADX extensions
3928 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3930 This issue was reported to OpenSSL by the OSS-Fuzz project.
3935 * Malformed X.509 IPAddressFamily could cause OOB read
3937 If an X.509 certificate has a malformed IPAddressFamily extension,
3938 OpenSSL could do a one-byte buffer overread. The most likely result
3939 would be an erroneous display of the certificate in text format.
3941 This issue was reported to OpenSSL by the OSS-Fuzz project.
3946 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
3948 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3949 platform rather than 'mingw'.
3953 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
3955 * Truncated packet could crash via OOB read
3957 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3958 cipher is being used, then a truncated packet can cause that host to
3959 perform an out-of-bounds read, usually resulting in a crash.
3961 This issue was reported to OpenSSL by Robert Święcki of Google.
3966 * BN_mod_exp may produce incorrect results on x86_64
3968 There is a carry propagating bug in the x86_64 Montgomery squaring
3969 procedure. No EC algorithms are affected. Analysis suggests that attacks
3970 against RSA and DSA as a result of this defect would be very difficult to
3971 perform and are not believed likely. Attacks against DH are considered just
3972 feasible (although very difficult) because most of the work necessary to
3973 deduce information about a private key may be performed offline. The amount
3974 of resources required for such an attack would be very significant and
3975 likely only accessible to a limited number of attackers. An attacker would
3976 additionally need online access to an unpatched system using the target
3977 private key in a scenario with persistent DH parameters and a private
3978 key that is shared between multiple clients. For example this can occur by
3979 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3980 similar to CVE-2015-3193 but must be treated as a separate problem.
3982 This issue was reported to OpenSSL by the OSS-Fuzz project.
3987 * Montgomery multiplication may produce incorrect results
3989 There is a carry propagating bug in the Broadwell-specific Montgomery
3990 multiplication procedure that handles input lengths divisible by, but
3991 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3992 and DH private keys are impossible. This is because the subroutine in
3993 question is not used in operations with the private key itself and an input
3994 of the attacker's direct choice. Otherwise the bug can manifest itself as
3995 transient authentication and key negotiation failures or reproducible
3996 erroneous outcome of public-key operations with specially crafted input.
3997 Among EC algorithms only Brainpool P-512 curves are affected and one
3998 presumably can attack ECDH key negotiation. Impact was not analyzed in
3999 detail, because pre-requisites for attack are considered unlikely. Namely
4000 multiple clients have to choose the curve in question and the server has to
4001 share the private key among them, neither of which is default behaviour.
4002 Even then only clients that chose the curve will be affected.
4004 This issue was publicly reported as transient failures and was not
4005 initially recognized as a security issue. Thanks to Richard Morgan for
4006 providing reproducible case.
4011 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4012 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4013 prevent issues where no progress is being made and the peer continually
4014 sends unrecognised record types, using up resources processing them.
4018 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4020 * Missing CRL sanity check
4022 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4023 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4024 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4026 This issue only affects the OpenSSL 1.0.2i
4031 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4033 * OCSP Status Request extension unbounded memory growth
4035 A malicious client can send an excessively large OCSP Status Request
4036 extension. If that client continually requests renegotiation, sending a
4037 large OCSP Status Request extension each time, then there will be unbounded
4038 memory growth on the server. This will eventually lead to a Denial Of
4039 Service attack through memory exhaustion. Servers with a default
4040 configuration are vulnerable even if they do not support OCSP. Builds using
4041 the "no-ocsp" build time option are not affected.
4043 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4048 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4051 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4057 * OOB write in MDC2_Update()
4059 An overflow can occur in MDC2_Update() either if called directly or
4060 through the EVP_DigestUpdate() function using MDC2. If an attacker
4061 is able to supply very large amounts of input data after a previous
4062 call to EVP_EncryptUpdate() with a partial block then a length check
4063 can overflow resulting in a heap corruption.
4065 The amount of data needed is comparable to SIZE_MAX which is impractical
4068 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4073 * Malformed SHA512 ticket DoS
4075 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4076 DoS attack where a malformed ticket will result in an OOB read which will
4079 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4080 a custom server callback and ticket lookup mechanism.
4082 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4087 * OOB write in BN_bn2dec()
4089 The function BN_bn2dec() does not check the return value of BN_div_word().
4090 This can cause an OOB write if an application uses this function with an
4091 overly large BIGNUM. This could be a problem if an overly large certificate
4092 or CRL is printed out from an untrusted source. TLS is not affected because
4093 record limits will reject an oversized certificate before it is parsed.
4095 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4100 * OOB read in TS_OBJ_print_bio()
4102 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4103 the total length the OID text representation would use and not the amount
4104 of data written. This will result in OOB reads when large OIDs are
4107 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4112 * Pointer arithmetic undefined behaviour
4114 Avoid some undefined pointer arithmetic
4116 A common idiom in the codebase is to check limits in the following manner:
4119 Where "p" points to some malloc'd data of SIZE bytes and
4122 "len" here could be from some externally supplied data (e.g. from a TLS
4125 The rules of C pointer arithmetic are such that "p + len" is only well
4126 defined where len <= SIZE. Therefore the above idiom is actually
4127 undefined behaviour.
4129 For example this could cause problems if some malloc implementation
4130 provides an address for "p" such that "p + len" actually overflows for
4131 values of len that are too big and therefore p + len < limit.
4133 This issue was reported to OpenSSL by Guido Vranken
4138 * Constant time flag not preserved in DSA signing
4140 Operations in the DSA signing algorithm should run in constant time in
4141 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4142 implementation means that a non-constant time codepath is followed for
4143 certain operations. This has been demonstrated through a cache-timing
4144 attack to be sufficient for an attacker to recover the private DSA key.
4146 This issue was reported by César Pereida (Aalto University), Billy Brumley
4147 (Tampere University of Technology), and Yuval Yarom (The University of
4148 Adelaide and NICTA).
4153 * DTLS buffered message DoS
4155 In a DTLS connection where handshake messages are delivered out-of-order
4156 those messages that OpenSSL is not yet ready to process will be buffered
4157 for later use. Under certain circumstances, a flaw in the logic means that
4158 those messages do not get removed from the buffer even though the handshake
4159 has been completed. An attacker could force up to approx. 15 messages to
4160 remain in the buffer when they are no longer required. These messages will
4161 be cleared when the DTLS connection is closed. The default maximum size for
4162 a message is 100k. Therefore the attacker could force an additional 1500k
4163 to be consumed per connection. By opening many simulataneous connections an
4164 attacker could cause a DoS attack through memory exhaustion.
4166 This issue was reported to OpenSSL by Quan Luo.
4171 * DTLS replay protection DoS
4173 A flaw in the DTLS replay attack protection mechanism means that records
4174 that arrive for future epochs update the replay protection "window" before
4175 the MAC for the record has been validated. This could be exploited by an
4176 attacker by sending a record for the next epoch (which does not have to
4177 decrypt or have a valid MAC), with a very large sequence number. This means
4178 that all subsequent legitimate packets are dropped causing a denial of
4179 service for a specific DTLS connection.
4181 This issue was reported to OpenSSL by the OCAP audit team.
4186 * Certificate message OOB reads
4188 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4189 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4190 theoretical DoS risk but this has not been observed in practice on common
4193 The messages affected are client certificate, client certificate request
4194 and server certificate. As a result the attack can only be performed
4195 against a client or a server which enables client authentication.
4197 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4202 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4204 * Prevent padding oracle in AES-NI CBC MAC check
4206 A MITM attacker can use a padding oracle attack to decrypt traffic
4207 when the connection uses an AES CBC cipher and the server support
4210 This issue was introduced as part of the fix for Lucky 13 padding
4211 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4212 constant time by making sure that always the same bytes are read and
4213 compared against either the MAC or padding bytes. But it no longer
4214 checked that there was enough data to have both the MAC and padding
4217 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4222 * Fix EVP_EncodeUpdate overflow
4224 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4225 Base64 encoding of binary data. If an attacker is able to supply very large
4226 amounts of input data then a length check can overflow resulting in a heap
4229 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4230 the `PEM_write_bio*` family of functions. These are mainly used within the
4231 OpenSSL command line applications, so any application which processes data
4232 from an untrusted source and outputs it as a PEM file should be considered
4233 vulnerable to this issue. User applications that call these APIs directly
4234 with large amounts of untrusted data may also be vulnerable.
4236 This issue was reported by Guido Vranken.
4241 * Fix EVP_EncryptUpdate overflow
4243 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4244 is able to supply very large amounts of input data after a previous call to
4245 EVP_EncryptUpdate() with a partial block then a length check can overflow
4246 resulting in a heap corruption. Following an analysis of all OpenSSL
4247 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4248 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4249 the first called function after an EVP_EncryptInit(), and therefore that
4250 specific call must be safe. The second form is where the length passed to
4251 EVP_EncryptUpdate() can be seen from the code to be some small value and
4252 therefore there is no possibility of an overflow. Since all instances are
4253 one of these two forms, it is believed that there can be no overflows in
4254 internal code due to this problem. It should be noted that
4255 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4256 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4257 of these calls have also been analysed too and it is believed there are no
4258 instances in internal usage where an overflow could occur.
4260 This issue was reported by Guido Vranken.
4265 * Prevent ASN.1 BIO excessive memory allocation
4267 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4268 a short invalid encoding can cause allocation of large amounts of memory
4269 potentially consuming excessive resources or exhausting memory.
4271 Any application parsing untrusted data through d2i BIO functions is
4272 affected. The memory based functions such as d2i_X509() are *not* affected.
4273 Since the memory based functions are used by the TLS library, TLS
4274 applications are not affected.
4276 This issue was reported by Brian Carpenter.
4283 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4284 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4285 in arbitrary stack data being returned in the buffer.
4287 This issue was reported by Guido Vranken.
4292 * Modify behavior of ALPN to invoke callback after SNI/servername
4293 callback, such that updates to the SSL_CTX affect ALPN.
4297 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4302 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4303 methods are enabled and ssl2 is disabled the methods return NULL.
4307 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4309 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4310 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4311 provide any "EXPORT" or "LOW" strength ciphers.
4315 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4316 is by default disabled at build-time. Builds that are not configured with
4317 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4318 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4319 will need to explicitly call either of:
4321 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4323 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4325 as appropriate. Even if either of those is used, or the application
4326 explicitly uses the version-specific SSLv2_method() or its client and
4327 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4328 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4329 ciphers, and SSLv2 56-bit DES are no longer available.
4334 * Fix a double-free in DSA code
4336 A double free bug was discovered when OpenSSL parses malformed DSA private
4337 keys and could lead to a DoS attack or memory corruption for applications
4338 that receive DSA private keys from untrusted sources. This scenario is
4341 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4347 * Disable SRP fake user seed to address a server memory leak.
4349 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4351 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4352 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4353 was changed to ignore the "fake user" SRP seed, even if the seed
4356 Users should use SRP_VBASE_get1_by_user instead. Note that in
4357 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4358 also that even though configuring the SRP seed attempts to hide
4359 invalid usernames by continuing the handshake with fake
4360 credentials, this behaviour is not constant time and no strong
4361 guarantees are made that the handshake is indistinguishable from
4362 that of a valid user.
4367 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4369 In the BN_hex2bn function the number of hex digits is calculated using an
4370 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4371 large values of |i| this can result in |bn_expand| not allocating any
4372 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4373 field as NULL leading to a subsequent NULL ptr deref. For very large values
4374 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4375 In this case memory is allocated to the internal BIGNUM data field, but it
4376 is insufficiently sized leading to heap corruption. A similar issue exists
4377 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4378 is ever called by user applications with very large untrusted hex/dec data.
4379 This is anticipated to be a rare occurrence.
4381 All OpenSSL internal usage of these functions use data that is not expected
4382 to be untrusted, e.g. config file data or application command line
4383 arguments. If user developed applications generate config file data based
4384 on untrusted data then it is possible that this could also lead to security
4385 consequences. This is also anticipated to be rare.
4387 This issue was reported to OpenSSL by Guido Vranken.
4392 * Fix memory issues in `BIO_*printf` functions
4394 The internal |fmtstr| function used in processing a "%s" format string in
4395 the `BIO_*printf` functions could overflow while calculating the length of a
4396 string and cause an OOB read when printing very long strings.
4398 Additionally the internal |doapr_outch| function can attempt to write to an
4399 OOB memory location (at an offset from the NULL pointer) in the event of a
4400 memory allocation failure. In 1.0.2 and below this could be caused where
4401 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4402 could be in processing a very long "%s" format string. Memory leaks can
4405 The first issue may mask the second issue dependent on compiler behaviour.
4406 These problems could enable attacks where large amounts of untrusted data
4407 is passed to the `BIO_*printf` functions. If applications use these functions
4408 in this way then they could be vulnerable. OpenSSL itself uses these
4409 functions when printing out human-readable dumps of ASN.1 data. Therefore
4410 applications that print this data could be vulnerable if the data is from
4411 untrusted sources. OpenSSL command line applications could also be
4412 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4413 as command line arguments.
4415 Libssl is not considered directly vulnerable. Additionally certificates etc
4416 received via remote connections via libssl are also unlikely to be able to
4417 trigger these issues because of message size limits enforced within libssl.
4419 This issue was reported to OpenSSL Guido Vranken.
4424 * Side channel attack on modular exponentiation
4426 A side-channel attack was found which makes use of cache-bank conflicts on
4427 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4428 of RSA keys. The ability to exploit this issue is limited as it relies on
4429 an attacker who has control of code in a thread running on the same
4430 hyper-threaded core as the victim thread which is performing decryptions.
4432 This issue was reported to OpenSSL by Yuval Yarom, The University of
4433 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4434 Nadia Heninger, University of Pennsylvania with more information at
4435 <http://cachebleed.info>.
4440 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4441 if no keysize is specified with default_bits. This fixes an
4442 omission in an earlier change that changed all RSA/DSA key generation
4443 apps to use 2048 bits by default.
4447 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4449 * DH small subgroups
4451 Historically OpenSSL only ever generated DH parameters based on "safe"
4452 primes. More recently (in version 1.0.2) support was provided for
4453 generating X9.42 style parameter files such as those required for RFC 5114
4454 support. The primes used in such files may not be "safe". Where an
4455 application is using DH configured with parameters based on primes that are
4456 not "safe" then an attacker could use this fact to find a peer's private
4457 DH exponent. This attack requires that the attacker complete multiple
4458 handshakes in which the peer uses the same private DH exponent. For example
4459 this could be used to discover a TLS server's private DH exponent if it's
4460 reusing the private DH exponent or it's using a static DH ciphersuite.
4462 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4463 TLS. It is not on by default. If the option is not set then the server
4464 reuses the same private DH exponent for the life of the server process and
4465 would be vulnerable to this attack. It is believed that many popular
4466 applications do set this option and would therefore not be at risk.
4468 The fix for this issue adds an additional check where a "q" parameter is
4469 available (as is the case in X9.42 based parameters). This detects the
4470 only known attack, and is the only possible defense for static DH
4471 ciphersuites. This could have some performance impact.
4473 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4474 default and cannot be disabled. This could have some performance impact.
4476 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4481 * SSLv2 doesn't block disabled ciphers
4483 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4484 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4485 been disabled, provided that the SSLv2 protocol was not also disabled via
4488 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4489 and Sebastian Schinzel.
4494 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4496 * BN_mod_exp may produce incorrect results on x86_64
4498 There is a carry propagating bug in the x86_64 Montgomery squaring
4499 procedure. No EC algorithms are affected. Analysis suggests that attacks
4500 against RSA and DSA as a result of this defect would be very difficult to
4501 perform and are not believed likely. Attacks against DH are considered just
4502 feasible (although very difficult) because most of the work necessary to
4503 deduce information about a private key may be performed offline. The amount
4504 of resources required for such an attack would be very significant and
4505 likely only accessible to a limited number of attackers. An attacker would
4506 additionally need online access to an unpatched system using the target
4507 private key in a scenario with persistent DH parameters and a private
4508 key that is shared between multiple clients. For example this can occur by
4509 default in OpenSSL DHE based SSL/TLS ciphersuites.
4511 This issue was reported to OpenSSL by Hanno Böck.
4516 * Certificate verify crash with missing PSS parameter
4518 The signature verification routines will crash with a NULL pointer
4519 dereference if presented with an ASN.1 signature using the RSA PSS
4520 algorithm and absent mask generation function parameter. Since these
4521 routines are used to verify certificate signature algorithms this can be
4522 used to crash any certificate verification operation and exploited in a
4523 DoS attack. Any application which performs certificate verification is
4524 vulnerable including OpenSSL clients and servers which enable client
4527 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4532 * X509_ATTRIBUTE memory leak
4534 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4535 memory. This structure is used by the PKCS#7 and CMS routines so any
4536 application which reads PKCS#7 or CMS data from untrusted sources is
4537 affected. SSL/TLS is not affected.
4539 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4545 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4546 This changes the decoding behaviour for some invalid messages,
4547 though the change is mostly in the more lenient direction, and
4548 legacy behaviour is preserved as much as possible.
4552 * In DSA_generate_parameters_ex, if the provided seed is too short,
4555 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4557 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4559 * Alternate chains certificate forgery
4561 During certificate verification, OpenSSL will attempt to find an
4562 alternative certificate chain if the first attempt to build such a chain
4563 fails. An error in the implementation of this logic can mean that an
4564 attacker could cause certain checks on untrusted certificates to be
4565 bypassed, such as the CA flag, enabling them to use a valid leaf
4566 certificate to act as a CA and "issue" an invalid certificate.
4568 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4573 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4575 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4576 incompatibility in the handling of HMAC. The previous ABI has now been
4581 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4583 * Malformed ECParameters causes infinite loop
4585 When processing an ECParameters structure OpenSSL enters an infinite loop
4586 if the curve specified is over a specially malformed binary polynomial
4589 This can be used to perform denial of service against any
4590 system which processes public keys, certificate requests or
4591 certificates. This includes TLS clients and TLS servers with
4592 client authentication enabled.
4594 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4599 * Exploitable out-of-bounds read in X509_cmp_time
4601 X509_cmp_time does not properly check the length of the ASN1_TIME
4602 string and can read a few bytes out of bounds. In addition,
4603 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4606 An attacker can use this to craft malformed certificates and CRLs of
4607 various sizes and potentially cause a segmentation fault, resulting in
4608 a DoS on applications that verify certificates or CRLs. TLS clients
4609 that verify CRLs are affected. TLS clients and servers with client
4610 authentication enabled may be affected if they use custom verification
4613 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4614 independently by Hanno Böck.
4619 * PKCS7 crash with missing EnvelopedContent
4621 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4622 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4623 with missing content and trigger a NULL pointer dereference on parsing.
4625 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4626 structures from untrusted sources are affected. OpenSSL clients and
4627 servers are not affected.
4629 This issue was reported to OpenSSL by Michal Zalewski (Google).
4634 * CMS verify infinite loop with unknown hash function
4636 When verifying a signedData message the CMS code can enter an infinite loop
4637 if presented with an unknown hash function OID. This can be used to perform
4638 denial of service against any system which verifies signedData messages using
4640 This issue was reported to OpenSSL by Johannes Bauer.
4645 * Race condition handling NewSessionTicket
4647 If a NewSessionTicket is received by a multi-threaded client when attempting to
4648 reuse a previous ticket then a race condition can occur potentially leading to
4649 a double free of the ticket data.
4654 * Only support 256-bit or stronger elliptic curves with the
4655 'ecdh_auto' setting (server) or by default (client). Of supported
4656 curves, prefer P-256 (both).
4660 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4662 * ClientHello sigalgs DoS fix
4664 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4665 invalid signature algorithms extension a NULL pointer dereference will
4666 occur. This can be exploited in a DoS attack against the server.
4668 This issue was was reported to OpenSSL by David Ramos of Stanford
4672 *Stephen Henson and Matt Caswell*
4674 * Multiblock corrupted pointer fix
4676 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4677 feature only applies on 64 bit x86 architecture platforms that support AES
4678 NI instructions. A defect in the implementation of "multiblock" can cause
4679 OpenSSL's internal write buffer to become incorrectly set to NULL when
4680 using non-blocking IO. Typically, when the user application is using a
4681 socket BIO for writing, this will only result in a failed connection.
4682 However if some other BIO is used then it is likely that a segmentation
4683 fault will be triggered, thus enabling a potential DoS attack.
4685 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4690 * Segmentation fault in DTLSv1_listen fix
4692 The DTLSv1_listen function is intended to be stateless and processes the
4693 initial ClientHello from many peers. It is common for user code to loop
4694 over the call to DTLSv1_listen until a valid ClientHello is received with
4695 an associated cookie. A defect in the implementation of DTLSv1_listen means
4696 that state is preserved in the SSL object from one invocation to the next
4697 that can lead to a segmentation fault. Errors processing the initial
4698 ClientHello can trigger this scenario. An example of such an error could be
4699 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4702 This issue was reported to OpenSSL by Per Allansson.
4707 * Segmentation fault in ASN1_TYPE_cmp fix
4709 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4710 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4711 certificate signature algorithm consistency this can be used to crash any
4712 certificate verification operation and exploited in a DoS attack. Any
4713 application which performs certificate verification is vulnerable including
4714 OpenSSL clients and servers which enable client authentication.
4719 * Segmentation fault for invalid PSS parameters fix
4721 The signature verification routines will crash with a NULL pointer
4722 dereference if presented with an ASN.1 signature using the RSA PSS
4723 algorithm and invalid parameters. Since these routines are used to verify
4724 certificate signature algorithms this can be used to crash any
4725 certificate verification operation and exploited in a DoS attack. Any
4726 application which performs certificate verification is vulnerable including
4727 OpenSSL clients and servers which enable client authentication.
4729 This issue was was reported to OpenSSL by Brian Carpenter.
4734 * ASN.1 structure reuse memory corruption fix
4736 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4737 memory corruption via an invalid write. Such reuse is and has been
4738 strongly discouraged and is believed to be rare.
4740 Applications that parse structures containing CHOICE or ANY DEFINED BY
4741 components may be affected. Certificate parsing (d2i_X509 and related
4742 functions) are however not affected. OpenSSL clients and servers are
4748 * PKCS7 NULL pointer dereferences fix
4750 The PKCS#7 parsing code does not handle missing outer ContentInfo
4751 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4752 missing content and trigger a NULL pointer dereference on parsing.
4754 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4755 otherwise parse PKCS#7 structures from untrusted sources are
4756 affected. OpenSSL clients and servers are not affected.
4758 This issue was reported to OpenSSL by Michal Zalewski (Google).
4763 * DoS via reachable assert in SSLv2 servers fix
4765 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4766 servers that both support SSLv2 and enable export cipher suites by sending
4767 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4769 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4770 (OpenSSL development team).
4775 * Empty CKE with client auth and DHE fix
4777 If client auth is used then a server can seg fault in the event of a DHE
4778 ciphersuite being selected and a zero length ClientKeyExchange message
4779 being sent by the client. This could be exploited in a DoS attack.
4784 * Handshake with unseeded PRNG fix
4786 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4787 with an unseeded PRNG. The conditions are:
4788 - The client is on a platform where the PRNG has not been seeded
4789 automatically, and the user has not seeded manually
4790 - A protocol specific client method version has been used (i.e. not
4791 SSL_client_methodv23)
4792 - A ciphersuite is used that does not require additional random data from
4793 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4795 If the handshake succeeds then the client random that has been used will
4796 have been generated from a PRNG with insufficient entropy and therefore the
4797 output may be predictable.
4799 For example using the following command with an unseeded openssl will
4800 succeed on an unpatched platform:
4802 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4807 * Use After Free following d2i_ECPrivatekey error fix
4809 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4810 could cause a use after free condition. This, in turn, could cause a double
4811 free in several private key parsing functions (such as d2i_PrivateKey
4812 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4813 for applications that receive EC private keys from untrusted
4814 sources. This scenario is considered rare.
4816 This issue was discovered by the BoringSSL project and fixed in their
4822 * X509_to_X509_REQ NULL pointer deref fix
4824 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4825 the certificate key is invalid. This function is rarely used in practice.
4827 This issue was discovered by Brian Carpenter.
4832 * Removed the export ciphers from the DEFAULT ciphers
4836 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4838 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4839 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4840 So far those who have to target multiple platforms would compromise
4841 and argue that binary targeting say ARMv5 would still execute on
4842 ARMv8. "Universal" build resolves this compromise by providing
4843 near-optimal performance even on newer platforms.
4847 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4848 (other platforms pending).
4850 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4852 * Add support for the SignedCertificateTimestampList certificate and
4853 OCSP response extensions from RFC6962.
4857 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4858 for corner cases. (Certain input points at infinity could lead to
4859 bogus results, with non-infinity inputs mapped to infinity too.)
4863 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4864 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4865 common cases are optimized and there still is room for further
4866 improvements. Vector Permutation AES for Altivec is also added.
4870 * Add support for little-endian ppc64 Linux target.
4872 *Marcelo Cerri (IBM)*
4874 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4875 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4876 are optimized and there still is room for further improvements.
4877 Both 32- and 64-bit modes are supported.
4879 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4881 * Improved ARMv7 NEON support.
4885 * Support for SPARC Architecture 2011 crypto extensions, first
4886 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4887 SHA256/512, MD5, GHASH and modular exponentiation.
4889 *Andy Polyakov, David Miller*
4891 * Accelerated modular exponentiation for Intel processors, a.k.a.
4894 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4896 * Support for new and upcoming Intel processors, including AVX2,
4897 BMI and SHA ISA extensions. This includes additional "stitched"
4898 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4901 This work was sponsored by Intel Corp.
4905 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4906 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4907 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4911 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4912 this fixes a limitation in previous versions of OpenSSL.
4916 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4917 MGF1 digest and OAEP label.
4921 * Add EVP support for key wrapping algorithms, to avoid problems with
4922 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4923 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4924 algorithms and include tests cases.
4928 * Add functions to allocate and set the fields of an ECDSA_METHOD
4931 *Douglas E. Engert, Steve Henson*
4933 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4934 difference in days and seconds between two tm or ASN1_TIME structures.
4938 * Add -rev test option to s_server to just reverse order of characters
4939 received by client and send back to server. Also prints an abbreviated
4940 summary of the connection parameters.
4944 * New option -brief for s_client and s_server to print out a brief summary
4945 of connection parameters.
4949 * Add callbacks for arbitrary TLS extensions.
4951 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4953 * New option -crl_download in several openssl utilities to download CRLs
4954 from CRLDP extension in certificates.
4958 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4962 * New function X509_CRL_diff to generate a delta CRL from the difference
4963 of two full CRLs. Add support to "crl" utility.
4967 * New functions to set lookup_crls function and to retrieve
4968 X509_STORE from X509_STORE_CTX.
4972 * Print out deprecated issuer and subject unique ID fields in
4977 * Extend OCSP I/O functions so they can be used for simple general purpose
4978 HTTP as well as OCSP. New wrapper function which can be used to download
4979 CRLs using the OCSP API.
4983 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4987 * `SSL_CONF*` functions. These provide a common framework for application
4988 configuration using configuration files or command lines.
4992 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4993 message callback and prints the results. Needs compile time option
4994 "enable-ssl-trace". New options to s_client and s_server to enable
4999 * New ctrl and macro to retrieve supported points extensions.
5000 Print out extension in s_server and s_client.
5004 * New functions to retrieve certificate signature and signature
5009 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5014 * New Suite B modes for TLS code. These use and enforce the requirements
5015 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5016 only use Suite B curves. The Suite B modes can be set by using the
5017 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5021 * New chain verification flags for Suite B levels of security. Check
5022 algorithms are acceptable when flags are set in X509_verify_cert.
5026 * Make tls1_check_chain return a set of flags indicating checks passed
5027 by a certificate chain. Add additional tests to handle client
5028 certificates: checks for matching certificate type and issuer name
5033 * If an attempt is made to use a signature algorithm not in the peer
5034 preference list abort the handshake. If client has no suitable
5035 signature algorithms in response to a certificate request do not
5036 use the certificate.
5040 * If server EC tmp key is not in client preference list abort handshake.
5044 * Add support for certificate stores in CERT structure. This makes it
5045 possible to have different stores per SSL structure or one store in
5046 the parent SSL_CTX. Include distinct stores for certificate chain
5047 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5048 to build and store a certificate chain in CERT structure: returning
5049 an error if the chain cannot be built: this will allow applications
5050 to test if a chain is correctly configured.
5052 Note: if the CERT based stores are not set then the parent SSL_CTX
5053 store is used to retain compatibility with existing behaviour.
5057 * New function ssl_set_client_disabled to set a ciphersuite disabled
5058 mask based on the current session, check mask when sending client
5059 hello and checking the requested ciphersuite.
5063 * New ctrls to retrieve and set certificate types in a certificate
5064 request message. Print out received values in s_client. If certificate
5065 types is not set with custom values set sensible values based on
5066 supported signature algorithms.
5070 * Support for distinct client and server supported signature algorithms.
5074 * Add certificate callback. If set this is called whenever a certificate
5075 is required by client or server. An application can decide which
5076 certificate chain to present based on arbitrary criteria: for example
5077 supported signature algorithms. Add very simple example to s_server.
5078 This fixes many of the problems and restrictions of the existing client
5079 certificate callback: for example you can now clear an existing
5080 certificate and specify the whole chain.
5084 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5085 the certificate can be used for (if anything). Set valid_flags field
5086 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5087 to have similar checks in it.
5089 Add new "cert_flags" field to CERT structure and include a "strict mode".
5090 This enforces some TLS certificate requirements (such as only permitting
5091 certificate signature algorithms contained in the supported algorithms
5092 extension) which some implementations ignore: this option should be used
5093 with caution as it could cause interoperability issues.
5097 * Update and tidy signature algorithm extension processing. Work out
5098 shared signature algorithms based on preferences and peer algorithms
5099 and print them out in s_client and s_server. Abort handshake if no
5100 shared signature algorithms.
5104 * Add new functions to allow customised supported signature algorithms
5105 for SSL and SSL_CTX structures. Add options to s_client and s_server
5110 * New function SSL_certs_clear() to delete all references to certificates
5111 from an SSL structure. Before this once a certificate had been added
5112 it couldn't be removed.
5116 * Integrate hostname, email address and IP address checking with certificate
5117 verification. New verify options supporting checking in openssl utility.
5121 * Fixes and wildcard matching support to hostname and email checking
5122 functions. Add manual page.
5124 *Florian Weimer (Red Hat Product Security Team)*
5126 * New functions to check a hostname email or IP address against a
5127 certificate. Add options x509 utility to print results of checks against
5132 * Fix OCSP checking.
5134 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5136 * Initial experimental support for explicitly trusted non-root CAs.
5137 OpenSSL still tries to build a complete chain to a root but if an
5138 intermediate CA has a trust setting included that is used. The first
5139 setting is used: whether to trust (e.g., -addtrust option to the x509
5144 * Add -trusted_first option which attempts to find certificates in the
5145 trusted store even if an untrusted chain is also supplied.
5149 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5150 platform support for Linux and Android.
5154 * Support for linux-x32, ILP32 environment in x86_64 framework.
5158 * Experimental multi-implementation support for FIPS capable OpenSSL.
5159 When in FIPS mode the approved implementations are used as normal,
5160 when not in FIPS mode the internal unapproved versions are used instead.
5161 This means that the FIPS capable OpenSSL isn't forced to use the
5162 (often lower performance) FIPS implementations outside FIPS mode.
5166 * Transparently support X9.42 DH parameters when calling
5167 PEM_read_bio_DHparameters. This means existing applications can handle
5168 the new parameter format automatically.
5172 * Initial experimental support for X9.42 DH parameter format: mainly
5173 to support use of 'q' parameter for RFC5114 parameters.
5177 * Add DH parameters from RFC5114 including test data to dhtest.
5181 * Support for automatic EC temporary key parameter selection. If enabled
5182 the most preferred EC parameters are automatically used instead of
5183 hardcoded fixed parameters. Now a server just has to call:
5184 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5185 support ECDH and use the most appropriate parameters.
5189 * Enhance and tidy EC curve and point format TLS extension code. Use
5190 static structures instead of allocation if default values are used.
5191 New ctrls to set curves we wish to support and to retrieve shared curves.
5192 Print out shared curves in s_server. New options to s_server and s_client
5193 to set list of supported curves.
5197 * New ctrls to retrieve supported signature algorithms and
5198 supported curve values as an array of NIDs. Extend openssl utility
5199 to print out received values.
5203 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5204 between NIDs and the more common NIST names such as "P-256". Enhance
5205 ecparam utility and ECC method to recognise the NIST names for curves.
5209 * Enhance SSL/TLS certificate chain handling to support different
5210 chains for each certificate instead of one chain in the parent SSL_CTX.
5214 * Support for fixed DH ciphersuite client authentication: where both
5215 server and client use DH certificates with common parameters.
5219 * Support for fixed DH ciphersuites: those requiring DH server
5224 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5226 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5227 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5228 X509_CINF_get_signature were reverted post internal team review.
5233 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5235 * OCSP Status Request extension unbounded memory growth
5237 A malicious client can send an excessively large OCSP Status Request
5238 extension. If that client continually requests renegotiation, sending a
5239 large OCSP Status Request extension each time, then there will be unbounded
5240 memory growth on the server. This will eventually lead to a Denial Of
5241 Service attack through memory exhaustion. Servers with a default
5242 configuration are vulnerable even if they do not support OCSP. Builds using
5243 the "no-ocsp" build time option are not affected.
5245 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5250 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5253 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5259 * OOB write in MDC2_Update()
5261 An overflow can occur in MDC2_Update() either if called directly or
5262 through the EVP_DigestUpdate() function using MDC2. If an attacker
5263 is able to supply very large amounts of input data after a previous
5264 call to EVP_EncryptUpdate() with a partial block then a length check
5265 can overflow resulting in a heap corruption.
5267 The amount of data needed is comparable to SIZE_MAX which is impractical
5270 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5275 * Malformed SHA512 ticket DoS
5277 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5278 DoS attack where a malformed ticket will result in an OOB read which will
5281 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5282 a custom server callback and ticket lookup mechanism.
5284 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5289 * OOB write in BN_bn2dec()
5291 The function BN_bn2dec() does not check the return value of BN_div_word().
5292 This can cause an OOB write if an application uses this function with an
5293 overly large BIGNUM. This could be a problem if an overly large certificate
5294 or CRL is printed out from an untrusted source. TLS is not affected because
5295 record limits will reject an oversized certificate before it is parsed.
5297 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5302 * OOB read in TS_OBJ_print_bio()
5304 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5305 the total length the OID text representation would use and not the amount
5306 of data written. This will result in OOB reads when large OIDs are
5309 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5314 * Pointer arithmetic undefined behaviour
5316 Avoid some undefined pointer arithmetic
5318 A common idiom in the codebase is to check limits in the following manner:
5321 Where "p" points to some malloc'd data of SIZE bytes and
5324 "len" here could be from some externally supplied data (e.g. from a TLS
5327 The rules of C pointer arithmetic are such that "p + len" is only well
5328 defined where len <= SIZE. Therefore the above idiom is actually
5329 undefined behaviour.
5331 For example this could cause problems if some malloc implementation
5332 provides an address for "p" such that "p + len" actually overflows for
5333 values of len that are too big and therefore p + len < limit.
5335 This issue was reported to OpenSSL by Guido Vranken
5340 * Constant time flag not preserved in DSA signing
5342 Operations in the DSA signing algorithm should run in constant time in
5343 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5344 implementation means that a non-constant time codepath is followed for
5345 certain operations. This has been demonstrated through a cache-timing
5346 attack to be sufficient for an attacker to recover the private DSA key.
5348 This issue was reported by César Pereida (Aalto University), Billy Brumley
5349 (Tampere University of Technology), and Yuval Yarom (The University of
5350 Adelaide and NICTA).
5355 * DTLS buffered message DoS
5357 In a DTLS connection where handshake messages are delivered out-of-order
5358 those messages that OpenSSL is not yet ready to process will be buffered
5359 for later use. Under certain circumstances, a flaw in the logic means that
5360 those messages do not get removed from the buffer even though the handshake
5361 has been completed. An attacker could force up to approx. 15 messages to
5362 remain in the buffer when they are no longer required. These messages will
5363 be cleared when the DTLS connection is closed. The default maximum size for
5364 a message is 100k. Therefore the attacker could force an additional 1500k
5365 to be consumed per connection. By opening many simulataneous connections an
5366 attacker could cause a DoS attack through memory exhaustion.
5368 This issue was reported to OpenSSL by Quan Luo.
5373 * DTLS replay protection DoS
5375 A flaw in the DTLS replay attack protection mechanism means that records
5376 that arrive for future epochs update the replay protection "window" before
5377 the MAC for the record has been validated. This could be exploited by an
5378 attacker by sending a record for the next epoch (which does not have to
5379 decrypt or have a valid MAC), with a very large sequence number. This means
5380 that all subsequent legitimate packets are dropped causing a denial of
5381 service for a specific DTLS connection.
5383 This issue was reported to OpenSSL by the OCAP audit team.
5388 * Certificate message OOB reads
5390 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5391 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5392 theoretical DoS risk but this has not been observed in practice on common
5395 The messages affected are client certificate, client certificate request
5396 and server certificate. As a result the attack can only be performed
5397 against a client or a server which enables client authentication.
5399 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5404 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5406 * Prevent padding oracle in AES-NI CBC MAC check
5408 A MITM attacker can use a padding oracle attack to decrypt traffic
5409 when the connection uses an AES CBC cipher and the server support
5412 This issue was introduced as part of the fix for Lucky 13 padding
5413 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5414 constant time by making sure that always the same bytes are read and
5415 compared against either the MAC or padding bytes. But it no longer
5416 checked that there was enough data to have both the MAC and padding
5419 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5424 * Fix EVP_EncodeUpdate overflow
5426 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5427 Base64 encoding of binary data. If an attacker is able to supply very large
5428 amounts of input data then a length check can overflow resulting in a heap
5431 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5432 the `PEM_write_bio*` family of functions. These are mainly used within the
5433 OpenSSL command line applications, so any application which processes data
5434 from an untrusted source and outputs it as a PEM file should be considered
5435 vulnerable to this issue. User applications that call these APIs directly
5436 with large amounts of untrusted data may also be vulnerable.
5438 This issue was reported by Guido Vranken.
5443 * Fix EVP_EncryptUpdate overflow
5445 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5446 is able to supply very large amounts of input data after a previous call to
5447 EVP_EncryptUpdate() with a partial block then a length check can overflow
5448 resulting in a heap corruption. Following an analysis of all OpenSSL
5449 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5450 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5451 the first called function after an EVP_EncryptInit(), and therefore that
5452 specific call must be safe. The second form is where the length passed to
5453 EVP_EncryptUpdate() can be seen from the code to be some small value and
5454 therefore there is no possibility of an overflow. Since all instances are
5455 one of these two forms, it is believed that there can be no overflows in
5456 internal code due to this problem. It should be noted that
5457 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5458 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5459 of these calls have also been analysed too and it is believed there are no
5460 instances in internal usage where an overflow could occur.
5462 This issue was reported by Guido Vranken.
5467 * Prevent ASN.1 BIO excessive memory allocation
5469 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5470 a short invalid encoding can casuse allocation of large amounts of memory
5471 potentially consuming excessive resources or exhausting memory.
5473 Any application parsing untrusted data through d2i BIO functions is
5474 affected. The memory based functions such as d2i_X509() are *not* affected.
5475 Since the memory based functions are used by the TLS library, TLS
5476 applications are not affected.
5478 This issue was reported by Brian Carpenter.
5485 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5486 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5487 in arbitrary stack data being returned in the buffer.
5489 This issue was reported by Guido Vranken.
5494 * Modify behavior of ALPN to invoke callback after SNI/servername
5495 callback, such that updates to the SSL_CTX affect ALPN.
5499 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5504 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5505 methods are enabled and ssl2 is disabled the methods return NULL.
5509 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5511 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5512 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5513 provide any "EXPORT" or "LOW" strength ciphers.
5517 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5518 is by default disabled at build-time. Builds that are not configured with
5519 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5520 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5521 will need to explicitly call either of:
5523 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5525 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5527 as appropriate. Even if either of those is used, or the application
5528 explicitly uses the version-specific SSLv2_method() or its client and
5529 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5530 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5531 ciphers, and SSLv2 56-bit DES are no longer available.
5536 * Fix a double-free in DSA code
5538 A double free bug was discovered when OpenSSL parses malformed DSA private
5539 keys and could lead to a DoS attack or memory corruption for applications
5540 that receive DSA private keys from untrusted sources. This scenario is
5543 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5549 * Disable SRP fake user seed to address a server memory leak.
5551 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5553 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5554 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5555 was changed to ignore the "fake user" SRP seed, even if the seed
5558 Users should use SRP_VBASE_get1_by_user instead. Note that in
5559 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5560 also that even though configuring the SRP seed attempts to hide
5561 invalid usernames by continuing the handshake with fake
5562 credentials, this behaviour is not constant time and no strong
5563 guarantees are made that the handshake is indistinguishable from
5564 that of a valid user.
5569 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5571 In the BN_hex2bn function the number of hex digits is calculated using an
5572 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5573 large values of |i| this can result in |bn_expand| not allocating any
5574 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5575 field as NULL leading to a subsequent NULL ptr deref. For very large values
5576 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5577 In this case memory is allocated to the internal BIGNUM data field, but it
5578 is insufficiently sized leading to heap corruption. A similar issue exists
5579 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5580 is ever called by user applications with very large untrusted hex/dec data.
5581 This is anticipated to be a rare occurrence.
5583 All OpenSSL internal usage of these functions use data that is not expected
5584 to be untrusted, e.g. config file data or application command line
5585 arguments. If user developed applications generate config file data based
5586 on untrusted data then it is possible that this could also lead to security
5587 consequences. This is also anticipated to be rare.
5589 This issue was reported to OpenSSL by Guido Vranken.
5594 * Fix memory issues in `BIO_*printf` functions
5596 The internal |fmtstr| function used in processing a "%s" format string in
5597 the `BIO_*printf` functions could overflow while calculating the length of a
5598 string and cause an OOB read when printing very long strings.
5600 Additionally the internal |doapr_outch| function can attempt to write to an
5601 OOB memory location (at an offset from the NULL pointer) in the event of a
5602 memory allocation failure. In 1.0.2 and below this could be caused where
5603 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5604 could be in processing a very long "%s" format string. Memory leaks can
5607 The first issue may mask the second issue dependent on compiler behaviour.
5608 These problems could enable attacks where large amounts of untrusted data
5609 is passed to the `BIO_*printf` functions. If applications use these functions
5610 in this way then they could be vulnerable. OpenSSL itself uses these
5611 functions when printing out human-readable dumps of ASN.1 data. Therefore
5612 applications that print this data could be vulnerable if the data is from
5613 untrusted sources. OpenSSL command line applications could also be
5614 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5615 as command line arguments.
5617 Libssl is not considered directly vulnerable. Additionally certificates etc
5618 received via remote connections via libssl are also unlikely to be able to
5619 trigger these issues because of message size limits enforced within libssl.
5621 This issue was reported to OpenSSL Guido Vranken.
5626 * Side channel attack on modular exponentiation
5628 A side-channel attack was found which makes use of cache-bank conflicts on
5629 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5630 of RSA keys. The ability to exploit this issue is limited as it relies on
5631 an attacker who has control of code in a thread running on the same
5632 hyper-threaded core as the victim thread which is performing decryptions.
5634 This issue was reported to OpenSSL by Yuval Yarom, The University of
5635 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5636 Nadia Heninger, University of Pennsylvania with more information at
5637 <http://cachebleed.info>.
5642 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5643 if no keysize is specified with default_bits. This fixes an
5644 omission in an earlier change that changed all RSA/DSA key generation
5645 apps to use 2048 bits by default.
5649 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5651 * Protection for DH small subgroup attacks
5653 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5654 switched on by default and cannot be disabled. This could have some
5659 * SSLv2 doesn't block disabled ciphers
5661 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5662 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5663 been disabled, provided that the SSLv2 protocol was not also disabled via
5666 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5667 and Sebastian Schinzel.
5672 * Reject DH handshakes with parameters shorter than 1024 bits.
5676 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5678 * Certificate verify crash with missing PSS parameter
5680 The signature verification routines will crash with a NULL pointer
5681 dereference if presented with an ASN.1 signature using the RSA PSS
5682 algorithm and absent mask generation function parameter. Since these
5683 routines are used to verify certificate signature algorithms this can be
5684 used to crash any certificate verification operation and exploited in a
5685 DoS attack. Any application which performs certificate verification is
5686 vulnerable including OpenSSL clients and servers which enable client
5689 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5694 * X509_ATTRIBUTE memory leak
5696 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5697 memory. This structure is used by the PKCS#7 and CMS routines so any
5698 application which reads PKCS#7 or CMS data from untrusted sources is
5699 affected. SSL/TLS is not affected.
5701 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5707 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5708 This changes the decoding behaviour for some invalid messages,
5709 though the change is mostly in the more lenient direction, and
5710 legacy behaviour is preserved as much as possible.
5714 * In DSA_generate_parameters_ex, if the provided seed is too short,
5715 use a random seed, as already documented.
5717 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5719 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5721 * Alternate chains certificate forgery
5723 During certificate verfification, OpenSSL will attempt to find an
5724 alternative certificate chain if the first attempt to build such a chain
5725 fails. An error in the implementation of this logic can mean that an
5726 attacker could cause certain checks on untrusted certificates to be
5727 bypassed, such as the CA flag, enabling them to use a valid leaf
5728 certificate to act as a CA and "issue" an invalid certificate.
5730 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5736 * Race condition handling PSK identify hint
5738 If PSK identity hints are received by a multi-threaded client then
5739 the values are wrongly updated in the parent SSL_CTX structure. This can
5740 result in a race condition potentially leading to a double free of the
5746 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5748 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5749 incompatibility in the handling of HMAC. The previous ABI has now been
5752 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5754 * Malformed ECParameters causes infinite loop
5756 When processing an ECParameters structure OpenSSL enters an infinite loop
5757 if the curve specified is over a specially malformed binary polynomial
5760 This can be used to perform denial of service against any
5761 system which processes public keys, certificate requests or
5762 certificates. This includes TLS clients and TLS servers with
5763 client authentication enabled.
5765 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5770 * Exploitable out-of-bounds read in X509_cmp_time
5772 X509_cmp_time does not properly check the length of the ASN1_TIME
5773 string and can read a few bytes out of bounds. In addition,
5774 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5777 An attacker can use this to craft malformed certificates and CRLs of
5778 various sizes and potentially cause a segmentation fault, resulting in
5779 a DoS on applications that verify certificates or CRLs. TLS clients
5780 that verify CRLs are affected. TLS clients and servers with client
5781 authentication enabled may be affected if they use custom verification
5784 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5785 independently by Hanno Böck.
5790 * PKCS7 crash with missing EnvelopedContent
5792 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5793 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5794 with missing content and trigger a NULL pointer dereference on parsing.
5796 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5797 structures from untrusted sources are affected. OpenSSL clients and
5798 servers are not affected.
5800 This issue was reported to OpenSSL by Michal Zalewski (Google).
5805 * CMS verify infinite loop with unknown hash function
5807 When verifying a signedData message the CMS code can enter an infinite loop
5808 if presented with an unknown hash function OID. This can be used to perform
5809 denial of service against any system which verifies signedData messages using
5811 This issue was reported to OpenSSL by Johannes Bauer.
5816 * Race condition handling NewSessionTicket
5818 If a NewSessionTicket is received by a multi-threaded client when attempting to
5819 reuse a previous ticket then a race condition can occur potentially leading to
5820 a double free of the ticket data.
5825 * Reject DH handshakes with parameters shorter than 768 bits.
5827 *Kurt Roeckx and Emilia Kasper*
5829 * dhparam: generate 2048-bit parameters by default.
5831 *Kurt Roeckx and Emilia Kasper*
5833 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5835 * Segmentation fault in ASN1_TYPE_cmp fix
5837 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5838 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5839 certificate signature algorithm consistency this can be used to crash any
5840 certificate verification operation and exploited in a DoS attack. Any
5841 application which performs certificate verification is vulnerable including
5842 OpenSSL clients and servers which enable client authentication.
5847 * ASN.1 structure reuse memory corruption fix
5849 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5850 memory corruption via an invalid write. Such reuse is and has been
5851 strongly discouraged and is believed to be rare.
5853 Applications that parse structures containing CHOICE or ANY DEFINED BY
5854 components may be affected. Certificate parsing (d2i_X509 and related
5855 functions) are however not affected. OpenSSL clients and servers are
5861 * PKCS7 NULL pointer dereferences fix
5863 The PKCS#7 parsing code does not handle missing outer ContentInfo
5864 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5865 missing content and trigger a NULL pointer dereference on parsing.
5867 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5868 otherwise parse PKCS#7 structures from untrusted sources are
5869 affected. OpenSSL clients and servers are not affected.
5871 This issue was reported to OpenSSL by Michal Zalewski (Google).
5876 * DoS via reachable assert in SSLv2 servers fix
5878 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5879 servers that both support SSLv2 and enable export cipher suites by sending
5880 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5882 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5883 (OpenSSL development team).
5888 * Use After Free following d2i_ECPrivatekey error fix
5890 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5891 could cause a use after free condition. This, in turn, could cause a double
5892 free in several private key parsing functions (such as d2i_PrivateKey
5893 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5894 for applications that receive EC private keys from untrusted
5895 sources. This scenario is considered rare.
5897 This issue was discovered by the BoringSSL project and fixed in their
5903 * X509_to_X509_REQ NULL pointer deref fix
5905 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5906 the certificate key is invalid. This function is rarely used in practice.
5908 This issue was discovered by Brian Carpenter.
5913 * Removed the export ciphers from the DEFAULT ciphers
5917 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
5919 * Build fixes for the Windows and OpenVMS platforms
5921 *Matt Caswell and Richard Levitte*
5923 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
5925 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5926 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5927 dereference. This could lead to a Denial Of Service attack. Thanks to
5928 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5933 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5934 dtls1_buffer_record function under certain conditions. In particular this
5935 could occur if an attacker sent repeated DTLS records with the same
5936 sequence number but for the next epoch. The memory leak could be exploited
5937 by an attacker in a Denial of Service attack through memory exhaustion.
5938 Thanks to Chris Mueller for reporting this issue.
5943 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5944 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5945 method would be set to NULL which could later result in a NULL pointer
5946 dereference. Thanks to Frank Schmirler for reporting this issue.
5951 * Abort handshake if server key exchange message is omitted for ephemeral
5954 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5955 reporting this issue.
5960 * Remove non-export ephemeral RSA code on client and server. This code
5961 violated the TLS standard by allowing the use of temporary RSA keys in
5962 non-export ciphersuites and could be used by a server to effectively
5963 downgrade the RSA key length used to a value smaller than the server
5964 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5965 INRIA or reporting this issue.
5970 * Fixed issue where DH client certificates are accepted without verification.
5971 An OpenSSL server will accept a DH certificate for client authentication
5972 without the certificate verify message. This effectively allows a client to
5973 authenticate without the use of a private key. This only affects servers
5974 which trust a client certificate authority which issues certificates
5975 containing DH keys: these are extremely rare and hardly ever encountered.
5976 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5982 * Ensure that the session ID context of an SSL is updated when its
5983 SSL_CTX is updated via SSL_set_SSL_CTX.
5985 The session ID context is typically set from the parent SSL_CTX,
5986 and can vary with the CTX.
5990 * Fix various certificate fingerprint issues.
5992 By using non-DER or invalid encodings outside the signed portion of a
5993 certificate the fingerprint can be changed without breaking the signature.
5994 Although no details of the signed portion of the certificate can be changed
5995 this can cause problems with some applications: e.g. those using the
5996 certificate fingerprint for blacklists.
5998 1. Reject signatures with non zero unused bits.
6000 If the BIT STRING containing the signature has non zero unused bits reject
6001 the signature. All current signature algorithms require zero unused bits.
6003 2. Check certificate algorithm consistency.
6005 Check the AlgorithmIdentifier inside TBS matches the one in the
6006 certificate signature. NB: this will result in signature failure
6007 errors for some broken certificates.
6009 Thanks to Konrad Kraszewski from Google for reporting this issue.
6011 3. Check DSA/ECDSA signatures use DER.
6013 Re-encode DSA/ECDSA signatures and compare with the original received
6014 signature. Return an error if there is a mismatch.
6016 This will reject various cases including garbage after signature
6017 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6018 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6019 (negative or with leading zeroes).
6021 Further analysis was conducted and fixes were developed by Stephen Henson
6022 of the OpenSSL core team.
6028 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6029 results on some platforms, including x86_64. This bug occurs at random
6030 with a very low probability, and is not known to be exploitable in any
6031 way, though its exact impact is difficult to determine. Thanks to Pieter
6032 Wuille (Blockstream) who reported this issue and also suggested an initial
6033 fix. Further analysis was conducted by the OpenSSL development team and
6034 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6035 the OpenSSL core team.
6040 * Do not resume sessions on the server if the negotiated protocol
6041 version does not match the session's version. Resuming with a different
6042 version, while not strictly forbidden by the RFC, is of questionable
6043 sanity and breaks all known clients.
6045 *David Benjamin, Emilia Käsper*
6047 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6048 early CCS messages during renegotiation. (Note that because
6049 renegotiation is encrypted, this early CCS was not exploitable.)
6053 * Tighten client-side session ticket handling during renegotiation:
6054 ensure that the client only accepts a session ticket if the server sends
6055 the extension anew in the ServerHello. Previously, a TLS client would
6056 reuse the old extension state and thus accept a session ticket if one was
6057 announced in the initial ServerHello.
6059 Similarly, ensure that the client requires a session ticket if one
6060 was advertised in the ServerHello. Previously, a TLS client would
6061 ignore a missing NewSessionTicket message.
6065 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6069 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6070 sends a carefully crafted handshake message, to cause OpenSSL to fail
6071 to free up to 64k of memory causing a memory leak. This could be
6072 exploited in a Denial Of Service attack. This issue affects OpenSSL
6073 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6074 whether SRTP is used or configured. Implementations of OpenSSL that
6075 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6077 The fix was developed by the OpenSSL team.
6082 * Session Ticket Memory Leak.
6084 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6085 integrity of that ticket is first verified. In the event of a session
6086 ticket integrity check failing, OpenSSL will fail to free memory
6087 causing a memory leak. By sending a large number of invalid session
6088 tickets an attacker could exploit this issue in a Denial Of Service
6094 * Build option no-ssl3 is incomplete.
6096 When OpenSSL is configured with "no-ssl3" as a build option, servers
6097 could accept and complete a SSL 3.0 handshake, and clients could be
6098 configured to send them.
6101 *Akamai and the OpenSSL team*
6103 * Add support for TLS_FALLBACK_SCSV.
6104 Client applications doing fallback retries should call
6105 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6108 *Adam Langley, Bodo Moeller*
6110 * Add additional DigestInfo checks.
6112 Re-encode DigestInto in DER and check against the original when
6113 verifying RSA signature: this will reject any improperly encoded
6114 DigestInfo structures.
6116 Note: this is a precautionary measure and no attacks are currently known.
6120 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6122 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6123 SRP code can be overrun an internal buffer. Add sanity check that
6124 g, A, B < N to SRP code.
6126 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6127 Group for discovering this issue.
6132 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6133 TLS 1.0 instead of higher protocol versions when the ClientHello message
6134 is badly fragmented. This allows a man-in-the-middle attacker to force a
6135 downgrade to TLS 1.0 even if both the server and the client support a
6136 higher protocol version, by modifying the client's TLS records.
6138 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6139 researching this issue.
6144 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6145 to a denial of service attack. A malicious server can crash the client
6146 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6147 ciphersuite and sending carefully crafted handshake messages.
6149 Thanks to Felix Gröbert (Google) for discovering and researching this
6155 * By sending carefully crafted DTLS packets an attacker could cause openssl
6156 to leak memory. This can be exploited through a Denial of Service attack.
6157 Thanks to Adam Langley for discovering and researching this issue.
6162 * An attacker can force openssl to consume large amounts of memory whilst
6163 processing DTLS handshake messages. This can be exploited through a
6164 Denial of Service attack.
6165 Thanks to Adam Langley for discovering and researching this issue.
6170 * An attacker can force an error condition which causes openssl to crash
6171 whilst processing DTLS packets due to memory being freed twice. This
6172 can be exploited through a Denial of Service attack.
6173 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6179 * If a multithreaded client connects to a malicious server using a resumed
6180 session and the server sends an ec point format extension it could write
6181 up to 255 bytes to freed memory.
6183 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6189 * A malicious server can crash an OpenSSL client with a null pointer
6190 dereference (read) by specifying an SRP ciphersuite even though it was not
6191 properly negotiated with the client. This can be exploited through a
6192 Denial of Service attack.
6194 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6195 discovering and researching this issue.
6200 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6201 X509_name_oneline, X509_name_print_ex et al. to leak some information
6202 from the stack. Applications may be affected if they echo pretty printing
6203 output to the attacker.
6205 Thanks to Ivan Fratric (Google) for discovering this issue.
6208 *Emilia Käsper, and Steve Henson*
6210 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6211 for corner cases. (Certain input points at infinity could lead to
6212 bogus results, with non-infinity inputs mapped to infinity too.)
6216 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6218 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6219 handshake can force the use of weak keying material in OpenSSL
6220 SSL/TLS clients and servers.
6222 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6223 researching this issue. [CVE-2014-0224][]
6225 *KIKUCHI Masashi, Steve Henson*
6227 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6228 OpenSSL DTLS client the code can be made to recurse eventually crashing
6231 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6234 *Imre Rad, Steve Henson*
6236 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6237 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6238 client or server. This is potentially exploitable to run arbitrary
6239 code on a vulnerable client or server.
6241 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6243 *Jüri Aedla, Steve Henson*
6245 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6246 are subject to a denial of service attack.
6248 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6249 this issue. [CVE-2014-3470][]
6251 *Felix Gröbert, Ivan Fratric, Steve Henson*
6253 * Harmonize version and its documentation. -f flag is used to display
6256 *mancha <mancha1@zoho.com>*
6258 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6259 in i2d_ECPrivateKey.
6261 *mancha <mancha1@zoho.com>*
6263 * Fix some double frees. These are not thought to be exploitable.
6265 *mancha <mancha1@zoho.com>*
6267 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6269 * A missing bounds check in the handling of the TLS heartbeat extension
6270 can be used to reveal up to 64k of memory to a connected client or
6273 Thanks for Neel Mehta of Google Security for discovering this bug and to
6274 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6275 preparing the fix [CVE-2014-0160][]
6277 *Adam Langley, Bodo Moeller*
6279 * Fix for the attack described in the paper "Recovering OpenSSL
6280 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6281 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6282 <http://eprint.iacr.org/2014/140>
6284 Thanks to Yuval Yarom and Naomi Benger for discovering this
6285 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6287 *Yuval Yarom and Naomi Benger*
6289 * TLS pad extension: draft-agl-tls-padding-03
6291 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6292 TLS client Hello record length value would otherwise be > 255 and
6293 less that 512 pad with a dummy extension containing zeroes so it
6294 is at least 512 bytes long.
6296 *Adam Langley, Steve Henson*
6298 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6300 * Fix for TLS record tampering bug. A carefully crafted invalid
6301 handshake could crash OpenSSL with a NULL pointer exception.
6302 Thanks to Anton Johansson for reporting this issues.
6305 * Keep original DTLS digest and encryption contexts in retransmission
6306 structures so we can use the previous session parameters if they need
6307 to be resent. [CVE-2013-6450][]
6311 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6312 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6313 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6314 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6315 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6316 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6318 *Rob Stradling, Adam Langley*
6320 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6322 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6323 supporting platforms or when small records were transferred.
6325 *Andy Polyakov, Steve Henson*
6327 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6329 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6331 This addresses the flaw in CBC record processing discovered by
6332 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6333 at: <http://www.isg.rhul.ac.uk/tls/>
6335 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6336 Security Group at Royal Holloway, University of London
6337 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6338 Emilia Käsper for the initial patch.
6341 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6343 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6344 ciphersuites which can be exploited in a denial of service attack.
6345 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6346 and detecting this bug and to Wolfgang Ettlinger
6347 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6352 * Return an error when checking OCSP signatures when key is NULL.
6353 This fixes a DoS attack. [CVE-2013-0166][]
6357 * Make openssl verify return errors.
6359 *Chris Palmer <palmer@google.com> and Ben Laurie*
6361 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6362 the right response is stapled. Also change SSL_get_certificate()
6363 so it returns the certificate actually sent.
6364 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6366 *Rob Stradling <rob.stradling@comodo.com>*
6368 * Fix possible deadlock when decoding public keys.
6372 * Don't use TLS 1.0 record version number in initial client hello
6377 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6379 * Sanity check record length before skipping explicit IV in TLS
6380 1.2, 1.1 and DTLS to fix DoS attack.
6382 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6383 fuzzing as a service testing platform.
6388 * Initialise tkeylen properly when encrypting CMS messages.
6389 Thanks to Solar Designer of Openwall for reporting this issue.
6393 * In FIPS mode don't try to use composite ciphers as they are not
6398 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6400 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6401 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6402 mean any application compiled against OpenSSL 1.0.0 headers setting
6403 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6404 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6405 0x10000000L Any application which was previously compiled against
6406 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6407 will need to be recompiled as a result. Letting be results in
6408 inability to disable specifically TLS 1.1 and in client context,
6409 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6413 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6414 disable just protocol X, but all protocols above X *if* there are
6415 protocols *below* X still enabled. In more practical terms it means
6416 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6417 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6418 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6423 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6425 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6426 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6427 in CRYPTO_realloc_clean.
6429 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6430 issue and to Adam Langley <agl@chromium.org> for fixing it.
6433 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6435 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6439 * Workarounds for some broken servers that "hang" if a client hello
6440 record length exceeds 255 bytes.
6442 1. Do not use record version number > TLS 1.0 in initial client
6443 hello: some (but not all) hanging servers will now work.
6444 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6445 the number of ciphers sent in the client hello. This should be
6446 set to an even number, such as 50, for example by passing:
6447 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6448 Most broken servers should now work.
6449 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6450 TLS 1.2 client support entirely.
6454 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6458 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6460 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6461 STRING form instead of a DigestInfo.
6465 * The format used for MDC2 RSA signatures is inconsistent between EVP
6466 and the RSA_sign/RSA_verify functions. This was made more apparent when
6467 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6468 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6469 the correct format in RSA_verify so both forms transparently work.
6473 * Some servers which support TLS 1.0 can choke if we initially indicate
6474 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6475 encrypted premaster secret. As a workaround use the maximum permitted
6476 client version in client hello, this should keep such servers happy
6477 and still work with previous versions of OpenSSL.
6481 * Add support for TLS/DTLS heartbeats.
6483 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6485 * Add support for SCTP.
6487 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6489 * Improved PRNG seeding for VOS.
6491 *Paul Green <Paul.Green@stratus.com>*
6493 * Extensive assembler packs updates, most notably:
6495 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6496 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6497 - x86_64: bit-sliced AES implementation;
6498 - ARM: NEON support, contemporary platforms optimizations;
6499 - s390x: z196 support;
6500 - `*`: GHASH and GF(2^m) multiplication implementations;
6504 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6505 (removal of unnecessary code)
6507 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6509 * Add TLS key material exporter from RFC 5705.
6513 * Add DTLS-SRTP negotiation from RFC 5764.
6517 * Add Next Protocol Negotiation,
6518 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6519 disabled with a no-npn flag to config or Configure. Code donated
6522 *Adam Langley <agl@google.com> and Ben Laurie*
6524 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6525 NIST-P256, NIST-P521, with constant-time single point multiplication on
6526 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6527 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6528 Code made available under Apache License version 2.0.
6530 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6531 line to include this in your build of OpenSSL, and run "make depend" (or
6532 "make update"). This enables the following EC_METHODs:
6534 EC_GFp_nistp224_method()
6535 EC_GFp_nistp256_method()
6536 EC_GFp_nistp521_method()
6538 EC_GROUP_new_by_curve_name() will automatically use these (while
6539 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6542 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6544 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6545 all platforms. Move ssize_t definition from e_os.h to the public
6546 header file e_os2.h as it now appears in public header file cms.h
6550 * New -sigopt option to the ca, req and x509 utilities. Additional
6551 signature parameters can be passed using this option and in
6556 * Add RSA PSS signing function. This will generate and set the
6557 appropriate AlgorithmIdentifiers for PSS based on those in the
6558 corresponding EVP_MD_CTX structure. No application support yet.
6562 * Support for companion algorithm specific ASN1 signing routines.
6563 New function ASN1_item_sign_ctx() signs a pre-initialised
6564 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6565 the appropriate parameters.
6569 * Add new algorithm specific ASN1 verification initialisation function
6570 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6571 handling will be the same no matter what EVP_PKEY_METHOD is used.
6572 Add a PSS handler to support verification of PSS signatures: checked
6573 against a number of sample certificates.
6577 * Add signature printing for PSS. Add PSS OIDs.
6579 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6581 * Add algorithm specific signature printing. An individual ASN1 method
6582 can now print out signatures instead of the standard hex dump.
6584 More complex signatures (e.g. PSS) can print out more meaningful
6585 information. Include DSA version that prints out the signature
6590 * Password based recipient info support for CMS library: implementing
6595 * Split password based encryption into PBES2 and PBKDF2 functions. This
6596 neatly separates the code into cipher and PBE sections and is required
6597 for some algorithms that split PBES2 into separate pieces (such as
6598 password based CMS).
6602 * Session-handling fixes:
6603 - Fix handling of connections that are resuming with a session ID,
6604 but also support Session Tickets.
6605 - Fix a bug that suppressed issuing of a new ticket if the client
6606 presented a ticket with an expired session.
6607 - Try to set the ticket lifetime hint to something reasonable.
6608 - Make tickets shorter by excluding irrelevant information.
6609 - On the client side, don't ignore renewed tickets.
6611 *Adam Langley, Bodo Moeller (Google)*
6613 * Fix PSK session representation.
6617 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6619 This work was sponsored by Intel.
6623 * Add GCM support to TLS library. Some custom code is needed to split
6624 the IV between the fixed (from PRF) and explicit (from TLS record)
6625 portions. This adds all GCM ciphersuites supported by RFC5288 and
6626 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6627 add a special AESGCM string for GCM only.
6631 * Expand range of ctrls for AES GCM. Permit setting invocation
6632 field on decrypt and retrieval of invocation field only on encrypt.
6636 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6637 As required by RFC5289 these ciphersuites cannot be used if for
6638 versions of TLS earlier than 1.2.
6642 * For FIPS capable OpenSSL interpret a NULL default public key method
6643 as unset and return the appropriate default but do *not* set the default.
6644 This means we can return the appropriate method in applications that
6645 switch between FIPS and non-FIPS modes.
6649 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6650 ENGINE is used then we cannot handle that in the FIPS module so we
6651 keep original code iff non-FIPS operations are allowed.
6655 * Add -attime option to openssl utilities.
6657 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6659 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6663 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6664 FIPS EC methods unconditionally for now.
6668 * New build option no-ec2m to disable characteristic 2 code.
6672 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6673 all cases can be covered as some introduce binary incompatibilities.
6677 * Redirect RSA operations to FIPS module including keygen,
6678 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6682 * Add similar low level API blocking to ciphers.
6686 * Low level digest APIs are not approved in FIPS mode: any attempt
6687 to use these will cause a fatal error. Applications that *really* want
6688 to use them can use the `private_*` version instead.
6692 * Redirect cipher operations to FIPS module for FIPS builds.
6696 * Redirect digest operations to FIPS module for FIPS builds.
6700 * Update build system to add "fips" flag which will link in fipscanister.o
6701 for static and shared library builds embedding a signature if needed.
6705 * Output TLS supported curves in preference order instead of numerical
6706 order. This is currently hardcoded for the highest order curves first.
6707 This should be configurable so applications can judge speed vs strength.
6711 * Add TLS v1.2 server support for client authentication.
6715 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6720 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6721 FIPS modules versions.
6725 * Add TLS v1.2 client side support for client authentication. Keep cache
6726 of handshake records longer as we don't know the hash algorithm to use
6727 until after the certificate request message is received.
6731 * Initial TLS v1.2 client support. Add a default signature algorithms
6732 extension including all the algorithms we support. Parse new signature
6733 format in client key exchange. Relax some ECC signing restrictions for
6734 TLS v1.2 as indicated in RFC5246.
6738 * Add server support for TLS v1.2 signature algorithms extension. Switch
6739 to new signature format when needed using client digest preference.
6740 All server ciphersuites should now work correctly in TLS v1.2. No client
6741 support yet and no support for client certificates.
6745 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6746 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6747 ciphersuites. At present only RSA key exchange ciphersuites work with
6748 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6749 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6750 and version checking.
6754 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6755 with this defined it will not be affected by any changes to ssl internal
6756 structures. Add several utility functions to allow openssl application
6757 to work with OPENSSL_NO_SSL_INTERN defined.
6761 * A long standing patch to add support for SRP from EdelWeb (Peter
6762 Sylvester and Christophe Renou) was integrated.
6763 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6764 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6767 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6771 * Permit abbreviated handshakes when renegotiating using the function
6772 SSL_renegotiate_abbreviated().
6774 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6776 * Add call to ENGINE_register_all_complete() to
6777 ENGINE_load_builtin_engines(), so some implementations get used
6778 automatically instead of needing explicit application support.
6782 * Add support for TLS key exporter as described in RFC5705.
6784 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6786 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6787 a few changes are required:
6789 Add SSL_OP_NO_TLSv1_1 flag.
6790 Add TLSv1_1 methods.
6791 Update version checking logic to handle version 1.1.
6792 Add explicit IV handling (ported from DTLS code).
6793 Add command line options to s_client/s_server.
6800 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6802 * X509_ATTRIBUTE memory leak
6804 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6805 memory. This structure is used by the PKCS#7 and CMS routines so any
6806 application which reads PKCS#7 or CMS data from untrusted sources is
6807 affected. SSL/TLS is not affected.
6809 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6815 * Race condition handling PSK identify hint
6817 If PSK identity hints are received by a multi-threaded client then
6818 the values are wrongly updated in the parent SSL_CTX structure. This can
6819 result in a race condition potentially leading to a double free of the
6825 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6827 * Malformed ECParameters causes infinite loop
6829 When processing an ECParameters structure OpenSSL enters an infinite loop
6830 if the curve specified is over a specially malformed binary polynomial
6833 This can be used to perform denial of service against any
6834 system which processes public keys, certificate requests or
6835 certificates. This includes TLS clients and TLS servers with
6836 client authentication enabled.
6838 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6843 * Exploitable out-of-bounds read in X509_cmp_time
6845 X509_cmp_time does not properly check the length of the ASN1_TIME
6846 string and can read a few bytes out of bounds. In addition,
6847 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6850 An attacker can use this to craft malformed certificates and CRLs of
6851 various sizes and potentially cause a segmentation fault, resulting in
6852 a DoS on applications that verify certificates or CRLs. TLS clients
6853 that verify CRLs are affected. TLS clients and servers with client
6854 authentication enabled may be affected if they use custom verification
6857 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6858 independently by Hanno Böck.
6863 * PKCS7 crash with missing EnvelopedContent
6865 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6866 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6867 with missing content and trigger a NULL pointer dereference on parsing.
6869 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6870 structures from untrusted sources are affected. OpenSSL clients and
6871 servers are not affected.
6873 This issue was reported to OpenSSL by Michal Zalewski (Google).
6878 * CMS verify infinite loop with unknown hash function
6880 When verifying a signedData message the CMS code can enter an infinite loop
6881 if presented with an unknown hash function OID. This can be used to perform
6882 denial of service against any system which verifies signedData messages using
6884 This issue was reported to OpenSSL by Johannes Bauer.
6889 * Race condition handling NewSessionTicket
6891 If a NewSessionTicket is received by a multi-threaded client when attempting to
6892 reuse a previous ticket then a race condition can occur potentially leading to
6893 a double free of the ticket data.
6898 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
6900 * Segmentation fault in ASN1_TYPE_cmp fix
6902 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6903 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6904 certificate signature algorithm consistency this can be used to crash any
6905 certificate verification operation and exploited in a DoS attack. Any
6906 application which performs certificate verification is vulnerable including
6907 OpenSSL clients and servers which enable client authentication.
6912 * ASN.1 structure reuse memory corruption fix
6914 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6915 memory corruption via an invalid write. Such reuse is and has been
6916 strongly discouraged and is believed to be rare.
6918 Applications that parse structures containing CHOICE or ANY DEFINED BY
6919 components may be affected. Certificate parsing (d2i_X509 and related
6920 functions) are however not affected. OpenSSL clients and servers are
6926 * PKCS7 NULL pointer dereferences fix
6928 The PKCS#7 parsing code does not handle missing outer ContentInfo
6929 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6930 missing content and trigger a NULL pointer dereference on parsing.
6932 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6933 otherwise parse PKCS#7 structures from untrusted sources are
6934 affected. OpenSSL clients and servers are not affected.
6936 This issue was reported to OpenSSL by Michal Zalewski (Google).
6941 * DoS via reachable assert in SSLv2 servers fix
6943 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6944 servers that both support SSLv2 and enable export cipher suites by sending
6945 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6947 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6948 (OpenSSL development team).
6953 * Use After Free following d2i_ECPrivatekey error fix
6955 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6956 could cause a use after free condition. This, in turn, could cause a double
6957 free in several private key parsing functions (such as d2i_PrivateKey
6958 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6959 for applications that receive EC private keys from untrusted
6960 sources. This scenario is considered rare.
6962 This issue was discovered by the BoringSSL project and fixed in their
6968 * X509_to_X509_REQ NULL pointer deref fix
6970 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6971 the certificate key is invalid. This function is rarely used in practice.
6973 This issue was discovered by Brian Carpenter.
6978 * Removed the export ciphers from the DEFAULT ciphers
6982 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
6984 * Build fixes for the Windows and OpenVMS platforms
6986 *Matt Caswell and Richard Levitte*
6988 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
6990 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6991 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6992 dereference. This could lead to a Denial Of Service attack. Thanks to
6993 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6998 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6999 dtls1_buffer_record function under certain conditions. In particular this
7000 could occur if an attacker sent repeated DTLS records with the same
7001 sequence number but for the next epoch. The memory leak could be exploited
7002 by an attacker in a Denial of Service attack through memory exhaustion.
7003 Thanks to Chris Mueller for reporting this issue.
7008 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7009 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7010 method would be set to NULL which could later result in a NULL pointer
7011 dereference. Thanks to Frank Schmirler for reporting this issue.
7016 * Abort handshake if server key exchange message is omitted for ephemeral
7019 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7020 reporting this issue.
7025 * Remove non-export ephemeral RSA code on client and server. This code
7026 violated the TLS standard by allowing the use of temporary RSA keys in
7027 non-export ciphersuites and could be used by a server to effectively
7028 downgrade the RSA key length used to a value smaller than the server
7029 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7030 INRIA or reporting this issue.
7035 * Fixed issue where DH client certificates are accepted without verification.
7036 An OpenSSL server will accept a DH certificate for client authentication
7037 without the certificate verify message. This effectively allows a client to
7038 authenticate without the use of a private key. This only affects servers
7039 which trust a client certificate authority which issues certificates
7040 containing DH keys: these are extremely rare and hardly ever encountered.
7041 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7047 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7048 results on some platforms, including x86_64. This bug occurs at random
7049 with a very low probability, and is not known to be exploitable in any
7050 way, though its exact impact is difficult to determine. Thanks to Pieter
7051 Wuille (Blockstream) who reported this issue and also suggested an initial
7052 fix. Further analysis was conducted by the OpenSSL development team and
7053 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7054 the OpenSSL core team.
7059 * Fix various certificate fingerprint issues.
7061 By using non-DER or invalid encodings outside the signed portion of a
7062 certificate the fingerprint can be changed without breaking the signature.
7063 Although no details of the signed portion of the certificate can be changed
7064 this can cause problems with some applications: e.g. those using the
7065 certificate fingerprint for blacklists.
7067 1. Reject signatures with non zero unused bits.
7069 If the BIT STRING containing the signature has non zero unused bits reject
7070 the signature. All current signature algorithms require zero unused bits.
7072 2. Check certificate algorithm consistency.
7074 Check the AlgorithmIdentifier inside TBS matches the one in the
7075 certificate signature. NB: this will result in signature failure
7076 errors for some broken certificates.
7078 Thanks to Konrad Kraszewski from Google for reporting this issue.
7080 3. Check DSA/ECDSA signatures use DER.
7082 Reencode DSA/ECDSA signatures and compare with the original received
7083 signature. Return an error if there is a mismatch.
7085 This will reject various cases including garbage after signature
7086 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7087 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7088 (negative or with leading zeroes).
7090 Further analysis was conducted and fixes were developed by Stephen Henson
7091 of the OpenSSL core team.
7097 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7099 * Session Ticket Memory Leak.
7101 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7102 integrity of that ticket is first verified. In the event of a session
7103 ticket integrity check failing, OpenSSL will fail to free memory
7104 causing a memory leak. By sending a large number of invalid session
7105 tickets an attacker could exploit this issue in a Denial Of Service
7111 * Build option no-ssl3 is incomplete.
7113 When OpenSSL is configured with "no-ssl3" as a build option, servers
7114 could accept and complete a SSL 3.0 handshake, and clients could be
7115 configured to send them.
7118 *Akamai and the OpenSSL team*
7120 * Add support for TLS_FALLBACK_SCSV.
7121 Client applications doing fallback retries should call
7122 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7125 *Adam Langley, Bodo Moeller*
7127 * Add additional DigestInfo checks.
7129 Reencode DigestInto in DER and check against the original when
7130 verifying RSA signature: this will reject any improperly encoded
7131 DigestInfo structures.
7133 Note: this is a precautionary measure and no attacks are currently known.
7137 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7139 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7140 to a denial of service attack. A malicious server can crash the client
7141 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7142 ciphersuite and sending carefully crafted handshake messages.
7144 Thanks to Felix Gröbert (Google) for discovering and researching this
7150 * By sending carefully crafted DTLS packets an attacker could cause openssl
7151 to leak memory. This can be exploited through a Denial of Service attack.
7152 Thanks to Adam Langley for discovering and researching this issue.
7157 * An attacker can force openssl to consume large amounts of memory whilst
7158 processing DTLS handshake messages. This can be exploited through a
7159 Denial of Service attack.
7160 Thanks to Adam Langley for discovering and researching this issue.
7165 * An attacker can force an error condition which causes openssl to crash
7166 whilst processing DTLS packets due to memory being freed twice. This
7167 can be exploited through a Denial of Service attack.
7168 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7174 * If a multithreaded client connects to a malicious server using a resumed
7175 session and the server sends an ec point format extension it could write
7176 up to 255 bytes to freed memory.
7178 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7184 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7185 X509_name_oneline, X509_name_print_ex et al. to leak some information
7186 from the stack. Applications may be affected if they echo pretty printing
7187 output to the attacker.
7189 Thanks to Ivan Fratric (Google) for discovering this issue.
7192 *Emilia Käsper, and Steve Henson*
7194 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7195 for corner cases. (Certain input points at infinity could lead to
7196 bogus results, with non-infinity inputs mapped to infinity too.)
7200 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7202 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7203 handshake can force the use of weak keying material in OpenSSL
7204 SSL/TLS clients and servers.
7206 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7207 researching this issue. [CVE-2014-0224][]
7209 *KIKUCHI Masashi, Steve Henson*
7211 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7212 OpenSSL DTLS client the code can be made to recurse eventually crashing
7215 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7218 *Imre Rad, Steve Henson*
7220 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7221 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7222 client or server. This is potentially exploitable to run arbitrary
7223 code on a vulnerable client or server.
7225 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7227 *Jüri Aedla, Steve Henson*
7229 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7230 are subject to a denial of service attack.
7232 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7233 this issue. [CVE-2014-3470][]
7235 *Felix Gröbert, Ivan Fratric, Steve Henson*
7237 * Harmonize version and its documentation. -f flag is used to display
7240 *mancha <mancha1@zoho.com>*
7242 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7243 in i2d_ECPrivateKey.
7245 *mancha <mancha1@zoho.com>*
7247 * Fix some double frees. These are not thought to be exploitable.
7249 *mancha <mancha1@zoho.com>*
7251 * Fix for the attack described in the paper "Recovering OpenSSL
7252 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7253 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7254 <http://eprint.iacr.org/2014/140>
7256 Thanks to Yuval Yarom and Naomi Benger for discovering this
7257 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7259 *Yuval Yarom and Naomi Benger*
7261 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7263 * Keep original DTLS digest and encryption contexts in retransmission
7264 structures so we can use the previous session parameters if they need
7265 to be resent. [CVE-2013-6450][]
7269 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7270 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7271 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7272 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7273 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7274 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7276 *Rob Stradling, Adam Langley*
7278 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7280 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7282 This addresses the flaw in CBC record processing discovered by
7283 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7284 at: <http://www.isg.rhul.ac.uk/tls/>
7286 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7287 Security Group at Royal Holloway, University of London
7288 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7289 Emilia Käsper for the initial patch.
7292 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7294 * Return an error when checking OCSP signatures when key is NULL.
7295 This fixes a DoS attack. [CVE-2013-0166][]
7299 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7300 the right response is stapled. Also change SSL_get_certificate()
7301 so it returns the certificate actually sent.
7302 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7303 (This is a backport)
7305 *Rob Stradling <rob.stradling@comodo.com>*
7307 * Fix possible deadlock when decoding public keys.
7311 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7313 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7316 * Sanity check record length before skipping explicit IV in DTLS
7319 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7320 fuzzing as a service testing platform.
7325 * Initialise tkeylen properly when encrypting CMS messages.
7326 Thanks to Solar Designer of Openwall for reporting this issue.
7330 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7332 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7333 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7334 in CRYPTO_realloc_clean.
7336 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7337 issue and to Adam Langley <agl@chromium.org> for fixing it.
7340 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7342 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7344 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7345 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7346 content decryption and always return the same error. Note: this attack
7347 needs on average 2^20 messages so it only affects automated senders. The
7348 old behaviour can be re-enabled in the CMS code by setting the
7349 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7350 an MMA defence is not necessary.
7351 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7352 this issue. [CVE-2012-0884][]
7356 * Fix CVE-2011-4619: make sure we really are receiving a
7357 client hello before rejecting multiple SGC restarts. Thanks to
7358 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7362 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7364 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7365 Thanks to Antonio Martin, Enterprise Secure Access Research and
7366 Development, Cisco Systems, Inc. for discovering this bug and
7367 preparing a fix. [CVE-2012-0050][]
7371 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7373 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7374 of the Vaudenay padding oracle attack on CBC mode encryption
7375 which enables an efficient plaintext recovery attack against
7376 the OpenSSL implementation of DTLS. Their attack exploits timing
7377 differences arising during decryption processing. A research
7378 paper describing this attack can be found at:
7379 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7380 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7381 Security Group at Royal Holloway, University of London
7382 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7383 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7384 for preparing the fix. [CVE-2011-4108][]
7386 *Robin Seggelmann, Michael Tuexen*
7388 * Clear bytes used for block padding of SSL 3.0 records.
7391 *Adam Langley (Google)*
7393 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7394 Kadianakis <desnacked@gmail.com> for discovering this issue and
7395 Adam Langley for preparing the fix. [CVE-2011-4619][]
7397 *Adam Langley (Google)*
7399 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7401 *Andrey Kulikov <amdeich@gmail.com>*
7403 * Prevent malformed RFC3779 data triggering an assertion failure.
7404 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7405 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7407 *Rob Austein <sra@hactrn.net>*
7409 * Improved PRNG seeding for VOS.
7411 *Paul Green <Paul.Green@stratus.com>*
7413 * Fix ssl_ciph.c set-up race.
7415 *Adam Langley (Google)*
7417 * Fix spurious failures in ecdsatest.c.
7419 *Emilia Käsper (Google)*
7421 * Fix the BIO_f_buffer() implementation (which was mixing different
7422 interpretations of the `..._len` fields).
7424 *Adam Langley (Google)*
7426 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7427 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7428 threads won't reuse the same blinding coefficients.
7430 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7431 lock to call BN_BLINDING_invert_ex, and avoids one use of
7432 BN_BLINDING_update for each BN_BLINDING structure (previously,
7433 the last update always remained unused).
7435 *Emilia Käsper (Google)*
7437 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7439 *Bob Buckholz (Google)*
7441 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7443 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7444 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7446 *Kaspar Brand <ossl@velox.ch>*
7448 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7449 for multi-threaded use of ECDH. [CVE-2011-3210][]
7451 *Adam Langley (Google)*
7453 * Fix x509_name_ex_d2i memory leak on bad inputs.
7457 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7458 signature public key algorithm by using OID xref utilities instead.
7459 Before this you could only use some ECC ciphersuites with SHA1 only.
7463 * Add protection against ECDSA timing attacks as mentioned in the paper
7464 by Billy Bob Brumley and Nicola Tuveri, see:
7465 <http://eprint.iacr.org/2011/232.pdf>
7467 *Billy Bob Brumley and Nicola Tuveri*
7469 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7471 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7473 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7475 * Fix bug in string printing code: if *any* escaping is enabled we must
7476 escape the escape character (backslash) or the resulting string is
7481 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7483 * Disable code workaround for ancient and obsolete Netscape browsers
7484 and servers: an attacker can use it in a ciphersuite downgrade attack.
7485 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7489 * Fixed J-PAKE implementation error, originally discovered by
7490 Sebastien Martini, further info and confirmation from Stefan
7491 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7495 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7497 * Fix extension code to avoid race conditions which can result in a buffer
7498 overrun vulnerability: resumed sessions must not be modified as they can
7499 be shared by multiple threads. CVE-2010-3864
7503 * Fix WIN32 build system to correctly link an ENGINE directory into
7508 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7510 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7513 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7515 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7517 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7518 context. The operation can be customised via the ctrl mechanism in
7519 case ENGINEs want to include additional functionality.
7523 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7527 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7528 output hashes compatible with older versions of OpenSSL.
7530 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7532 * Fix compression algorithm handling: if resuming a session use the
7533 compression algorithm of the resumed session instead of determining
7534 it from client hello again. Don't allow server to change algorithm.
7538 * Add load_crls() function to apps tidying load_certs() too. Add option
7539 to verify utility to allow additional CRLs to be included.
7543 * Update OCSP request code to permit adding custom headers to the request:
7544 some responders need this.
7548 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7551 *Julia Lawall <julia@diku.dk>*
7553 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7554 needlessly dereferenced structures, used obsolete functions and
7555 didn't handle all updated verify codes correctly.
7559 * Disable MD2 in the default configuration.
7563 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7564 indicate the initial BIO being pushed or popped. This makes it possible
7565 to determine whether the BIO is the one explicitly called or as a result
7566 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7567 it handles reference counts correctly and doesn't zero out the I/O bio
7568 when it is not being explicitly popped. WARNING: applications which
7569 included workarounds for the old buggy behaviour will need to be modified
7570 or they could free up already freed BIOs.
7574 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7575 renaming to all platforms (within the 0.9.8 branch, this was
7576 done conditionally on Netware platforms to avoid a name clash).
7578 *Guenter <lists@gknw.net>*
7580 * Add ECDHE and PSK support to DTLS.
7582 *Michael Tuexen <tuexen@fh-muenster.de>*
7584 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7589 * Add "missing" function EVP_MD_flags() (without this the only way to
7590 retrieve a digest flags is by accessing the structure directly. Update
7591 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7592 or cipher is registered as in the "from" argument. Print out all
7593 registered digests in the dgst usage message instead of manually
7594 attempting to work them out.
7598 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7599 this allows the use of compression and extensions. Change default cipher
7600 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7601 by default unless an application cipher string requests it.
7605 * Alter match criteria in PKCS12_parse(). It used to try to use local
7606 key ids to find matching certificates and keys but some PKCS#12 files
7607 don't follow the (somewhat unwritten) rules and this strategy fails.
7608 Now just gather all certificates together and the first private key
7609 then look for the first certificate that matches the key.
7613 * Support use of registered digest and cipher names for dgst and cipher
7614 commands instead of having to add each one as a special case. So now
7621 openssl dgst -sha256 foo
7623 and this works for ENGINE based algorithms too.
7627 * Update Gost ENGINE to support parameter files.
7629 *Victor B. Wagner <vitus@cryptocom.ru>*
7631 * Support GeneralizedTime in ca utility.
7633 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7635 * Enhance the hash format used for certificate directory links. The new
7636 form uses the canonical encoding (meaning equivalent names will work
7637 even if they aren't identical) and uses SHA1 instead of MD5. This form
7638 is incompatible with the older format and as a result c_rehash should
7639 be used to rebuild symbolic links.
7643 * Make PKCS#8 the default write format for private keys, replacing the
7644 traditional format. This form is standardised, more secure and doesn't
7645 include an implicit MD5 dependency.
7649 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7650 committed to OpenSSL should pass this lot as a minimum.
7654 * Add session ticket override functionality for use by EAP-FAST.
7656 *Jouni Malinen <j@w1.fi>*
7658 * Modify HMAC functions to return a value. Since these can be implemented
7659 in an ENGINE errors can occur.
7663 * Type-checked OBJ_bsearch_ex.
7667 * Type-checked OBJ_bsearch. Also some constification necessitated
7668 by type-checking. Still to come: TXT_DB, bsearch(?),
7669 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7674 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7675 seconds to a tm structure directly, instead of going through OS
7676 specific date routines. This avoids any issues with OS routines such
7677 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7678 and X509_time_adj_ex() to cover the extended range. The existing
7679 X509_time_adj() is still usable and will no longer have any date issues.
7683 * Delta CRL support. New use deltas option which will attempt to locate
7684 and search any appropriate delta CRLs available.
7686 This work was sponsored by Google.
7690 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7691 code and add additional score elements. Validate alternate CRL paths
7692 as part of the CRL checking and indicate a new error "CRL path validation
7693 error" in this case. Applications wanting additional details can use
7694 the verify callback and check the new "parent" field. If this is not
7695 NULL CRL path validation is taking place. Existing applications won't
7696 see this because it requires extended CRL support which is off by
7699 This work was sponsored by Google.
7703 * Support for freshest CRL extension.
7705 This work was sponsored by Google.
7709 * Initial indirect CRL support. Currently only supported in the CRLs
7710 passed directly and not via lookup. Process certificate issuer
7711 CRL entry extension and lookup CRL entries by bother issuer name
7712 and serial number. Check and process CRL issuer entry in IDP extension.
7714 This work was sponsored by Google.
7718 * Add support for distinct certificate and CRL paths. The CRL issuer
7719 certificate is validated separately in this case. Only enabled if
7720 an extended CRL support flag is set: this flag will enable additional
7721 CRL functionality in future.
7723 This work was sponsored by Google.
7727 * Add support for policy mappings extension.
7729 This work was sponsored by Google.
7733 * Fixes to pathlength constraint, self issued certificate handling,
7734 policy processing to align with RFC3280 and PKITS tests.
7736 This work was sponsored by Google.
7740 * Support for name constraints certificate extension. DN, email, DNS
7741 and URI types are currently supported.
7743 This work was sponsored by Google.
7747 * To cater for systems that provide a pointer-based thread ID rather
7748 than numeric, deprecate the current numeric thread ID mechanism and
7749 replace it with a structure and associated callback type. This
7750 mechanism allows a numeric "hash" to be extracted from a thread ID in
7751 either case, and on platforms where pointers are larger than 'long',
7752 mixing is done to help ensure the numeric 'hash' is usable even if it
7753 can't be guaranteed unique. The default mechanism is to use "&errno"
7754 as a pointer-based thread ID to distinguish between threads.
7756 Applications that want to provide their own thread IDs should now use
7757 CRYPTO_THREADID_set_callback() to register a callback that will call
7758 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7760 Note that ERR_remove_state() is now deprecated, because it is tied
7761 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7762 to free the current thread's error state should be replaced by
7763 ERR_remove_thread_state(NULL).
7765 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7766 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7767 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7768 application was previously providing a numeric thread callback that
7769 was inappropriate for distinguishing threads, then uniqueness might
7770 have been obtained with &errno that happened immediately in the
7771 intermediate development versions of OpenSSL; this is no longer the
7772 case, the numeric thread callback will now override the automatic use
7775 *Geoff Thorpe, with help from Bodo Moeller*
7777 * Initial support for different CRL issuing certificates. This covers a
7778 simple case where the self issued certificates in the chain exist and
7779 the real CRL issuer is higher in the existing chain.
7781 This work was sponsored by Google.
7785 * Removed effectively defunct crypto/store from the build.
7789 * Revamp of STACK to provide stronger type-checking. Still to come:
7790 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7791 ASN1_STRING, CONF_VALUE.
7795 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7796 RAM on SSL connections. This option can save about 34k per idle SSL.
7800 * Revamp of LHASH to provide stronger type-checking. Still to come:
7801 STACK, TXT_DB, bsearch, qsort.
7805 * Initial support for Cryptographic Message Syntax (aka CMS) based
7806 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7807 support for data, signedData, compressedData, digestedData and
7808 encryptedData, envelopedData types included. Scripts to check against
7809 RFC4134 examples draft and interop and consistency checks of many
7810 content types and variants.
7814 * Add options to enc utility to support use of zlib compression BIO.
7818 * Extend mk1mf to support importing of options and assembly language
7819 files from Configure script, currently only included in VC-WIN32.
7820 The assembly language rules can now optionally generate the source
7821 files from the associated perl scripts.
7825 * Implement remaining functionality needed to support GOST ciphersuites.
7826 Interop testing has been performed using CryptoPro implementations.
7828 *Victor B. Wagner <vitus@cryptocom.ru>*
7830 * s390x assembler pack.
7834 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7839 * Implement Opaque PRF Input TLS extension as specified in
7840 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7841 official specification yet and no extension type assignment by
7842 IANA exists, this extension (for now) will have to be explicitly
7843 enabled when building OpenSSL by providing the extension number
7844 to use. For example, specify an option
7846 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7848 to the "config" or "Configure" script to enable the extension,
7849 assuming extension number 0x9527 (which is a completely arbitrary
7850 and unofficial assignment based on the MD5 hash of the Internet
7851 Draft). Note that by doing so, you potentially lose
7852 interoperability with other TLS implementations since these might
7853 be using the same extension number for other purposes.
7855 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7856 opaque PRF input value to use in the handshake. This will create
7857 an internal copy of the length-'len' string at 'src', and will
7858 return non-zero for success.
7860 To get more control and flexibility, provide a callback function
7863 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7864 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7868 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7871 Callback function 'cb' will be called in handshakes, and is
7872 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7873 Argument 'arg' is for application purposes (the value as given to
7874 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7875 be provided to the callback function). The callback function
7876 has to return non-zero to report success: usually 1 to use opaque
7877 PRF input just if possible, or 2 to enforce use of the opaque PRF
7878 input. In the latter case, the library will abort the handshake
7879 if opaque PRF input is not successfully negotiated.
7881 Arguments 'peerinput' and 'len' given to the callback function
7882 will always be NULL and 0 in the case of a client. A server will
7883 see the client's opaque PRF input through these variables if
7884 available (NULL and 0 otherwise). Note that if the server
7885 provides an opaque PRF input, the length must be the same as the
7886 length of the client's opaque PRF input.
7888 Note that the callback function will only be called when creating
7889 a new session (session resumption can resume whatever was
7890 previously negotiated), and will not be called in SSL 2.0
7891 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7892 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7893 for applications that need to enforce opaque PRF input.
7897 * Update ssl code to support digests other than SHA1+MD5 for handshake
7900 *Victor B. Wagner <vitus@cryptocom.ru>*
7902 * Add RFC4507 support to OpenSSL. This includes the corrections in
7903 RFC4507bis. The encrypted ticket format is an encrypted encoded
7904 SSL_SESSION structure, that way new session features are automatically
7907 If a client application caches session in an SSL_SESSION structure
7908 support is transparent because tickets are now stored in the encoded
7911 The SSL_CTX structure automatically generates keys for ticket
7912 protection in servers so again support should be possible
7913 with no application modification.
7915 If a client or server wishes to disable RFC4507 support then the option
7916 SSL_OP_NO_TICKET can be set.
7918 Add a TLS extension debugging callback to allow the contents of any client
7919 or server extensions to be examined.
7921 This work was sponsored by Google.
7925 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7926 OpenSSL should now compile cleanly on gcc 4.2
7928 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7930 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7931 support including streaming MAC support: this is required for GOST
7932 ciphersuite support.
7934 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7936 * Add option -stream to use PKCS#7 streaming in smime utility. New
7937 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7938 to output in BER and PEM format.
7942 * Experimental support for use of HMAC via EVP_PKEY interface. This
7943 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
7944 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7945 ENGINE support for HMAC keys which are unextractable. New -mac and
7946 -macopt options to dgst utility.
7950 * New option -sigopt to dgst utility. Update dgst to use
7951 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
7952 alternative signing parameters such as X9.31 or PSS in the dgst
7957 * Change ssl_cipher_apply_rule(), the internal function that does
7958 the work each time a ciphersuite string requests enabling
7959 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7960 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7961 the order of disabled ciphersuites such that those ciphersuites
7962 that most recently went from enabled to disabled not only stay
7963 in order with respect to each other, but also have higher priority
7964 than other disabled ciphersuites the next time ciphersuites are
7967 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7968 the same ciphersuites as with "HIGH" alone, but in a specific
7969 order where the PSK ciphersuites come first (since they are the
7970 most recently disabled ciphersuites when "HIGH" is parsed).
7972 Also, change ssl_create_cipher_list() (using this new
7973 functionality) such that between otherwise identical
7974 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7979 * Change ssl_create_cipher_list() so that it automatically
7980 arranges the ciphersuites in reasonable order before starting
7981 to process the rule string. Thus, the definition for "DEFAULT"
7982 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7983 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7984 This makes it much easier to arrive at a reasonable default order
7985 in applications for which anonymous ciphers are OK (meaning
7986 that you can't actually use DEFAULT).
7988 *Bodo Moeller; suggested by Victor Duchovni*
7990 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7991 processing) into multiple integers instead of setting
7992 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7993 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7994 (These masks as well as the individual bit definitions are hidden
7995 away into the non-exported interface ssl/ssl_locl.h, so this
7996 change to the definition of the SSL_CIPHER structure shouldn't
7997 affect applications.) This give us more bits for each of these
7998 categories, so there is no longer a need to coagulate AES128 and
7999 AES256 into a single algorithm bit, and to coagulate Camellia128
8000 and Camellia256 into a single algorithm bit, which has led to all
8003 Thus, among other things, the kludge introduced in 0.9.7m and
8004 0.9.8e for masking out AES256 independently of AES128 or masking
8005 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8007 With the change, we also introduce new ciphersuite aliases that
8008 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8013 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8014 Use the leftmost N bytes of the signature input if the input is
8015 larger than the prime q (with N being the size in bytes of q).
8019 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8020 it yet and it is largely untested.
8024 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8028 * Initial incomplete changes to avoid need for function casts in OpenSSL
8029 some compilers (gcc 4.2 and later) reject their use. Safestack is
8030 reimplemented. Update ASN1 to avoid use of legacy functions.
8034 * Win32/64 targets are linked with Winsock2.
8038 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8039 to external functions. This can be used to increase CRL handling
8040 efficiency especially when CRLs are very large by (for example) storing
8041 the CRL revoked certificates in a database.
8045 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8046 new CRLs added to a directory can be used. New command line option
8047 -verify_return_error to s_client and s_server. This causes real errors
8048 to be returned by the verify callback instead of carrying on no matter
8049 what. This reflects the way a "real world" verify callback would behave.
8053 * GOST engine, supporting several GOST algorithms and public key formats.
8054 Kindly donated by Cryptocom.
8058 * Partial support for Issuing Distribution Point CRL extension. CRLs
8059 partitioned by DP are handled but no indirect CRL or reason partitioning
8060 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8061 selected via a scoring technique which handles IDP and AKID in CRLs.
8065 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8066 will ultimately be used for all verify operations: this will remove the
8067 X509_STORE dependency on certificate verification and allow alternative
8068 lookup methods. X509_STORE based implementations of these two callbacks.
8072 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8073 Modify get_crl() to find a valid (unexpired) CRL if possible.
8077 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8078 this would be called X509_CRL_cmp() but that name is already used by
8079 a function that just compares CRL issuer names. Cache several CRL
8080 extensions in X509_CRL structure and cache CRLDP in X509.
8084 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8085 this maps equivalent X509_NAME structures into a consistent structure.
8086 Name comparison can then be performed rapidly using memcmp().
8090 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8095 * Allow digests to supply their own micalg string for S/MIME type using
8096 the ctrl EVP_MD_CTRL_MICALG.
8100 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8101 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8102 ctrl. It can then customise the structure before and/or after signing
8107 * New function OBJ_add_sigid() to allow application defined signature OIDs
8108 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8109 to free up any added signature OIDs.
8113 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8114 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8115 digest and cipher tables. New options added to openssl utility:
8116 list-message-digest-algorithms and list-cipher-algorithms.
8120 * Change the array representation of binary polynomials: the list
8121 of degrees of non-zero coefficients is now terminated with -1.
8122 Previously it was terminated with 0, which was also part of the
8123 value; thus, the array representation was not applicable to
8124 polynomials where t^0 has coefficient zero. This change makes
8125 the array representation useful in a more general context.
8129 * Various modifications and fixes to SSL/TLS cipher string
8130 handling. For ECC, the code now distinguishes between fixed ECDH
8131 with RSA certificates on the one hand and with ECDSA certificates
8132 on the other hand, since these are separate ciphersuites. The
8133 unused code for Fortezza ciphersuites has been removed.
8135 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8136 (not "ECDHE"). For consistency with the code for DH
8137 certificates, use of ECDH certificates is now considered ECDH
8138 authentication, not RSA or ECDSA authentication (the latter is
8139 merely the CA's signing algorithm and not actively used in the
8142 The temporary ciphersuite alias "ECCdraft" is no longer
8143 available, and ECC ciphersuites are no longer excluded from "ALL"
8144 and "DEFAULT". The following aliases now exist for RFC 4492
8145 ciphersuites, most of these by analogy with the DH case:
8147 kECDHr - ECDH cert, signed with RSA
8148 kECDHe - ECDH cert, signed with ECDSA
8149 kECDH - ECDH cert (signed with either RSA or ECDSA)
8150 kEECDH - ephemeral ECDH
8151 ECDH - ECDH cert or ephemeral ECDH
8157 AECDH - anonymous ECDH
8158 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8162 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8163 Use correct micalg parameters depending on digest(s) in signed message.
8167 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8168 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8172 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8173 an engine to register a method. Add ENGINE lookups for methods and
8174 functional reference processing.
8178 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8179 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8184 * New -resign option to smime utility. This adds one or more signers
8185 to an existing PKCS#7 signedData structure. Also -md option to use an
8186 alternative message digest algorithm for signing.
8190 * Tidy up PKCS#7 routines and add new functions to make it easier to
8191 create PKCS7 structures containing multiple signers. Update smime
8192 application to support multiple signers.
8196 * New -macalg option to pkcs12 utility to allow setting of an alternative
8201 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8202 Reorganize PBE internals to lookup from a static table using NIDs,
8203 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8204 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8205 PRF which will be automatically used with PBES2.
8209 * Replace the algorithm specific calls to generate keys in "req" with the
8214 * Update PKCS#7 enveloped data routines to use new API. This is now
8215 supported by any public key method supporting the encrypt operation. A
8216 ctrl is added to allow the public key algorithm to examine or modify
8217 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8222 * Add a ctrl to asn1 method to allow a public key algorithm to express
8223 a default digest type to use. In most cases this will be SHA1 but some
8224 algorithms (such as GOST) need to specify an alternative digest. The
8225 return value indicates how strong the preference is 1 means optional and
8226 2 is mandatory (that is it is the only supported type). Modify
8227 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8228 use the default md. Update openssl utilities to use the default digest
8229 type for signing if it is not explicitly indicated.
8233 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8234 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8235 signing method from the key type. This effectively removes the link
8236 between digests and public key types.
8240 * Add an OID cross reference table and utility functions. Its purpose is to
8241 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8242 rsaEncryption. This will allow some of the algorithm specific hackery
8243 needed to use the correct OID to be removed.
8247 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8248 structures for PKCS7_sign(). They are now set up by the relevant public
8253 * Add provisional EC pkey method with support for ECDSA and ECDH.
8257 * Add support for key derivation (agreement) in the API, DH method and
8262 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8263 public and private key formats. As a side effect these add additional
8264 command line functionality not previously available: DSA signatures can be
8265 generated and verified using pkeyutl and DH key support and generation in
8272 *Oliver Tappe <zooey@hirschkaefer.de>*
8274 * New make target "install_html_docs" installs HTML renditions of the
8277 *Oliver Tappe <zooey@hirschkaefer.de>*
8279 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8280 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8281 support key and parameter generation and add initial key generation
8282 functionality for RSA.
8286 * Add functions for main EVP_PKEY_method operations. The undocumented
8287 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8288 `EVP_PKEY_{encrypt,decrypt}_old`.
8292 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8293 key API, doesn't do much yet.
8297 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8298 public key algorithms. New option to openssl utility:
8299 "list-public-key-algorithms" to print out info.
8303 * Implement the Supported Elliptic Curves Extension for
8304 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8308 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8309 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8313 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8314 utilities such as rsa, dsa, dsaparam etc except they process any key
8319 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8320 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8321 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8326 * Initial support for pluggable public key ASN1.
8327 De-spaghettify the public key ASN1 handling. Move public and private
8328 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8329 algorithm specific handling to a single module within the relevant
8330 algorithm directory. Add functions to allow (near) opaque processing
8331 of public and private key structures.
8335 * Implement the Supported Point Formats Extension for
8336 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8340 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8341 for the psk identity [hint] and the psk callback functions to the
8342 SSL_SESSION, SSL and SSL_CTX structure.
8345 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8349 SSL_CTX_use_psk_identity_hint
8350 SSL_get_psk_identity_hint
8351 SSL_get_psk_identity
8352 SSL_use_psk_identity_hint
8354 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8356 * Add RFC 3161 compliant time stamp request creation, response generation
8357 and response verification functionality.
8359 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8361 * Add initial support for TLS extensions, specifically for the server_name
8362 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8363 have new members for a host name. The SSL data structure has an
8364 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8365 stored in that context to allow for session resumption, even after the
8366 SSL has been switched to a new SSL_CTX in reaction to a client's
8367 server_name extension.
8369 New functions (subject to change):
8371 SSL_get_servername()
8372 SSL_get_servername_type()
8375 New CTRL codes and macros (subject to change):
8377 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8378 - SSL_CTX_set_tlsext_servername_callback()
8379 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8380 - SSL_CTX_set_tlsext_servername_arg()
8381 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8383 openssl s_client has a new '-servername ...' option.
8385 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8386 '-key2 ...', '-servername_fatal' (subject to change). This allows
8387 testing the HostName extension for a specific single host name ('-cert'
8388 and '-key' remain fallbacks for handshakes without HostName
8389 negotiation). If the unrecognized_name alert has to be sent, this by
8390 default is a warning; it becomes fatal with the '-servername_fatal'
8393 *Peter Sylvester, Remy Allais, Christophe Renou*
8395 * Whirlpool hash implementation is added.
8399 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8400 bn(64,32). Because of instruction set limitations it doesn't have
8401 any negative impact on performance. This was done mostly in order
8402 to make it possible to share assembler modules, such as bn_mul_mont
8403 implementations, between 32- and 64-bit builds without hassle.
8407 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8408 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8413 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8414 dedicated Montgomery multiplication procedure, is introduced.
8415 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8416 "64-bit" performance on certain 32-bit targets.
8420 * New option SSL_OP_NO_COMP to disable use of compression selectively
8421 in SSL structures. New SSL ctrl to set maximum send fragment size.
8422 Save memory by setting the I/O buffer sizes dynamically instead of
8423 using the maximum available value.
8427 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8428 in addition to the text details.
8432 * Very, very preliminary EXPERIMENTAL support for printing of general
8433 ASN1 structures. This currently produces rather ugly output and doesn't
8434 handle several customised structures at all.
8438 * Integrated support for PVK file format and some related formats such
8439 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8440 these in the 'rsa' and 'dsa' utilities.
8444 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8448 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8449 place for the (very old) "NETSCAPE" format certificates which are now
8450 handled using new ASN1 code equivalents.
8454 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8455 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8456 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8460 * Modify CRL distribution points extension code to print out previously
8461 unsupported fields. Enhance extension setting code to allow setting of
8466 * Add print and set support for Issuing Distribution Point CRL extension.
8470 * Change 'Configure' script to enable Camellia by default.
8477 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8479 * When rejecting SSL/TLS records due to an incorrect version number, never
8480 update s->server with a new major version number. As of
8481 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8482 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8483 the previous behavior could result in a read attempt at NULL when
8484 receiving specific incorrect SSL/TLS records once record payload
8485 protection is active. [CVE-2010-0740][]
8487 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8489 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8490 could be crashed if the relevant tables were not present (e.g. chrooted).
8492 *Tomas Hoger <thoger@redhat.com>*
8494 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8496 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8498 *Martin Olsson, Neel Mehta*
8500 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8501 accommodate for stack sorting, always a write lock!).
8505 * On some versions of WIN32 Heap32Next is very slow. This can cause
8506 excessive delays in the RAND_poll(): over a minute. As a workaround
8507 include a time check in the inner Heap32Next loop too.
8511 * The code that handled flushing of data in SSL/TLS originally used the
8512 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8513 the problem outlined in PR#1949. The fix suggested there however can
8514 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8515 of Apache). So instead simplify the code to flush unconditionally.
8516 This should be fine since flushing with no data to flush is a no op.
8520 * Handle TLS versions 2.0 and later properly and correctly use the
8521 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8522 off ancient servers have a habit of sticking around for a while...
8526 * Modify compression code so it frees up structures without using the
8527 ex_data callbacks. This works around a problem where some applications
8528 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8529 restarting) then use compression (e.g. SSL with compression) later.
8530 This results in significant per-connection memory leaks and
8531 has caused some security issues including CVE-2008-1678 and
8536 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8537 change when encrypting or decrypting.
8541 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8542 connect and renegotiate with servers which do not support RI.
8543 Until RI is more widely deployed this option is enabled by default.
8547 * Add "missing" ssl ctrls to clear options and mode.
8551 * If client attempts to renegotiate and doesn't support RI respond with
8552 a no_renegotiation alert as required by RFC5746. Some renegotiating
8553 TLS clients will continue a connection gracefully when they receive
8554 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8555 waiting for a server hello which it will never receive. Now we treat a
8556 received no_renegotiation alert as a fatal error. This is because
8557 applications requesting a renegotiation might well expect it to succeed
8558 and would have no code in place to handle the server denying it so the
8559 only safe thing to do is to terminate the connection.
8563 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8564 peer supports secure renegotiation and 0 otherwise. Print out peer
8565 renegotiation support in s_client/s_server.
8569 * Replace the highly broken and deprecated SPKAC certification method with
8570 the updated NID creation version. This should correctly handle UTF8.
8574 * Implement RFC5746. Re-enable renegotiation but require the extension
8575 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8576 turns out to be a bad idea. It has been replaced by
8577 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8578 SSL_CTX_set_options(). This is really not recommended unless you
8579 know what you are doing.
8581 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8583 * Fixes to stateless session resumption handling. Use initial_ctx when
8584 issuing and attempting to decrypt tickets in case it has changed during
8585 servername handling. Use a non-zero length session ID when attempting
8586 stateless session resumption: this makes it possible to determine if
8587 a resumption has occurred immediately after receiving server hello
8588 (several places in OpenSSL subtly assume this) instead of later in
8593 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8594 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8595 fixes for a few places where the return code is not checked
8598 *Julia Lawall <julia@diku.dk>*
8600 * Add --strict-warnings option to Configure script to include devteam
8601 warnings in other configurations.
8605 * Add support for --libdir option and LIBDIR variable in makefiles. This
8606 makes it possible to install openssl libraries in locations which
8607 have names other than "lib", for example "/usr/lib64" which some
8610 *Steve Henson, based on patch from Jeremy Utley*
8612 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8613 X690 8.9.12 and can produce some misleading textual output of OIDs.
8615 *Steve Henson, reported by Dan Kaminsky*
8617 * Delete MD2 from algorithm tables. This follows the recommendation in
8618 several standards that it is not used in new applications due to
8619 several cryptographic weaknesses. For binary compatibility reasons
8620 the MD2 API is still compiled in by default.
8624 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8629 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8630 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8633 *Guenter <lists@gknw.net>*
8635 * Fix the server certificate chain building code to use X509_verify_cert(),
8636 it used to have an ad-hoc builder which was unable to cope with anything
8637 other than a simple chain.
8639 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8641 * Don't check self signed certificate signatures in X509_verify_cert()
8642 by default (a flag can override this): it just wastes time without
8643 adding any security. As a useful side effect self signed root CAs
8644 with non-FIPS digests are now usable in FIPS mode.
8648 * In dtls1_process_out_of_seq_message() the check if the current message
8649 is already buffered was missing. For every new message was memory
8650 allocated, allowing an attacker to perform an denial of service attack
8651 with sending out of seq handshake messages until there is no memory
8652 left. Additionally every future message was buffered, even if the
8653 sequence number made no sense and would be part of another handshake.
8654 So only messages with sequence numbers less than 10 in advance will be
8655 buffered. [CVE-2009-1378][]
8657 *Robin Seggelmann, discovered by Daniel Mentz*
8659 * Records are buffered if they arrive with a future epoch to be
8660 processed after finishing the corresponding handshake. There is
8661 currently no limitation to this buffer allowing an attacker to perform
8662 a DOS attack with sending records with future epochs until there is no
8663 memory left. This patch adds the pqueue_size() function to determine
8664 the size of a buffer and limits the record buffer to 100 entries.
8667 *Robin Seggelmann, discovered by Daniel Mentz*
8669 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8670 parent structure is freed. [CVE-2009-1379][]
8674 * Handle non-blocking I/O properly in SSL_shutdown() call.
8676 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8678 * Add `2.5.4.*` OIDs
8680 *Ilya O. <vrghost@gmail.com>*
8682 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8684 * Disable renegotiation completely - this fixes a severe security
8685 problem [CVE-2009-3555][] at the cost of breaking all
8686 renegotiation. Renegotiation can be re-enabled by setting
8687 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8688 run-time. This is really not recommended unless you know what
8693 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8695 * Don't set val to NULL when freeing up structures, it is freed up by
8696 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8697 zeroing past the valid field. [CVE-2009-0789][]
8699 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8701 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8702 checked correctly. This would allow some invalid signed attributes to
8703 appear to verify correctly. [CVE-2009-0591][]
8705 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8707 * Reject UniversalString and BMPString types with invalid lengths. This
8708 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8709 a legal length. [CVE-2009-0590][]
8713 * Set S/MIME signing as the default purpose rather than setting it
8714 unconditionally. This allows applications to override it at the store
8719 * Permit restricted recursion of ASN1 strings. This is needed in practice
8720 to handle some structures.
8724 * Improve efficiency of mem_gets: don't search whole buffer each time
8727 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8729 * New -hex option for openssl rand.
8733 * Print out UTF8String and NumericString when parsing ASN1.
8737 * Support NumericString type for name components.
8741 * Allow CC in the environment to override the automatically chosen
8742 compiler. Note that nothing is done to ensure flags work with the
8747 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8749 * Properly check EVP_VerifyFinal() and similar return values
8752 *Ben Laurie, Bodo Moeller, Google Security Team*
8754 * Enable TLS extensions by default.
8758 * Allow the CHIL engine to be loaded, whether the application is
8759 multithreaded or not. (This does not release the developer from the
8760 obligation to set up the dynamic locking callbacks.)
8762 *Sander Temme <sander@temme.net>*
8764 * Use correct exit code if there is an error in dgst command.
8766 *Steve Henson; problem pointed out by Roland Dirlewanger*
8768 * Tweak Configure so that you need to say "experimental-jpake" to enable
8769 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8773 * Add experimental JPAKE support, including demo authentication in
8774 s_client and s_server.
8778 * Set the comparison function in v3_addr_canonize().
8780 *Rob Austein <sra@hactrn.net>*
8782 * Add support for XMPP STARTTLS in s_client.
8784 *Philip Paeps <philip@freebsd.org>*
8786 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8787 to ensure that even with this option, only ciphersuites in the
8788 server's preference list will be accepted. (Note that the option
8789 applies only when resuming a session, so the earlier behavior was
8790 just about the algorithm choice for symmetric cryptography.)
8794 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8796 * Fix NULL pointer dereference if a DTLS server received
8797 ChangeCipherSpec as first record [CVE-2009-1386][].
8801 * Fix a state transition in s3_srvr.c and d1_srvr.c
8802 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8806 * The fix in 0.9.8c that supposedly got rid of unsafe
8807 double-checked locking was incomplete for RSA blinding,
8808 addressing just one layer of what turns out to have been
8809 doubly unsafe triple-checked locking.
8811 So now fix this for real by retiring the MONT_HELPER macro
8812 in crypto/rsa/rsa_eay.c.
8814 *Bodo Moeller; problem pointed out by Marius Schilder*
8816 * Various precautionary measures:
8818 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8820 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8821 (NB: This would require knowledge of the secret session ticket key
8822 to exploit, in which case you'd be SOL either way.)
8824 - Change bn_nist.c so that it will properly handle input BIGNUMs
8825 outside the expected range.
8827 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8830 *Neel Mehta, Bodo Moeller*
8832 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8833 the load fails. Useful for distros.
8835 *Ben Laurie and the FreeBSD team*
8837 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8841 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8845 * Expand ENGINE to support engine supplied SSL client certificate functions.
8847 This work was sponsored by Logica.
8851 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8852 keystores. Support for SSL/TLS client authentication too.
8853 Not compiled unless enable-capieng specified to Configure.
8855 This work was sponsored by Logica.
8859 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8860 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8861 attribute creation routines such as certificate requests and PKCS#12
8866 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8868 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8869 handshake which could lead to a client crash as found using the
8870 Codenomicon TLS test suite [CVE-2008-1672][]
8872 *Steve Henson, Mark Cox*
8874 * Fix double free in TLS server name extensions which could lead to
8875 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8879 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8881 Clear the error queue to ensure that error entries left from
8882 older function calls do not interfere with the correct operation.
8884 *Lutz Jaenicke, Erik de Castro Lopo*
8886 * Remove root CA certificates of commercial CAs:
8888 The OpenSSL project does not recommend any specific CA and does not
8889 have any policy with respect to including or excluding any CA.
8890 Therefore it does not make any sense to ship an arbitrary selection
8891 of root CA certificates with the OpenSSL software.
8895 * RSA OAEP patches to fix two separate invalid memory reads.
8896 The first one involves inputs when 'lzero' is greater than
8897 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8898 before the beginning of from). The second one involves inputs where
8899 the 'db' section contains nothing but zeroes (there is a one-byte
8900 invalid read after the end of 'db').
8902 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8904 * Partial backport from 0.9.9-dev:
8906 Introduce bn_mul_mont (dedicated Montgomery multiplication
8907 procedure) as a candidate for BIGNUM assembler implementation.
8908 While 0.9.9-dev uses assembler for various architectures, only
8909 x86_64 is available by default here in the 0.9.8 branch, and
8910 32-bit x86 is available through a compile-time setting.
8912 To try the 32-bit x86 assembler implementation, use Configure
8913 option "enable-montasm" (which exists only for this backport).
8915 As "enable-montasm" for 32-bit x86 disclaims code stability
8916 anyway, in this constellation we activate additional code
8917 backported from 0.9.9-dev for further performance improvements,
8918 namely BN_from_montgomery_word. (To enable this otherwise,
8919 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8921 *Andy Polyakov (backport partially by Bodo Moeller)*
8923 * Add TLS session ticket callback. This allows an application to set
8924 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8925 values. This is useful for key rollover for example where several key
8926 sets may exist with different names.
8930 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8931 This was broken until now in 0.9.8 releases, such that the only way
8932 a registered ENGINE could be used (assuming it initialises
8933 successfully on the host) was to explicitly set it as the default
8934 for the relevant algorithms. This is in contradiction with 0.9.7
8935 behaviour and the documentation. With this fix, when an ENGINE is
8936 registered into a given algorithm's table of implementations, the
8937 'uptodate' flag is reset so that auto-discovery will be used next
8938 time a new context for that algorithm attempts to select an
8941 *Ian Lister (tweaked by Geoff Thorpe)*
8943 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8944 implementation in the following ways:
8946 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8949 Lack of BER streaming support means one pass streaming processing is
8950 only supported if data is detached: setting the streaming flag is
8951 ignored for embedded content.
8953 CMS support is disabled by default and must be explicitly enabled
8954 with the enable-cms configuration option.
8958 * Update the GMP engine glue to do direct copies between BIGNUM and
8959 mpz_t when openssl and GMP use the same limb size. Otherwise the
8960 existing "conversion via a text string export" trick is still used.
8962 *Paul Sheer <paulsheer@gmail.com>*
8964 * Zlib compression BIO. This is a filter BIO which compressed and
8965 uncompresses any data passed through it.
8969 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8970 RFC3394 compatible AES key wrapping.
8974 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8975 sets string data without copying. X509_ALGOR_set0() and
8976 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8977 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8978 from an X509_ATTRIBUTE structure optionally checking it occurs only
8979 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8984 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8985 to get the expected BN_FLG_CONSTTIME behavior.
8987 *Bodo Moeller (Google)*
8991 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8992 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8993 - added some more tests to do_tests.pl
8994 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8995 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8996 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8997 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8998 - various changes to netware.pl to enable gcc-cross builds on Win32
9000 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9001 - various changes to fix missing prototype warnings
9002 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9003 - added AES, WHIRLPOOL and CPUID assembler code to build files
9004 - added missing AES assembler make rules to mk1mf.pl
9005 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9007 *Guenter Knauf <eflash@gmx.net>*
9009 * Implement certificate status request TLS extension defined in RFC3546.
9010 A client can set the appropriate parameters and receive the encoded
9011 OCSP response via a callback. A server can query the supplied parameters
9012 and set the encoded OCSP response in the callback. Add simplified examples
9013 to s_client and s_server.
9017 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9020 + Binary incompatibility of ssl_ctx_st structure
9021 + DTLS interoperation with non-compliant servers
9022 + Don't call get_session_cb() without proposed session
9023 + Fix ia64 assembler code
9025 *Andy Polyakov, Steve Henson*
9027 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9029 * DTLS Handshake overhaul. There were longstanding issues with
9030 OpenSSL DTLS implementation, which were making it impossible for
9031 RFC 4347 compliant client to communicate with OpenSSL server.
9032 Unfortunately just fixing these incompatibilities would "cut off"
9033 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9034 server keeps tolerating non RFC compliant syntax. The opposite is
9035 not true, 0.9.8f client can not communicate with earlier server.
9036 This update even addresses CVE-2007-4995.
9040 * Changes to avoid need for function casts in OpenSSL: some compilers
9041 (gcc 4.2 and later) reject their use.
9042 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9045 * Add RFC4507 support to OpenSSL. This includes the corrections in
9046 RFC4507bis. The encrypted ticket format is an encrypted encoded
9047 SSL_SESSION structure, that way new session features are automatically
9050 If a client application caches session in an SSL_SESSION structure
9051 support is transparent because tickets are now stored in the encoded
9054 The SSL_CTX structure automatically generates keys for ticket
9055 protection in servers so again support should be possible
9056 with no application modification.
9058 If a client or server wishes to disable RFC4507 support then the option
9059 SSL_OP_NO_TICKET can be set.
9061 Add a TLS extension debugging callback to allow the contents of any client
9062 or server extensions to be examined.
9064 This work was sponsored by Google.
9068 * Add initial support for TLS extensions, specifically for the server_name
9069 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9070 have new members for a host name. The SSL data structure has an
9071 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9072 stored in that context to allow for session resumption, even after the
9073 SSL has been switched to a new SSL_CTX in reaction to a client's
9074 server_name extension.
9076 New functions (subject to change):
9078 SSL_get_servername()
9079 SSL_get_servername_type()
9082 New CTRL codes and macros (subject to change):
9084 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9085 - SSL_CTX_set_tlsext_servername_callback()
9086 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9087 - SSL_CTX_set_tlsext_servername_arg()
9088 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9090 openssl s_client has a new '-servername ...' option.
9092 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9093 '-key2 ...', '-servername_fatal' (subject to change). This allows
9094 testing the HostName extension for a specific single host name ('-cert'
9095 and '-key' remain fallbacks for handshakes without HostName
9096 negotiation). If the unrecognized_name alert has to be sent, this by
9097 default is a warning; it becomes fatal with the '-servername_fatal'
9100 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9102 * Add AES and SSE2 assembly language support to VC++ build.
9106 * Mitigate attack on final subtraction in Montgomery reduction.
9110 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9111 (which previously caused an internal error).
9115 * Squeeze another 10% out of IGE mode when in != out.
9119 * AES IGE mode speedup.
9121 *Dean Gaudet (Google)*
9123 * Add the Korean symmetric 128-bit cipher SEED (see
9124 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9125 add SEED ciphersuites from RFC 4162:
9127 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9128 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9129 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9130 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9132 To minimize changes between patchlevels in the OpenSSL 0.9.8
9133 series, SEED remains excluded from compilation unless OpenSSL
9134 is configured with 'enable-seed'.
9136 *KISA, Bodo Moeller*
9138 * Mitigate branch prediction attacks, which can be practical if a
9139 single processor is shared, allowing a spy process to extract
9140 information. For detailed background information, see
9141 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9142 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9143 and Necessary Software Countermeasures"). The core of the change
9144 are new versions BN_div_no_branch() and
9145 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9146 respectively, which are slower, but avoid the security-relevant
9147 conditional branches. These are automatically called by BN_div()
9148 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9149 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9150 remove a conditional branch.
9152 BN_FLG_CONSTTIME is the new name for the previous
9153 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9154 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9155 in the exponent causes BN_mod_exp_mont() to use the alternative
9156 implementation in BN_mod_exp_mont_consttime().) The old name
9157 remains as a deprecated alias.
9159 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9160 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9161 constant-time implementations for more than just exponentiation.
9162 Here too the old name is kept as a deprecated alias.
9164 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9165 the BN_BLINDING structure gets an independent copy of the
9166 modulus. This means that the previous `BIGNUM *m` argument to
9167 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9168 essentially becomes `const BIGNUM *m`, although we can't actually
9169 change this in the header file before 0.9.9. It allows
9170 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9171 enable BN_FLG_CONSTTIME.
9173 *Matthew D Wood (Intel Corp)*
9175 * In the SSL/TLS server implementation, be strict about session ID
9176 context matching (which matters if an application uses a single
9177 external cache for different purposes). Previously,
9178 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9179 set. This did ensure strict client verification, but meant that,
9180 with applications using a single external cache for quite
9181 different requirements, clients could circumvent ciphersuite
9182 restrictions for a given session ID context by starting a session
9183 in a different context.
9187 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9188 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9189 authentication-only ciphersuites.
9193 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9194 not complete and could lead to a possible single byte overflow
9195 [CVE-2007-5135][] [Ben Laurie]
9197 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9199 * Since AES128 and AES256 (and similarly Camellia128 and
9200 Camellia256) share a single mask bit in the logic of
9201 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9202 kludge to work properly if AES128 is available and AES256 isn't
9203 (or if Camellia128 is available and Camellia256 isn't).
9207 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9208 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9209 When a point or a seed is encoded in a BIT STRING, we need to
9210 prevent the removal of trailing zero bits to get the proper DER
9211 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9212 of a NamedBitList, for which trailing 0 bits need to be removed.)
9216 * Have SSL/TLS server implementation tolerate "mismatched" record
9217 protocol version while receiving ClientHello even if the
9218 ClientHello is fragmented. (The server can't insist on the
9219 particular protocol version it has chosen before the ServerHello
9220 message has informed the client about his choice.)
9224 * Add RFC 3779 support.
9226 *Rob Austein for ARIN, Ben Laurie*
9228 * Load error codes if they are not already present instead of using a
9229 static variable. This allows them to be cleanly unloaded and reloaded.
9230 Improve header file function name parsing.
9234 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9235 or CAPABILITY handshake as required by RFCs.
9239 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9241 * Introduce limits to prevent malicious keys being able to
9242 cause a denial of service. [CVE-2006-2940][]
9244 *Steve Henson, Bodo Moeller*
9246 * Fix ASN.1 parsing of certain invalid structures that can result
9247 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9249 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9250 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9252 * Fix SSL client code which could crash if connecting to a
9253 malicious SSLv2 server. [CVE-2006-4343][]
9255 *Tavis Ormandy and Will Drewry, Google Security Team*
9257 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9258 match only those. Before that, "AES256-SHA" would be interpreted
9259 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9260 the same strength classification in 0.9.7h) as we currently only
9261 have a single AES bit in the ciphersuite description bitmap.
9262 That change, however, also applied to ciphersuite strings such as
9263 "RC4-MD5" that intentionally matched multiple ciphersuites --
9264 namely, SSL 2.0 ciphersuites in addition to the more common ones
9265 from SSL 3.0/TLS 1.0.
9267 So we change the selection algorithm again: Naming an explicit
9268 ciphersuite selects this one ciphersuite, and any other similar
9269 ciphersuite (same bitmap) from *other* protocol versions.
9270 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9271 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9273 Since SSL 2.0 does not have any ciphersuites for which the
9274 128/256 bit distinction would be relevant, this works for now.
9275 The proper fix will be to use different bits for AES128 and
9276 AES256, which would have avoided the problems from the beginning;
9277 however, bits are scarce, so we can only do this in a new release
9278 (not just a patchlevel) when we can change the SSL_CIPHER
9279 definition to split the single 'unsigned long mask' bitmap into
9280 multiple values to extend the available space.
9284 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9286 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9287 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9289 * Add AES IGE and biIGE modes.
9293 * Change the Unix randomness entropy gathering to use poll() when
9294 possible instead of select(), since the latter has some
9295 undesirable limitations.
9297 *Darryl Miles via Richard Levitte and Bodo Moeller*
9299 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9300 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9301 cannot be implicitly activated as part of, e.g., the "AES" alias.
9302 However, please upgrade to OpenSSL 0.9.9[-dev] for
9303 non-experimental use of the ECC ciphersuites to get TLS extension
9304 support, which is required for curve and point format negotiation
9305 to avoid potential handshake problems.
9309 * Disable rogue ciphersuites:
9311 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9312 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9313 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9315 The latter two were purportedly from
9316 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9319 Also deactivate the remaining ciphersuites from
9320 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9321 unofficial, and the ID has long expired.
9325 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9326 dual-core machines) and other potential thread-safety issues.
9330 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9331 versions), which is now available for royalty-free use
9332 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9333 Also, add Camellia TLS ciphersuites from RFC 4132.
9335 To minimize changes between patchlevels in the OpenSSL 0.9.8
9336 series, Camellia remains excluded from compilation unless OpenSSL
9337 is configured with 'enable-camellia'.
9341 * Disable the padding bug check when compression is in use. The padding
9342 bug check assumes the first packet is of even length, this is not
9343 necessarily true if compression is enabled and can result in false
9344 positives causing handshake failure. The actual bug test is ancient
9345 code so it is hoped that implementations will either have fixed it by
9346 now or any which still have the bug do not support compression.
9350 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9352 * When applying a cipher rule check to see if string match is an explicit
9353 cipher suite and only match that one cipher suite if it is.
9357 * Link in manifests for VC++ if needed.
9359 *Austin Ziegler <halostatue@gmail.com>*
9361 * Update support for ECC-based TLS ciphersuites according to
9362 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9363 TLS extensions, which are supported starting with the 0.9.9
9364 branch, not in the OpenSSL 0.9.8 branch).
9368 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9369 opaque EVP_CIPHER_CTX handling.
9373 * Fixes and enhancements to zlib compression code. We now only use
9374 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9375 to conform with the standards mentioned here:
9376 <http://www.zlib.net/DLL_FAQ.txt>
9377 Static zlib linking now works on Windows and the new --with-zlib-include
9378 --with-zlib-lib options to Configure can be used to supply the location
9379 of the headers and library. Gracefully handle case where zlib library
9384 * Several fixes and enhancements to the OID generation code. The old code
9385 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9386 handle numbers larger than ULONG_MAX, truncated printing and had a
9387 non standard OBJ_obj2txt() behaviour.
9391 * Add support for building of engines under engine/ as shared libraries
9392 under VC++ build system.
9396 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9397 Hopefully, we will not see any false combination of paths any more.
9401 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9403 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9404 (part of SSL_OP_ALL). This option used to disable the
9405 countermeasure against man-in-the-middle protocol-version
9406 rollback in the SSL 2.0 server implementation, which is a bad
9407 idea. [CVE-2005-2969][]
9409 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9410 for Information Security, National Institute of Advanced Industrial
9411 Science and Technology [AIST], Japan)*
9413 * Add two function to clear and return the verify parameter flags.
9417 * Keep cipherlists sorted in the source instead of sorting them at
9418 runtime, thus removing the need for a lock.
9422 * Avoid some small subgroup attacks in Diffie-Hellman.
9424 *Nick Mathewson and Ben Laurie*
9426 * Add functions for well-known primes.
9430 * Extended Windows CE support.
9432 *Satoshi Nakamura and Andy Polyakov*
9434 * Initialize SSL_METHOD structures at compile time instead of during
9435 runtime, thus removing the need for a lock.
9439 * Make PKCS7_decrypt() work even if no certificate is supplied by
9440 attempting to decrypt each encrypted key in turn. Add support to
9445 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9447 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9450 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9454 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9455 key into the same file any more.
9459 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9463 * Add -utf8 command line and config file option to 'ca'.
9465 *Stefan <stf@udoma.org*
9467 * Removed the macro des_crypt(), as it seems to conflict with some
9468 libraries. Use DES_crypt().
9472 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9473 involves renaming the source and generated shared-libs for
9474 both. The engines will accept the corrected or legacy ids
9475 ('ncipher' and '4758_cca' respectively) when binding. NB,
9476 this only applies when building 'shared'.
9478 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9480 * Add attribute functions to EVP_PKEY structure. Modify
9481 PKCS12_create() to recognize a CSP name attribute and
9482 use it. Make -CSP option work again in pkcs12 utility.
9486 * Add new functionality to the bn blinding code:
9487 - automatic re-creation of the BN_BLINDING parameters after
9488 a fixed number of uses (currently 32)
9489 - add new function for parameter creation
9490 - introduce flags to control the update behaviour of the
9491 BN_BLINDING parameters
9492 - hide BN_BLINDING structure
9493 Add a second BN_BLINDING slot to the RSA structure to improve
9494 performance when a single RSA object is shared among several
9499 * Add support for DTLS.
9501 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9503 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9504 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9508 * Remove buggy and incomplete DH cert support from
9509 ssl/ssl_rsa.c and ssl/s3_both.c
9513 * Use SHA-1 instead of MD5 as the default digest algorithm for
9514 the apps/openssl applications.
9518 * Compile clean with "-Wall -Wmissing-prototypes
9519 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9520 DEBUG_SAFESTACK must also be set.
9524 * Change ./Configure so that certain algorithms can be disabled by default.
9525 The new counterpiece to "no-xxx" is "enable-xxx".
9527 The patented RC5 and MDC2 algorithms will now be disabled unless
9528 "enable-rc5" and "enable-mdc2", respectively, are specified.
9530 (IDEA remains enabled despite being patented. This is because IDEA
9531 is frequently required for interoperability, and there is no license
9532 fee for non-commercial use. As before, "no-idea" can be used to
9533 avoid this algorithm.)
9537 * Add processing of proxy certificates (see RFC 3820). This work was
9538 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9539 EGEE (Enabling Grids for E-science in Europe).
9543 * RC4 performance overhaul on modern architectures/implementations, such
9544 as Intel P4, IA-64 and AMD64.
9548 * New utility extract-section.pl. This can be used specify an alternative
9549 section number in a pod file instead of having to treat each file as
9550 a separate case in Makefile. This can be done by adding two lines to the
9553 =for comment openssl_section:XXX
9555 The blank line is mandatory.
9559 * New arguments -certform, -keyform and -pass for s_client and s_server
9560 to allow alternative format key and certificate files and passphrase
9565 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9566 update associated structures and add various utility functions.
9568 Add new policy related verify parameters, include policy checking in
9569 standard verify code. Enhance 'smime' application with extra parameters
9570 to support policy checking and print out.
9574 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9575 Nehemiah processors. These extensions support AES encryption in hardware
9576 as well as RNG (though RNG support is currently disabled).
9578 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9580 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9584 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9586 *Andy Polyakov and a number of other people*
9588 * Improved PowerPC platform support. Most notably BIGNUM assembler
9589 implementation contributed by IBM.
9591 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9593 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9594 exponent rather than 'unsigned long'. There is a corresponding change to
9595 the new 'rsa_keygen' element of the RSA_METHOD structure.
9597 *Jelte Jansen, Geoff Thorpe*
9599 * Functionality for creating the initial serial number file is now
9600 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9602 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9603 number file to 1, which is bound to cause problems. To avoid
9604 the problems while respecting compatibility between different 0.9.7
9605 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9606 CA.pl for serial number initialization. With the new release 0.9.8,
9607 we can fix the problem directly in the 'ca' utility.)
9611 * Reduced header interdependencies by declaring more opaque objects in
9612 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9613 give fewer recursive includes, which could break lazy source code - so
9614 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9615 developers should define this symbol when building and using openssl to
9616 ensure they track the recommended behaviour, interfaces, [etc], but
9617 backwards-compatible behaviour prevails when this isn't defined.
9621 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9625 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9626 This will generate a random key of the appropriate length based on the
9627 cipher context. The EVP_CIPHER can provide its own random key generation
9628 routine to support keys of a specific form. This is used in the des and
9629 3des routines to generate a key of the correct parity. Update S/MIME
9630 code to use new functions and hence generate correct parity DES keys.
9631 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9632 valid (weak or incorrect parity).
9636 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9637 as looking them up. This is useful when the verified structure may contain
9638 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9639 present unless the new PKCS7_NO_CRL flag is asserted.
9643 * Extend ASN1 oid configuration module. It now additionally accepts the
9646 shortName = some long name, 1.2.3.4
9650 * Reimplemented the BN_CTX implementation. There is now no more static
9651 limitation on the number of variables it can handle nor the depth of the
9652 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9653 information can now expand as required, and rather than having a single
9654 static array of bignums, BN_CTX now uses a linked-list of such arrays
9655 allowing it to expand on demand whilst maintaining the usefulness of
9656 BN_CTX's "bundling".
9660 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9661 to allow all RSA operations to function using a single BN_CTX.
9665 * Preliminary support for certificate policy evaluation and checking. This
9666 is initially intended to pass the tests outlined in "Conformance Testing
9667 of Relying Party Client Certificate Path Processing Logic" v1.07.
9671 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9672 remained unused and not that useful. A variety of other little bignum
9673 tweaks and fixes have also been made continuing on from the audit (see
9678 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9679 associated ASN1, EVP and SSL functions and old ASN1 macros.
9683 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9684 and this should never fail. So the return value from the use of
9685 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9686 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9690 * BN_CTX_get() should return zero-valued bignums, providing the same
9691 initialised value as BN_new().
9693 *Geoff Thorpe, suggested by Ulf Möller*
9695 * Support for inhibitAnyPolicy certificate extension.
9699 * An audit of the BIGNUM code is underway, for which debugging code is
9700 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9701 is considered valid when processing BIGNUMs, and causes execution to
9702 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9703 further steps are taken to deliberately pollute unused data in BIGNUM
9704 structures to try and expose faulty code further on. For now, openssl will
9705 (in its default mode of operation) continue to tolerate the inconsistent
9706 forms that it has tolerated in the past, but authors and packagers should
9707 consider trying openssl and their own applications when compiled with
9708 these debugging symbols defined. It will help highlight potential bugs in
9709 their own code, and will improve the test coverage for OpenSSL itself. At
9710 some point, these tighter rules will become openssl's default to improve
9711 maintainability, though the assert()s and other overheads will remain only
9712 in debugging configurations. See bn.h for more details.
9714 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9716 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9717 that can only be obtained through BN_CTX_new() (which implicitly
9718 initialises it). The presence of this function only made it possible
9719 to overwrite an existing structure (and cause memory leaks).
9723 * Because of the callback-based approach for implementing LHASH as a
9724 template type, lh_insert() adds opaque objects to hash-tables and
9725 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9726 to clean up those corresponding objects before destroying the hash table
9727 (and losing the object pointers). So some over-zealous constifications in
9728 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9729 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9730 prototyped to have "const" restrictions on the object pointers they are
9731 given (and so aren't required to cast them away any more).
9735 * The tmdiff.h API was so ugly and minimal that our own timing utility
9736 (speed) prefers to use its own implementation. The two implementations
9737 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9738 its object type properly exposed (MS_TM) instead of casting to/from
9739 `char *`. This may still change yet if someone realises MS_TM and
9741 aren't necessarily the greatest nomenclatures - but this is what was used
9742 internally to the implementation so I've used that for now.
9746 * Ensure that deprecated functions do not get compiled when
9747 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9748 the self-tests were still using deprecated key-generation functions so
9749 these have been updated also.
9753 * Reorganise PKCS#7 code to separate the digest location functionality
9754 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9755 New function PKCS7_set_digest() to set the digest type for PKCS#7
9756 digestedData type. Add additional code to correctly generate the
9757 digestedData type and add support for this type in PKCS7 initialization
9762 * New function PKCS7_set0_type_other() this initializes a PKCS7
9763 structure of type "other".
9767 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9768 sure the loop does correctly stop and breaking ("division by zero")
9769 modulus operations are not performed. The (pre-generated) prime
9770 table crypto/bn/bn_prime.h was already correct, but it could not be
9771 re-generated on some platforms because of the "division by zero"
9772 situation in the script.
9774 *Ralf S. Engelschall*
9776 * Update support for ECC-based TLS ciphersuites according to
9777 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9778 SHA-1 now is only used for "small" curves (where the
9779 representation of a field element takes up to 24 bytes); for
9780 larger curves, the field element resulting from ECDH is directly
9781 used as premaster secret.
9783 *Douglas Stebila (Sun Microsystems Laboratories)*
9785 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9786 curve secp160r1 to the tests.
9788 *Douglas Stebila (Sun Microsystems Laboratories)*
9790 * Add the possibility to load symbols globally with DSO.
9792 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9794 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9795 control of the error stack.
9799 * Add support for STORE in ENGINE.
9803 * Add the STORE type. The intention is to provide a common interface
9804 to certificate and key stores, be they simple file-based stores, or
9805 HSM-type store, or LDAP stores, or...
9806 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9810 * Add a generic structure called OPENSSL_ITEM. This can be used to
9811 pass a list of arguments to any function as well as provide a way
9812 for a function to pass data back to the caller.
9816 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9817 works like BUF_strdup() but can be used to duplicate a portion of
9818 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9823 * Add the function sk_find_ex() which works like sk_find(), but will
9824 return an index to an element even if an exact match couldn't be
9825 found. The index is guaranteed to point at the element where the
9826 searched-for key would be inserted to preserve sorting order.
9830 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9831 takes an extra flags argument for optional functionality. Currently,
9832 the following flags are defined:
9834 OBJ_BSEARCH_VALUE_ON_NOMATCH
9835 This one gets OBJ_bsearch_ex() to return a pointer to the first
9836 element where the comparing function returns a negative or zero
9839 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9840 This one gets OBJ_bsearch_ex() to return a pointer to the first
9841 element where the comparing function returns zero. This is useful
9842 if there are more than one element where the comparing function
9847 * Make it possible to create self-signed certificates with 'openssl ca'
9848 in such a way that the self-signed certificate becomes part of the
9849 CA database and uses the same mechanisms for serial number generation
9850 as all other certificate signing. The new flag '-selfsign' enables
9851 this functionality. Adapt CA.sh and CA.pl.in.
9855 * Add functionality to check the public key of a certificate request
9856 against a given private. This is useful to check that a certificate
9857 request can be signed by that key (self-signing).
9861 * Make it possible to have multiple active certificates with the same
9862 subject in the CA index file. This is done only if the keyword
9863 'unique_subject' is set to 'no' in the main CA section (default
9864 if 'CA_default') of the configuration file. The value is saved
9865 with the database itself in a separate index attribute file,
9866 named like the index file with '.attr' appended to the name.
9870 * Generate multi-valued AVAs using '+' notation in config files for
9875 * Support for nameConstraints certificate extension.
9879 * Support for policyConstraints certificate extension.
9883 * Support for policyMappings certificate extension.
9887 * Make sure the default DSA_METHOD implementation only uses its
9888 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9889 and change its own handlers to be NULL so as to remove unnecessary
9890 indirection. This lets alternative implementations fallback to the
9891 default implementation more easily.
9895 * Support for directoryName in GeneralName related extensions
9900 * Make it possible to link applications using Makefile.shared.
9901 Make that possible even when linking against static libraries!
9905 * Support for single pass processing for S/MIME signing. This now
9906 means that S/MIME signing can be done from a pipe, in addition
9907 cleartext signing (multipart/signed type) is effectively streaming
9908 and the signed data does not need to be all held in memory.
9910 This is done with a new flag PKCS7_STREAM. When this flag is set
9911 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9912 is done after the data is output (and digests calculated) in
9913 SMIME_write_PKCS7().
9917 * Add full support for -rpath/-R, both in shared libraries and
9918 applications, at least on the platforms where it's known how
9923 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9924 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9925 will now compute a table of multiples of the generator that
9926 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9927 faster (notably in the case of a single point multiplication,
9928 scalar * generator).
9930 *Nils Larsch, Bodo Moeller*
9932 * IPv6 support for certificate extensions. The various extensions
9933 which use the IP:a.b.c.d can now take IPv6 addresses using the
9934 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9939 * Added an ENGINE that implements RSA by performing private key
9940 exponentiations with the GMP library. The conversions to and from
9941 GMP's mpz_t format aren't optimised nor are any montgomery forms
9942 cached, and on x86 it appears OpenSSL's own performance has caught up.
9943 However there are likely to be other architectures where GMP could
9944 provide a boost. This ENGINE is not built in by default, but it can be
9945 specified at Configure time and should be accompanied by the necessary
9946 linker additions, eg;
9947 ./config -DOPENSSL_USE_GMP -lgmp
9951 * "openssl engine" will not display ENGINE/DSO load failure errors when
9952 testing availability of engines with "-t" - the old behaviour is
9953 produced by increasing the feature's verbosity with "-tt".
9957 * ECDSA routines: under certain error conditions uninitialized BN objects
9958 could be freed. Solution: make sure initialization is performed early
9959 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9964 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9965 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9966 software implementations. For DSA and DH, parameter generation can
9967 also be overridden by providing the appropriate method callbacks.
9971 * Change the "progress" mechanism used in key-generation and
9972 primality testing to functions that take a new BN_GENCB pointer in
9973 place of callback/argument pairs. The new API functions have `_ex`
9974 postfixes and the older functions are reimplemented as wrappers for
9975 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9976 declarations of the old functions to help (graceful) attempts to
9977 migrate to the new functions. Also, the new key-generation API
9978 functions operate on a caller-supplied key-structure and return
9979 success/failure rather than returning a key or NULL - this is to
9980 help make "keygen" another member function of RSA_METHOD etc.
9982 Example for using the new callback interface:
9984 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9988 BN_GENCB_set(&my_cb, my_callback, my_arg);
9990 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9991 /* For the meaning of a, b in calls to my_callback(), see the
9992 * documentation of the function that calls the callback.
9993 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9994 * my_callback should return 1 if it wants BN_is_prime_ex()
9995 * to continue, or 0 to stop.
10000 * Change the ZLIB compression method to be stateful, and make it
10001 available to TLS with the number defined in
10002 draft-ietf-tls-compression-04.txt.
10006 * Add the ASN.1 structures and functions for CertificatePair, which
10007 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10009 CertificatePair ::= SEQUENCE {
10010 forward [0] Certificate OPTIONAL,
10011 reverse [1] Certificate OPTIONAL,
10012 -- at least one of the pair shall be present -- }
10014 Also implement the PEM functions to read and write certificate
10015 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10017 This needed to be defined, mostly for the sake of the LDAP
10018 attribute crossCertificatePair, but may prove useful elsewhere as
10023 * Make it possible to inhibit symlinking of shared libraries in
10024 Makefile.shared, for Cygwin's sake.
10028 * Extend the BIGNUM API by creating a function
10029 void BN_set_negative(BIGNUM *a, int neg);
10030 and a macro that behave like
10031 int BN_is_negative(const BIGNUM *a);
10033 to avoid the need to access 'a->neg' directly in applications.
10037 * Implement fast modular reduction for pseudo-Mersenne primes
10038 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10039 EC_GROUP_new_curve_GFp() will now automatically use this
10042 *Nils Larsch <nla@trustcenter.de>*
10044 * Add new lock type (CRYPTO_LOCK_BN).
10048 * Change the ENGINE framework to automatically load engines
10049 dynamically from specific directories unless they could be
10050 found to already be built in or loaded. Move all the
10051 current engines except for the cryptodev one to a new
10052 directory engines/.
10053 The engines in engines/ are built as shared libraries if
10054 the "shared" options was given to ./Configure or ./config.
10055 Otherwise, they are inserted in libcrypto.a.
10056 /usr/local/ssl/engines is the default directory for dynamic
10057 engines, but that can be overridden at configure time through
10058 the usual use of --prefix and/or --openssldir, and at run
10059 time with the environment variable OPENSSL_ENGINES.
10061 *Geoff Thorpe and Richard Levitte*
10063 * Add Makefile.shared, a helper makefile to build shared
10064 libraries. Adapt Makefile.org.
10068 * Add version info to Win32 DLLs.
10070 *Peter 'Luna' Runestig" <peter@runestig.com>*
10072 * Add new 'medium level' PKCS#12 API. Certificates and keys
10073 can be added using this API to created arbitrary PKCS#12
10074 files while avoiding the low level API.
10076 New options to PKCS12_create(), key or cert can be NULL and
10077 will then be omitted from the output file. The encryption
10078 algorithm NIDs can be set to -1 for no encryption, the mac
10079 iteration count can be set to 0 to omit the mac.
10081 Enhance pkcs12 utility by making the -nokeys and -nocerts
10082 options work when creating a PKCS#12 file. New option -nomac
10083 to omit the mac, NONE can be set for an encryption algorithm.
10084 New code is modified to use the enhanced PKCS12_create()
10085 instead of the low level API.
10089 * Extend ASN1 encoder to support indefinite length constructed
10090 encoding. This can output sequences tags and octet strings in
10091 this form. Modify pk7_asn1.c to support indefinite length
10092 encoding. This is experimental and needs additional code to
10093 be useful, such as an ASN1 bio and some enhanced streaming
10096 Extend template encode functionality so that tagging is passed
10097 down to the template encoder.
10101 * Let 'openssl req' fail if an argument to '-newkey' is not
10102 recognized instead of using RSA as a default.
10106 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10107 As these are not official, they are not included in "ALL";
10108 the "ECCdraft" ciphersuite group alias can be used to select them.
10110 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10112 * Add ECDH engine support.
10114 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10116 * Add ECDH in new directory crypto/ecdh/.
10118 *Douglas Stebila (Sun Microsystems Laboratories)*
10120 * Let BN_rand_range() abort with an error after 100 iterations
10121 without success (which indicates a broken PRNG).
10125 * Change BN_mod_sqrt() so that it verifies that the input value
10126 is really the square of the return value. (Previously,
10127 BN_mod_sqrt would show GIGO behaviour.)
10131 * Add named elliptic curves over binary fields from X9.62, SECG,
10132 and WAP/WTLS; add OIDs that were still missing.
10134 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10136 * Extend the EC library for elliptic curves over binary fields
10137 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10140 EC_GF2m_simple_method
10144 EC_GROUP_new_curve_GF2m
10145 EC_GROUP_set_curve_GF2m
10146 EC_GROUP_get_curve_GF2m
10147 EC_POINT_set_affine_coordinates_GF2m
10148 EC_POINT_get_affine_coordinates_GF2m
10149 EC_POINT_set_compressed_coordinates_GF2m
10151 Point compression for binary fields is disabled by default for
10152 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10155 As binary polynomials are represented as BIGNUMs, various members
10156 of the EC_GROUP and EC_POINT data structures can be shared
10157 between the implementations for prime fields and binary fields;
10158 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10159 are essentially identical to their `..._GFp` counterparts.
10160 (For simplicity, the `..._GFp` prefix has been dropped from
10161 various internal method names.)
10163 An internal 'field_div' method (similar to 'field_mul' and
10164 'field_sqr') has been added; this is used only for binary fields.
10166 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10168 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10169 through methods ('mul', 'precompute_mult').
10171 The generic implementations (now internally called 'ec_wNAF_mul'
10172 and 'ec_wNAF_precomputed_mult') remain the default if these
10173 methods are undefined.
10175 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10177 * New function EC_GROUP_get_degree, which is defined through
10178 EC_METHOD. For curves over prime fields, this returns the bit
10179 length of the modulus.
10181 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10183 * New functions EC_GROUP_dup, EC_POINT_dup.
10184 (These simply call ..._new and ..._copy).
10186 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10188 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10189 Polynomials are represented as BIGNUMs (where the sign bit is not
10190 used) in the following functions [macros]:
10193 BN_GF2m_sub [= BN_GF2m_add]
10194 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10195 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10196 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10198 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10199 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10200 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10201 BN_GF2m_cmp [= BN_ucmp]
10203 (Note that only the 'mod' functions are actually for fields GF(2^m).
10204 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10206 For some functions, an the irreducible polynomial defining a
10207 field can be given as an 'unsigned int[]' with strictly
10208 decreasing elements giving the indices of those bits that are set;
10209 i.e., p[] represents the polynomial
10210 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10212 p[0] > p[1] > ... > p[k] = 0.
10213 This applies to the following functions:
10216 BN_GF2m_mod_mul_arr
10217 BN_GF2m_mod_sqr_arr
10218 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10219 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10220 BN_GF2m_mod_exp_arr
10221 BN_GF2m_mod_sqrt_arr
10222 BN_GF2m_mod_solve_quad_arr
10226 Conversion can be performed by the following functions:
10231 bntest.c has additional tests for binary polynomial arithmetic.
10233 Two implementations for BN_GF2m_mod_div() are available.
10234 The default algorithm simply uses BN_GF2m_mod_inv() and
10235 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10236 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10237 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10239 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10241 * Add new error code 'ERR_R_DISABLED' that can be used when some
10242 functionality is disabled at compile-time.
10244 *Douglas Stebila <douglas.stebila@sun.com>*
10246 * Change default behaviour of 'openssl asn1parse' so that more
10247 information is visible when viewing, e.g., a certificate:
10249 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10250 mode the content of non-printable OCTET STRINGs is output in a
10251 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10252 avoid the appearance of a printable string.
10254 *Nils Larsch <nla@trustcenter.de>*
10256 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10258 EC_GROUP_set_asn1_flag()
10259 EC_GROUP_get_asn1_flag()
10260 EC_GROUP_set_point_conversion_form()
10261 EC_GROUP_get_point_conversion_form()
10262 These control ASN1 encoding details:
10263 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10264 has been set to OPENSSL_EC_NAMED_CURVE.
10265 - Points are encoded in uncompressed form by default; options for
10266 asn1_for are as for point2oct, namely
10267 POINT_CONVERSION_COMPRESSED
10268 POINT_CONVERSION_UNCOMPRESSED
10269 POINT_CONVERSION_HYBRID
10271 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10273 EC_GROUP_set_seed()
10274 EC_GROUP_get0_seed()
10275 EC_GROUP_get_seed_len()
10276 This is used only for ASN1 purposes (so far).
10278 *Nils Larsch <nla@trustcenter.de>*
10280 * Add 'field_type' member to EC_METHOD, which holds the NID
10281 of the appropriate field type OID. The new function
10282 EC_METHOD_get_field_type() returns this value.
10284 *Nils Larsch <nla@trustcenter.de>*
10287 EC_POINT_point2bn()
10288 EC_POINT_bn2point()
10289 EC_POINT_point2hex()
10290 EC_POINT_hex2point()
10291 providing useful interfaces to EC_POINT_point2oct() and
10292 EC_POINT_oct2point().
10294 *Nils Larsch <nla@trustcenter.de>*
10296 * Change internals of the EC library so that the functions
10297 EC_GROUP_set_generator()
10298 EC_GROUP_get_generator()
10299 EC_GROUP_get_order()
10300 EC_GROUP_get_cofactor()
10301 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10302 to methods, which would lead to unnecessary code duplication when
10303 adding different types of curves.
10305 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10307 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10308 arithmetic, and such that modified wNAFs are generated
10309 (which avoid length expansion in many cases).
10313 * Add a function EC_GROUP_check_discriminant() (defined via
10314 EC_METHOD) that verifies that the curve discriminant is non-zero.
10316 Add a function EC_GROUP_check() that makes some sanity tests
10317 on a EC_GROUP, its generator and order. This includes
10318 EC_GROUP_check_discriminant().
10320 *Nils Larsch <nla@trustcenter.de>*
10322 * Add ECDSA in new directory crypto/ecdsa/.
10324 Add applications 'openssl ecparam' and 'openssl ecdsa'
10325 (these are based on 'openssl dsaparam' and 'openssl dsa').
10327 ECDSA support is also included in various other files across the
10328 library. Most notably,
10329 - 'openssl req' now has a '-newkey ecdsa:file' option;
10330 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10331 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10332 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10333 them suitable for ECDSA where domain parameters must be
10334 extracted before the specific public key;
10335 - ECDSA engine support has been added.
10337 *Nils Larsch <nla@trustcenter.de>*
10339 * Include some named elliptic curves, and add OIDs from X9.62,
10340 SECG, and WAP/WTLS. Each curve can be obtained from the new
10342 EC_GROUP_new_by_curve_name(),
10343 and the list of available named curves can be obtained with
10344 EC_get_builtin_curves().
10345 Also add a 'curve_name' member to EC_GROUP objects, which can be
10347 EC_GROUP_set_curve_name()
10348 EC_GROUP_get_curve_name()
10350 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10352 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10353 was actually never needed) and in BN_mul(). The removal in BN_mul()
10354 required a small change in bn_mul_part_recursive() and the addition
10355 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10356 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10357 bn_sub_words() and bn_add_words() except they take arrays with
10362 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10364 * Cleanse PEM buffers before freeing them since they may contain
10367 *Benjamin Bennett <ben@psc.edu>*
10369 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10370 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10371 authentication-only ciphersuites.
10375 * Since AES128 and AES256 share a single mask bit in the logic of
10376 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10377 kludge to work properly if AES128 is available and AES256 isn't.
10381 * Expand security boundary to match 1.1.1 module.
10385 * Remove redundant features: hash file source, editing of test vectors
10386 modify fipsld to use external fips_premain.c signature.
10390 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10391 run algorithm test programs.
10395 * Make algorithm test programs more tolerant of whitespace.
10399 * Have SSL/TLS server implementation tolerate "mismatched" record
10400 protocol version while receiving ClientHello even if the
10401 ClientHello is fragmented. (The server can't insist on the
10402 particular protocol version it has chosen before the ServerHello
10403 message has informed the client about his choice.)
10407 * Load error codes if they are not already present instead of using a
10408 static variable. This allows them to be cleanly unloaded and reloaded.
10412 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10414 * Introduce limits to prevent malicious keys being able to
10415 cause a denial of service. [CVE-2006-2940][]
10417 *Steve Henson, Bodo Moeller*
10419 * Fix ASN.1 parsing of certain invalid structures that can result
10420 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10422 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10423 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10425 * Fix SSL client code which could crash if connecting to a
10426 malicious SSLv2 server. [CVE-2006-4343][]
10428 *Tavis Ormandy and Will Drewry, Google Security Team*
10430 * Change ciphersuite string processing so that an explicit
10431 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10432 will no longer include "AES128-SHA"), and any other similar
10433 ciphersuite (same bitmap) from *other* protocol versions (so that
10434 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10435 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10436 changes from 0.9.8b and 0.9.8d.
10440 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10442 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10443 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10445 * Change the Unix randomness entropy gathering to use poll() when
10446 possible instead of select(), since the latter has some
10447 undesirable limitations.
10449 *Darryl Miles via Richard Levitte and Bodo Moeller*
10451 * Disable rogue ciphersuites:
10453 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10454 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10455 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10457 The latter two were purportedly from
10458 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10461 Also deactivate the remaining ciphersuites from
10462 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10463 unofficial, and the ID has long expired.
10467 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10468 dual-core machines) and other potential thread-safety issues.
10472 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10474 * Adapt fipsld and the build system to link against the validated FIPS
10475 module in FIPS mode.
10479 * Fixes for VC++ 2005 build under Windows.
10483 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10484 from a Windows bash shell such as MSYS. It is autodetected from the
10485 "config" script when run from a VC++ environment. Modify standard VC++
10486 build to use fipscanister.o from the GNU make build.
10490 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10492 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10493 The value now differs depending on if you build for FIPS or not.
10494 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10495 safely run with a non-FIPSed libcrypto, as it may crash because of
10496 the difference induced by this change.
10500 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10502 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10503 (part of SSL_OP_ALL). This option used to disable the
10504 countermeasure against man-in-the-middle protocol-version
10505 rollback in the SSL 2.0 server implementation, which is a bad
10506 idea. [CVE-2005-2969][]
10508 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10509 for Information Security, National Institute of Advanced Industrial
10510 Science and Technology [AIST, Japan)]*
10512 * Minimal support for X9.31 signatures and PSS padding modes. This is
10513 mainly for FIPS compliance and not fully integrated at this stage.
10517 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10518 the exponentiation using a fixed-length exponent. (Otherwise,
10519 the information leaked through timing could expose the secret key
10520 after many signatures; cf. Bleichenbacher's attack on DSA with
10525 * Make a new fixed-window mod_exp implementation the default for
10526 RSA, DSA, and DH private-key operations so that the sequence of
10527 squares and multiplies and the memory access pattern are
10528 independent of the particular secret key. This will mitigate
10529 cache-timing and potential related attacks.
10531 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10532 and this is automatically used by BN_mod_exp_mont() if the new flag
10533 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10534 will use this BN flag for private exponents unless the flag
10535 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10536 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10538 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10540 * Change the client implementation for SSLv23_method() and
10541 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10542 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10543 (Previously, the SSL 2.0 backwards compatible Client Hello
10544 message format would be used even with SSL_OP_NO_SSLv2.)
10548 * Add support for smime-type MIME parameter in S/MIME messages which some
10553 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10554 a threadsafe manner. Modify rsa code to use new function and add calls
10555 to dsa and dh code (which had race conditions before).
10559 * Include the fixed error library code in the C error file definitions
10560 instead of fixing them up at runtime. This keeps the error code
10561 structures constant.
10565 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10567 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10570 * Fixes for newer kerberos headers. NB: the casts are needed because
10571 the 'length' field is signed on one version and unsigned on another
10572 with no (?) obvious way to tell the difference, without these VC++
10573 complains. Also the "definition" of FAR (blank) is no longer included
10574 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10575 some needed definitions.
10579 * Undo Cygwin change.
10583 * Added support for proxy certificates according to RFC 3820.
10584 Because they may be a security thread to unaware applications,
10585 they must be explicitly allowed in run-time. See
10586 docs/HOWTO/proxy_certificates.txt for further information.
10590 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10592 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10593 server and client random values. Previously
10594 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10595 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10597 This change has negligible security impact because:
10599 1. Server and client random values still have 24 bytes of pseudo random
10602 2. Server and client random values are sent in the clear in the initial
10605 3. The master secret is derived using the premaster secret (48 bytes in
10606 size for static RSA ciphersuites) as well as client server and random
10609 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10612 *Stephen Henson, reported by UK NISCC*
10614 * Use Windows randomness collection on Cygwin.
10618 * Fix hang in EGD/PRNGD query when communication socket is closed
10619 prematurely by EGD/PRNGD.
10621 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10623 * Prompt for pass phrases when appropriate for PKCS12 input format.
10627 * Back-port of selected performance improvements from development
10628 branch, as well as improved support for PowerPC platforms.
10632 * Add lots of checks for memory allocation failure, error codes to indicate
10633 failure and freeing up memory if a failure occurs.
10635 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10637 * Add new -passin argument to dgst.
10641 * Perform some character comparisons of different types in X509_NAME_cmp:
10642 this is needed for some certificates that re-encode DNs into UTF8Strings
10643 (in violation of RFC3280) and can't or won't issue name rollover
10648 * Make an explicit check during certificate validation to see that
10649 the CA setting in each certificate on the chain is correct. As a
10650 side effect always do the following basic checks on extensions,
10651 not just when there's an associated purpose to the check:
10653 - if there is an unhandled critical extension (unless the user
10654 has chosen to ignore this fault)
10655 - if the path length has been exceeded (if one is set at all)
10656 - that certain extensions fit the associated purpose (if one has
10661 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10663 * Avoid a race condition when CRLs are checked in a multi threaded
10664 environment. This would happen due to the reordering of the revoked
10665 entries during signature checking and serial number lookup. Now the
10666 encoding is cached and the serial number sort performed under a lock.
10667 Add new STACK function sk_is_sorted().
10671 * Add Delta CRL to the extension code.
10675 * Various fixes to s3_pkt.c so alerts are sent properly.
10677 *David Holmes <d.holmes@f5.com>*
10679 * Reduce the chances of duplicate issuer name and serial numbers (in
10680 violation of RFC3280) using the OpenSSL certificate creation utilities.
10681 This is done by creating a random 64 bit value for the initial serial
10682 number when a serial number file is created or when a self signed
10683 certificate is created using 'openssl req -x509'. The initial serial
10684 number file is created using 'openssl x509 -next_serial' in CA.pl
10685 rather than being initialized to 1.
10689 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10691 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10692 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10694 *Joe Orton, Steve Henson*
10696 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10699 *Joe Orton, Steve Henson*
10701 * Make it possible to have multiple active certificates with the same
10702 subject in the CA index file. This is done only if the keyword
10703 'unique_subject' is set to 'no' in the main CA section (default
10704 if 'CA_default') of the configuration file. The value is saved
10705 with the database itself in a separate index attribute file,
10706 named like the index file with '.attr' appended to the name.
10710 * X509 verify fixes. Disable broken certificate workarounds when
10711 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10712 keyUsage extension present. Don't accept CRLs with unhandled critical
10713 extensions: since verify currently doesn't process CRL extensions this
10714 rejects a CRL with *any* critical extensions. Add new verify error codes
10719 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10720 A clarification of RFC2560 will require the use of OCTET STRINGs and
10721 some implementations cannot handle the current raw format. Since OpenSSL
10722 copies and compares OCSP nonces as opaque blobs without any attempt at
10723 parsing them this should not create any compatibility issues.
10727 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10728 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10729 this HMAC (and other) operations are several times slower than OpenSSL
10734 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10736 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10738 * Use the correct content when signing type "other".
10742 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10744 * Fix various bugs revealed by running the NISCC test suite:
10746 Stop out of bounds reads in the ASN1 code when presented with
10747 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10749 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10751 If verify callback ignores invalid public key errors don't try to check
10752 certificate signature with the NULL public key.
10756 * New -ignore_err option in ocsp application to stop the server
10757 exiting on the first error in a request.
10761 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10762 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10767 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10768 extra data after the compression methods not only for TLS 1.0
10769 but also for SSL 3.0 (as required by the specification).
10771 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10773 * Change X509_certificate_type() to mark the key as exported/exportable
10774 when it's 512 *bits* long, not 512 bytes.
10778 * Change AES_cbc_encrypt() so it outputs exact multiple of
10779 blocks during encryption.
10783 * Various fixes to base64 BIO and non blocking I/O. On write
10784 flushes were not handled properly if the BIO retried. On read
10785 data was not being buffered properly and had various logic bugs.
10786 This also affects blocking I/O when the data being decoded is a
10791 * Various S/MIME bugfixes and compatibility changes:
10792 output correct application/pkcs7 MIME type if
10793 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10794 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10795 of files as .eml work). Correctly handle very long lines in MIME
10800 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10802 * Countermeasure against the Klima-Pokorny-Rosa extension of
10803 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10804 a protocol version number mismatch like a decryption error
10805 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10809 * Turn on RSA blinding by default in the default implementation
10810 to avoid a timing attack. Applications that don't want it can call
10811 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10812 They would be ill-advised to do so in most cases.
10814 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10816 * Change RSA blinding code so that it works when the PRNG is not
10817 seeded (in this case, the secret RSA exponent is abused as
10818 an unpredictable seed -- if it is not unpredictable, there
10819 is no point in blinding anyway). Make RSA blinding thread-safe
10820 by remembering the creator's thread ID in rsa->blinding and
10821 having all other threads use local one-time blinding factors
10822 (this requires more computation than sharing rsa->blinding, but
10823 avoids excessive locking; and if an RSA object is not shared
10824 between threads, blinding will still be very fast).
10828 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10829 ENGINE as defaults for all supported algorithms irrespective of
10830 the 'flags' parameter. 'flags' is now honoured, so applications
10831 should make sure they are passing it correctly.
10835 * Target "mingw" now allows native Windows code to be generated in
10836 the Cygwin environment as well as with the MinGW compiler.
10840 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10842 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10843 via timing by performing a MAC computation even if incorrect
10844 block cipher padding has been found. This is a countermeasure
10845 against active attacks where the attacker has to distinguish
10846 between bad padding and a MAC verification error. [CVE-2003-0078][]
10848 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10849 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10850 Martin Vuagnoux (EPFL, Ilion)*
10852 * Make the no-err option work as intended. The intention with no-err
10853 is not to have the whole error stack handling routines removed from
10854 libcrypto, it's only intended to remove all the function name and
10855 reason texts, thereby removing some of the footprint that may not
10856 be interesting if those errors aren't displayed anyway.
10858 NOTE: it's still possible for any application or module to have its
10859 own set of error texts inserted. The routines are there, just not
10860 used by default when no-err is given.
10864 * Add support for FreeBSD on IA64.
10866 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10868 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10869 Kerberos function mit_des_cbc_cksum(). Before this change,
10870 the value returned by DES_cbc_cksum() was like the one from
10871 mit_des_cbc_cksum(), except the bytes were swapped.
10873 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10875 * Allow an application to disable the automatic SSL chain building.
10876 Before this a rather primitive chain build was always performed in
10877 ssl3_output_cert_chain(): an application had no way to send the
10878 correct chain if the automatic operation produced an incorrect result.
10880 Now the chain builder is disabled if either:
10882 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10884 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10886 The reasoning behind this is that an application would not want the
10887 auto chain building to take place if extra chain certificates are
10888 present and it might also want a means of sending no additional
10889 certificates (for example the chain has two certificates and the
10894 * Add the possibility to build without the ENGINE framework.
10896 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10898 * Under Win32 gmtime() can return NULL: check return value in
10899 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10903 * DSA routines: under certain error conditions uninitialized BN objects
10904 could be freed. Solution: make sure initialization is performed early
10905 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10906 Nils Larsch <nla@trustcenter.de> via PR#459)
10910 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10911 checked on reconnect on the client side, therefore session resumption
10912 could still fail with a "ssl session id is different" error. This
10913 behaviour is masked when SSL_OP_ALL is used due to
10914 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10915 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10916 followup to PR #377.
10920 * IA-32 assembler support enhancements: unified ELF targets, support
10921 for SCO/Caldera platforms, fix for Cygwin shared build.
10925 * Add support for FreeBSD on sparc64. As a consequence, support for
10926 FreeBSD on non-x86 processors is separate from x86 processors on
10927 the config script, much like the NetBSD support.
10929 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10931 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
10933 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10936 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10937 code (06) was taken as the first octet of the session ID and the last
10938 octet was ignored consequently. As a result SSLv2 client side session
10939 caching could not have worked due to the session ID mismatch between
10941 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10946 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10947 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10952 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10953 seems that in spite of existing for more than a year, many application
10954 author have done nothing to provide the necessary callbacks, which
10955 means that this particular engine will not work properly anywhere.
10956 This is a very unfortunate situation which forces us, in the name
10957 of usability, to give the hw_ncipher.c a static lock, which is part
10959 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10960 appear in 0.9.8 or later. We EXPECT application authors to have
10961 dealt properly with this when 0.9.8 is released (unless we actually
10962 make such changes in the libcrypto locking code that changes will
10963 have to be made anyway).
10967 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10968 octets have been read, EOF or an error occurs. Without this change
10969 some truncated ASN1 structures will not produce an error.
10973 * Disable Heimdal support, since it hasn't been fully implemented.
10974 Still give the possibility to force the use of Heimdal, but with
10975 warnings and a request that patches get sent to openssl-dev.
10979 * Add the VC-CE target, introduce the WINCE sysname, and add
10980 INSTALL.WCE and appropriate conditionals to make it build.
10982 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10984 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10985 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10986 edit numbers of the version.
10988 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10990 * Introduce safe string copy and catenation functions
10991 (BUF_strlcpy() and BUF_strlcat()).
10993 *Ben Laurie (CHATS) and Richard Levitte*
10995 * Avoid using fixed-size buffers for one-line DNs.
10997 *Ben Laurie (CHATS)*
10999 * Add BUF_MEM_grow_clean() to avoid information leakage when
11000 resizing buffers containing secrets, and use where appropriate.
11002 *Ben Laurie (CHATS)*
11004 * Avoid using fixed size buffers for configuration file location.
11006 *Ben Laurie (CHATS)*
11008 * Avoid filename truncation for various CA files.
11010 *Ben Laurie (CHATS)*
11012 * Use sizeof in preference to magic numbers.
11014 *Ben Laurie (CHATS)*
11016 * Avoid filename truncation in cert requests.
11018 *Ben Laurie (CHATS)*
11020 * Add assertions to check for (supposedly impossible) buffer
11023 *Ben Laurie (CHATS)*
11025 * Don't cache truncated DNS entries in the local cache (this could
11026 potentially lead to a spoofing attack).
11028 *Ben Laurie (CHATS)*
11030 * Fix various buffers to be large enough for hex/decimal
11031 representations in a platform independent manner.
11033 *Ben Laurie (CHATS)*
11035 * Add CRYPTO_realloc_clean() to avoid information leakage when
11036 resizing buffers containing secrets, and use where appropriate.
11038 *Ben Laurie (CHATS)*
11040 * Add BIO_indent() to avoid much slightly worrying code to do
11043 *Ben Laurie (CHATS)*
11045 * Convert sprintf()/BIO_puts() to BIO_printf().
11047 *Ben Laurie (CHATS)*
11049 * buffer_gets() could terminate with the buffer only half
11052 *Ben Laurie (CHATS)*
11054 * Add assertions to prevent user-supplied crypto functions from
11055 overflowing internal buffers by having large block sizes, etc.
11057 *Ben Laurie (CHATS)*
11059 * New OPENSSL_assert() macro (similar to assert(), but enabled
11062 *Ben Laurie (CHATS)*
11064 * Eliminate unused copy of key in RC4.
11066 *Ben Laurie (CHATS)*
11068 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11070 *Ben Laurie (CHATS)*
11072 * Fix off-by-one error in EGD path.
11074 *Ben Laurie (CHATS)*
11076 * If RANDFILE path is too long, ignore instead of truncating.
11078 *Ben Laurie (CHATS)*
11080 * Eliminate unused and incorrectly sized X.509 structure
11083 *Ben Laurie (CHATS)*
11085 * Eliminate unused and dangerous function knumber().
11087 *Ben Laurie (CHATS)*
11089 * Eliminate unused and dangerous structure, KSSL_ERR.
11091 *Ben Laurie (CHATS)*
11093 * Protect against overlong session ID context length in an encoded
11094 session object. Since these are local, this does not appear to be
11097 *Ben Laurie (CHATS)*
11099 * Change from security patch (see 0.9.6e below) that did not affect
11100 the 0.9.6 release series:
11102 Remote buffer overflow in SSL3 protocol - an attacker could
11103 supply an oversized master key in Kerberos-enabled versions.
11106 *Ben Laurie (CHATS)*
11108 * Change the SSL kerb5 codes to match RFC 2712.
11112 * Make -nameopt work fully for req and add -reqopt switch.
11114 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11116 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11118 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11120 * Make sure tests can be performed even if the corresponding algorithms
11121 have been removed entirely. This was also the last step to make
11122 OpenSSL compilable with DJGPP under all reasonable conditions.
11124 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11126 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11127 to allow version independent disabling of normally unselected ciphers,
11128 which may be activated as a side-effect of selecting a single cipher.
11130 (E.g., cipher list string "RSA" enables ciphersuites that are left
11131 out of "ALL" because they do not provide symmetric encryption.
11132 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11134 *Lutz Jaenicke, Bodo Moeller*
11136 * Add appropriate support for separate platform-dependent build
11137 directories. The recommended way to make a platform-dependent
11138 build directory is the following (tested on Linux), maybe with
11141 # Place yourself outside of the OpenSSL source tree. In
11142 # this example, the environment variable OPENSSL_SOURCE
11143 # is assumed to contain the absolute OpenSSL source directory.
11144 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11145 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11146 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11147 mkdir -p `dirname $F`
11148 ln -s $OPENSSL_SOURCE/$F $F
11151 To be absolutely sure not to disturb the source tree, a "make clean"
11152 is a good thing. If it isn't successful, don't worry about it,
11153 it probably means the source directory is very clean.
11157 * Make sure any ENGINE control commands make local copies of string
11158 pointers passed to them whenever necessary. Otherwise it is possible
11159 the caller may have overwritten (or deallocated) the original string
11160 data when a later ENGINE operation tries to use the stored values.
11162 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11164 * Improve diagnostics in file reading and command-line digests.
11166 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11168 * Add AES modes CFB and OFB to the object database. Correct an
11169 error in AES-CFB decryption.
11173 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11174 allows existing EVP_CIPHER_CTX structures to be reused after
11175 calling `EVP_*Final()`. This behaviour is used by encryption
11176 BIOs and some applications. This has the side effect that
11177 applications must explicitly clean up cipher contexts with
11178 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11182 * Check the values of dna and dnb in bn_mul_recursive before calling
11183 bn_mul_comba (a non zero value means the a or b arrays do not contain
11184 n2 elements) and fallback to bn_mul_normal if either is not zero.
11188 * Fix escaping of non-ASCII characters when using the -subj option
11189 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11193 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11194 form for "surname", serialNumber has no short form.
11195 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11196 therefore remove "mail" short name for "internet 7".
11197 The OID for unique identifiers in X509 certificates is
11198 x500UniqueIdentifier, not uniqueIdentifier.
11199 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11203 * Add an "init" command to the ENGINE config module and auto initialize
11204 ENGINEs. Without any "init" command the ENGINE will be initialized
11205 after all ctrl commands have been executed on it. If init=1 the
11206 ENGINE is initialized at that point (ctrls before that point are run
11207 on the uninitialized ENGINE and after on the initialized one). If
11208 init=0 then the ENGINE will not be initialized at all.
11212 * Fix the 'app_verify_callback' interface so that the user-defined
11213 argument is actually passed to the callback: In the
11214 SSL_CTX_set_cert_verify_callback() prototype, the callback
11215 declaration has been changed from
11218 int (*cb)(X509_STORE_CTX *,void *);
11219 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11220 i=s->ctx->app_verify_callback(&ctx)
11221 has been changed into
11222 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11224 To update applications using SSL_CTX_set_cert_verify_callback(),
11225 a dummy argument can be added to their callback functions.
11227 *D. K. Smetters <smetters@parc.xerox.com>*
11229 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11231 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11233 * Add and OPENSSL_LOAD_CONF define which will cause
11234 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11235 This allows older applications to transparently support certain
11236 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11237 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11238 load the config file and OPENSSL_add_all_algorithms_conf() which will
11239 always load it have also been added.
11243 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11244 Adjust NIDs and EVP layer.
11246 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11248 * Config modules support in openssl utility.
11250 Most commands now load modules from the config file,
11251 though in a few (such as version) this isn't done
11252 because it couldn't be used for anything.
11254 In the case of ca and req the config file used is
11255 the same as the utility itself: that is the -config
11256 command line option can be used to specify an
11261 * Move default behaviour from OPENSSL_config(). If appname is NULL
11262 use "openssl_conf" if filename is NULL use default openssl config file.
11266 * Add an argument to OPENSSL_config() to allow the use of an alternative
11267 config section name. Add a new flag to tolerate a missing config file
11268 and move code to CONF_modules_load_file().
11272 * Support for crypto accelerator cards from Accelerated Encryption
11273 Processing, www.aep.ie. (Use engine 'aep')
11274 The support was copied from 0.9.6c [engine] and adapted/corrected
11275 to work with the new engine framework.
11277 *AEP Inc. and Richard Levitte*
11279 * Support for SureWare crypto accelerator cards from Baltimore
11280 Technologies. (Use engine 'sureware')
11281 The support was copied from 0.9.6c [engine] and adapted
11282 to work with the new engine framework.
11286 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11287 make the newer ENGINE framework commands for the CHIL engine work.
11289 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11291 * Make it possible to produce shared libraries on ReliantUNIX.
11293 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11295 * Add the configuration target debug-linux-ppro.
11296 Make 'openssl rsa' use the general key loading routines
11297 implemented in apps.c, and make those routines able to
11298 handle the key format FORMAT_NETSCAPE and the variant
11301 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11303 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11305 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11307 * Add -keyform to rsautl, and document -engine.
11309 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11311 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11312 BIO_R_NO_SUCH_FILE error code rather than the generic
11313 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11317 * Add new functions
11318 ERR_peek_last_error
11319 ERR_peek_last_error_line
11320 ERR_peek_last_error_line_data.
11321 These are similar to
11323 ERR_peek_error_line
11324 ERR_peek_error_line_data,
11325 but report on the latest error recorded rather than the first one
11326 still in the error queue.
11328 *Ben Laurie, Bodo Moeller*
11330 * default_algorithms option in ENGINE config module. This allows things
11332 default_algorithms = ALL
11333 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11337 * Preliminary ENGINE config module.
11341 * New experimental application configuration code.
11345 * Change the AES code to follow the same name structure as all other
11346 symmetric ciphers, and behave the same way. Move everything to
11347 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11349 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11351 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11353 *Ben Laurie and Theo de Raadt*
11355 * Add option to output public keys in req command.
11357 *Massimiliano Pala madwolf@openca.org*
11359 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11360 (up to about 10% better than before for P-192 and P-224).
11364 * New functions/macros
11366 SSL_CTX_set_msg_callback(ctx, cb)
11367 SSL_CTX_set_msg_callback_arg(ctx, arg)
11368 SSL_set_msg_callback(ssl, cb)
11369 SSL_set_msg_callback_arg(ssl, arg)
11371 to request calling a callback function
11373 void cb(int write_p, int version, int content_type,
11374 const void *buf, size_t len, SSL *ssl, void *arg)
11376 whenever a protocol message has been completely received
11377 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11378 protocol version according to which the SSL library interprets
11379 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11380 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11381 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11382 specification (change_cipher_spec(20), alert(21), handshake(22)).
11383 'buf' and 'len' point to the actual message, 'ssl' to the
11384 SSL object, and 'arg' is the application-defined value set by
11385 SSL[_CTX]_set_msg_callback_arg().
11387 'openssl s_client' and 'openssl s_server' have new '-msg' options
11388 to enable a callback that displays all protocol messages.
11392 * Change the shared library support so shared libraries are built as
11393 soon as the corresponding static library is finished, and thereby get
11394 openssl and the test programs linked against the shared library.
11395 This still only happens when the keyword "shard" has been given to
11396 the configuration scripts.
11398 NOTE: shared library support is still an experimental thing, and
11399 backward binary compatibility is still not guaranteed.
11401 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11403 * Add support for Subject Information Access extension.
11405 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11407 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11408 additional bytes when new memory had to be allocated, not just
11409 when reusing an existing buffer.
11413 * New command line and configuration option 'utf8' for the req command.
11414 This allows field values to be specified as UTF8 strings.
11418 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11419 runs for the former and machine-readable output for the latter.
11423 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11424 of the e-mail address in the DN (i.e., it will go into a certificate
11425 extension only). The new configuration file option 'email_in_dn = no'
11426 has the same effect.
11428 *Massimiliano Pala madwolf@openca.org*
11430 * Change all functions with names starting with `des_` to be starting
11431 with `DES_` instead. Add wrappers that are compatible with libdes,
11432 but are named `_ossl_old_des_*`. Finally, add macros that map the
11433 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11434 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11435 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11438 Since we provide two compatibility mappings, the user needs to
11439 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11440 compatibility is desired. The default (i.e., when that macro
11441 isn't defined) is OpenSSL 0.9.6c compatibility.
11443 There are also macros that enable and disable the support of old
11444 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11445 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11446 are defined, the default will apply: to support the old des routines.
11448 In either case, one must include openssl/des.h to get the correct
11449 definitions. Do not try to just include openssl/des_old.h, that
11452 NOTE: This is a major break of an old API into a new one. Software
11453 authors are encouraged to switch to the `DES_` style functions. Some
11454 time in the future, des_old.h and the libdes compatibility functions
11455 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11456 default), and then completely removed.
11460 * Test for certificates which contain unsupported critical extensions.
11461 If such a certificate is found during a verify operation it is
11462 rejected by default: this behaviour can be overridden by either
11463 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11464 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11465 X509_supported_extension() has also been added which returns 1 if a
11466 particular extension is supported.
11470 * Modify the behaviour of EVP cipher functions in similar way to digests
11471 to retain compatibility with existing code.
11475 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11476 compatibility with existing code. In particular the 'ctx' parameter does
11477 not have to be to be initialized before the call to EVP_DigestInit() and
11478 it is tidied up after a call to EVP_DigestFinal(). New function
11479 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11480 EVP_MD_CTX_copy() changed to not require the destination to be
11481 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11482 requires the destination to be valid.
11484 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11485 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11489 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11490 so that complete 'Handshake' protocol structures are kept in memory
11491 instead of overwriting 'msg_type' and 'length' with 'body' data.
11495 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11497 *Massimo Santin via Richard Levitte*
11499 * Major restructuring to the underlying ENGINE code. This includes
11500 reduction of linker bloat, separation of pure "ENGINE" manipulation
11501 (initialisation, etc) from functionality dealing with implementations
11502 of specific crypto interfaces. This change also introduces integrated
11503 support for symmetric ciphers and digest implementations - so ENGINEs
11504 can now accelerate these by providing EVP_CIPHER and EVP_MD
11505 implementations of their own. This is detailed in crypto/engine/README
11506 as it couldn't be adequately described here. However, there are a few
11507 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11508 were changed in the original introduction of ENGINE code have now
11509 reverted back - the hooking from this code to ENGINE is now a good
11510 deal more passive and at run-time, operations deal directly with
11511 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11512 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11513 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11514 they were not being used by the framework as there is no concept of a
11515 BIGNUM_METHOD and they could not be generalised to the new
11516 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11517 ENGINE_cpy() has been removed as it cannot be consistently defined in
11522 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11526 * Change mkdef.pl to sort symbols that get the same entry number,
11527 and make sure the automatically generated functions `ERR_load_*`
11528 become part of libeay.num as well.
11532 * New function SSL_renegotiate_pending(). This returns true once
11533 renegotiation has been requested (either SSL_renegotiate() call
11534 or HelloRequest/ClientHello received from the peer) and becomes
11535 false once a handshake has been completed.
11536 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11537 sends a HelloRequest, but does not ensure that a handshake takes
11538 place. SSL_renegotiate_pending() is useful for checking if the
11539 client has followed the request.)
11543 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11544 By default, clients may request session resumption even during
11545 renegotiation (if session ID contexts permit); with this option,
11546 session resumption is possible only in the first handshake.
11548 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11549 more bits available for options that should not be part of
11550 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11554 * Add some demos for certificate and certificate request creation.
11558 * Make maximum certificate chain size accepted from the peer application
11559 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11560 "Douglas E. Engert" <deengert@anl.gov>.
11564 * Add support for shared libraries for Unixware-7
11565 (Boyd Lynn Gerber <gerberb@zenez.com>).
11569 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11570 be done prior to destruction. Use this to unload error strings from
11571 ENGINEs that load their own error strings. NB: This adds two new API
11572 functions to "get" and "set" this destroy handler in an ENGINE.
11576 * Alter all existing ENGINE implementations (except "openssl" and
11577 "openbsd") to dynamically instantiate their own error strings. This
11578 makes them more flexible to be built both as statically-linked ENGINEs
11579 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11580 Also, add stub code to each that makes building them as self-contained
11581 shared-libraries easier (see README.ENGINE).
11585 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11586 implementations into applications that are completely implemented in
11587 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11588 commands that can be used to configure what shared-library to load and
11589 to control aspects of the way it is handled. Also, made an update to
11590 the README.ENGINE file that brings its information up-to-date and
11591 provides some information and instructions on the "dynamic" ENGINE
11592 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11596 * Make it possible to unload ranges of ERR strings with a new
11597 "ERR_unload_strings" function.
11601 * Add a copy() function to EVP_MD.
11605 * Make EVP_MD routines take a context pointer instead of just the
11606 md_data void pointer.
11610 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11611 that the digest can only process a single chunk of data
11612 (typically because it is provided by a piece of
11613 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11614 is only going to provide a single chunk of data, and hence the
11615 framework needn't accumulate the data for oneshot drivers.
11619 * As with "ERR", make it possible to replace the underlying "ex_data"
11620 functions. This change also alters the storage and management of global
11621 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11622 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11623 index counters. The API functions that use this state have been changed
11624 to take a "class_index" rather than pointers to the class's local STACK
11625 and counter, and there is now an API function to dynamically create new
11626 classes. This centralisation allows us to (a) plug a lot of the
11627 thread-safety problems that existed, and (b) makes it possible to clean
11628 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11629 such data would previously have always leaked in application code and
11630 workarounds were in place to make the memory debugging turn a blind eye
11631 to it. Application code that doesn't use this new function will still
11632 leak as before, but their memory debugging output will announce it now
11633 rather than letting it slide.
11635 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11636 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11637 has a return value to indicate success or failure.
11641 * Make it possible to replace the underlying "ERR" functions such that the
11642 global state (2 LHASH tables and 2 locks) is only used by the "default"
11643 implementation. This change also adds two functions to "get" and "set"
11644 the implementation prior to it being automatically set the first time
11645 any other ERR function takes place. Ie. an application can call "get",
11646 pass the return value to a module it has just loaded, and that module
11647 can call its own "set" function using that value. This means the
11648 module's "ERR" operations will use (and modify) the error state in the
11649 application and not in its own statically linked copy of OpenSSL code.
11653 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11654 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11655 the operation, and provides a more encapsulated way for external code
11656 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11657 to use these functions rather than manually incrementing the counts.
11659 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11663 * Add EVP test program.
11667 * Add symmetric cipher support to ENGINE. Expect the API to change!
11671 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11672 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11673 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11674 These allow a CRL to be built without having to access X509_CRL fields
11675 directly. Modify 'ca' application to use new functions.
11679 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11680 bug workarounds. Rollback attack detection is a security feature.
11681 The problem will only arise on OpenSSL servers when TLSv1 is not
11682 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11683 Software authors not wanting to support TLSv1 will have special reasons
11684 for their choice and can explicitly enable this option.
11686 *Bodo Moeller, Lutz Jaenicke*
11688 * Rationalise EVP so it can be extended: don't include a union of
11689 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11690 (similar to those existing for EVP_CIPHER_CTX).
11695 EVP_MD_CTX_init(&md); /* new function call */
11696 EVP_DigestInit(&md, EVP_sha1());
11697 EVP_DigestUpdate(&md, in, len);
11698 EVP_DigestFinal(&md, out, NULL);
11699 EVP_MD_CTX_cleanup(&md); /* new function call */
11703 * Make DES key schedule conform to the usual scheme, as well as
11704 correcting its structure. This means that calls to DES functions
11705 now have to pass a pointer to a des_key_schedule instead of a
11706 plain des_key_schedule (which was actually always a pointer
11709 des_key_schedule ks;
11711 des_set_key_checked(..., &ks);
11712 des_ncbc_encrypt(..., &ks, ...);
11714 (Note that a later change renames 'des_...' into 'DES_...'.)
11718 * Initial reduction of linker bloat: the use of some functions, such as
11719 PEM causes large amounts of unused functions to be linked in due to
11720 poor organisation. For example pem_all.c contains every PEM function
11721 which has a knock on effect of linking in large amounts of (unused)
11722 ASN1 code. Grouping together similar functions and splitting unrelated
11723 functions prevents this.
11727 * Cleanup of EVP macros.
11731 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11732 correct `_ecb suffix`.
11736 * Add initial OCSP responder support to ocsp application. The
11737 revocation information is handled using the text based index
11738 use by the ca application. The responder can either handle
11739 requests generated internally, supplied in files (for example
11740 via a CGI script) or using an internal minimal server.
11744 * Add configuration choices to get zlib compression for TLS.
11748 * Changes to Kerberos SSL for RFC 2712 compliance:
11749 1. Implemented real KerberosWrapper, instead of just using
11750 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11751 2. Implemented optional authenticator field of KerberosWrapper.
11753 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11754 and authenticator structs; see crypto/krb5/.
11756 Generalized Kerberos calls to support multiple Kerberos libraries.
11757 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11758 via Richard Levitte*
11760 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11761 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11762 values for each of the key sizes rather than having just
11763 parameters (and 'speed' generating keys each time).
11767 * Speed up EVP routines.
11770 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11771 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11772 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11773 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11775 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11776 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11777 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11780 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11782 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11786 * Added the OS2-EMX target.
11788 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11790 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11791 to support NCONF routines in extension code. New function CONF_set_nconf()
11792 to allow functions which take an NCONF to also handle the old LHASH
11793 structure: this means that the old CONF compatible routines can be
11794 retained (in particular wrt extensions) without having to duplicate the
11795 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11799 * Enhance the general user interface with mechanisms for inner control
11800 and with possibilities to have yes/no kind of prompts.
11804 * Change all calls to low level digest routines in the library and
11805 applications to use EVP. Add missing calls to HMAC_cleanup() and
11806 don't assume HMAC_CTX can be copied using memcpy().
11808 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11810 * Add the possibility to control engines through control names but with
11811 arbitrary arguments instead of just a string.
11812 Change the key loaders to take a UI_METHOD instead of a callback
11813 function pointer. NOTE: this breaks binary compatibility with earlier
11814 versions of OpenSSL [engine].
11815 Adapt the nCipher code for these new conditions and add a card insertion
11820 * Enhance the general user interface with mechanisms to better support
11821 dialog box interfaces, application-defined prompts, the possibility
11822 to use defaults (for example default passwords from somewhere else)
11823 and interrupts/cancellations.
11827 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11828 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11832 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11833 tidy up some unnecessarily weird code in 'sk_new()').
11835 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11837 * Change the key loading routines for ENGINEs to use the same kind
11838 callback (pem_password_cb) as all other routines that need this
11843 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11844 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11845 than this minimum value is recommended.
11849 * New random seeder for OpenVMS, using the system process statistics
11850 that are easily reachable.
11854 * Windows apparently can't transparently handle global
11855 variables defined in DLLs. Initialisations such as:
11857 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11859 won't compile. This is used by the any applications that need to
11860 declare their own ASN1 modules. This was fixed by adding the option
11861 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11862 needed for static libraries under Win32.
11866 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11867 setting of purpose and trust fields. New X509_STORE trust and
11868 purpose functions and tidy up setting in other SSL functions.
11872 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11873 structure. These are inherited by X509_STORE_CTX when it is
11874 initialised. This allows various defaults to be set in the
11875 X509_STORE structure (such as flags for CRL checking and custom
11876 purpose or trust settings) for functions which only use X509_STORE_CTX
11877 internally such as S/MIME.
11879 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11880 trust settings if they are not set in X509_STORE. This allows X509_STORE
11881 purposes and trust (in S/MIME for example) to override any set by default.
11883 Add command line options for CRL checking to smime, s_client and s_server
11888 * Initial CRL based revocation checking. If the CRL checking flag(s)
11889 are set then the CRL is looked up in the X509_STORE structure and
11890 its validity and signature checked, then if the certificate is found
11891 in the CRL the verify fails with a revoked error.
11893 Various new CRL related callbacks added to X509_STORE_CTX structure.
11895 Command line options added to 'verify' application to support this.
11897 This needs some additional work, such as being able to handle multiple
11898 CRLs with different times, extension based lookup (rather than just
11899 by subject name) and ultimately more complete V2 CRL extension
11904 * Add a general user interface API (crypto/ui/). This is designed
11905 to replace things like des_read_password and friends (backward
11906 compatibility functions using this new API are provided).
11907 The purpose is to remove prompting functions from the DES code
11908 section as well as provide for prompting through dialog boxes in
11909 a window system and the like.
11913 * Add "ex_data" support to ENGINE so implementations can add state at a
11914 per-structure level rather than having to store it globally.
11918 * Make it possible for ENGINE structures to be copied when retrieved by
11919 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11920 This causes the "original" ENGINE structure to act like a template,
11921 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11922 operational state can be localised to each ENGINE structure, despite the
11923 fact they all share the same "methods". New ENGINE structures returned in
11924 this case have no functional references and the return value is the single
11925 structural reference. This matches the single structural reference returned
11926 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11931 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11932 needs to match any other type at all we need to manually clear the
11937 * Changes to the "openssl engine" utility to include;
11938 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11939 about an ENGINE's available control commands.
11940 - executing control commands from command line arguments using the
11941 '-pre' and '-post' switches. '-post' is only used if '-t' is
11942 specified and the ENGINE is successfully initialised. The syntax for
11943 the individual commands are colon-separated, for example;
11944 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11948 * New dynamic control command support for ENGINEs. ENGINEs can now
11949 declare their own commands (numbers), names (strings), descriptions,
11950 and input types for run-time discovery by calling applications. A
11951 subset of these commands are implicitly classed as "executable"
11952 depending on their input type, and only these can be invoked through
11953 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11954 can be based on user input, config files, etc). The distinction is
11955 that "executable" commands cannot return anything other than a boolean
11956 result and can only support numeric or string input, whereas some
11957 discoverable commands may only be for direct use through
11958 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11959 pointers, or other custom uses. The "executable" commands are to
11960 support parameterisations of ENGINE behaviour that can be
11961 unambiguously defined by ENGINEs and used consistently across any
11962 OpenSSL-based application. Commands have been added to all the
11963 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11964 control over shared-library paths without source code alterations.
11968 * Changed all ENGINE implementations to dynamically allocate their
11969 ENGINEs rather than declaring them statically. Apart from this being
11970 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11971 this also allows the implementations to compile without using the
11972 internal engine_int.h header.
11976 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11977 'const' value. Any code that should be able to modify a RAND_METHOD
11978 should already have non-const pointers to it (ie. they should only
11979 modify their own ones).
11983 * Made a variety of little tweaks to the ENGINE code.
11984 - "atalla" and "ubsec" string definitions were moved from header files
11985 to C code. "nuron" string definitions were placed in variables
11986 rather than hard-coded - allowing parameterisation of these values
11987 later on via ctrl() commands.
11988 - Removed unused "#if 0"'d code.
11989 - Fixed engine list iteration code so it uses ENGINE_free() to release
11990 structural references.
11991 - Constified the RAND_METHOD element of ENGINE structures.
11992 - Constified various get/set functions as appropriate and added
11993 missing functions (including a catch-all ENGINE_cpy that duplicates
11994 all ENGINE values onto a new ENGINE except reference counts/state).
11995 - Removed NULL parameter checks in get/set functions. Setting a method
11996 or function to NULL is a way of cancelling out a previously set
11997 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11998 and doesn't justify the extra error symbols and code.
11999 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12000 flags from engine_int.h to engine.h.
12001 - Changed prototypes for ENGINE handler functions (init(), finish(),
12002 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12006 * Implement binary inversion algorithm for BN_mod_inverse in addition
12007 to the algorithm using long division. The binary algorithm can be
12008 used only if the modulus is odd. On 32-bit systems, it is faster
12009 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12010 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12011 up to 450 bits. In 64-bit environments, the binary algorithm
12012 appears to be advantageous for much longer moduli; here we use it
12013 for moduli up to 2048 bits.
12017 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12018 could not support the combine flag in choice fields.
12022 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12023 extensions from a certificate request to the certificate.
12027 * Allow multiple 'certopt' and 'nameopt' options to be separated
12028 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12029 file: this allows the display of the certificate about to be
12030 signed to be customised, to allow certain fields to be included
12031 or excluded and extension details. The old system didn't display
12032 multicharacter strings properly, omitted fields not in the policy
12033 and couldn't display additional details such as extensions.
12037 * Function EC_POINTs_mul for multiple scalar multiplication
12038 of an arbitrary number of elliptic curve points
12039 \sum scalars[i]*points[i],
12040 optionally including the generator defined for the EC_GROUP:
12041 scalar*generator + \sum scalars[i]*points[i].
12043 EC_POINT_mul is a simple wrapper function for the typical case
12044 that the point list has just one item (besides the optional
12049 * First EC_METHODs for curves over GF(p):
12051 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12052 operations and provides various method functions that can also
12053 operate with faster implementations of modular arithmetic.
12055 EC_GFp_mont_method() reuses most functions that are part of
12056 EC_GFp_simple_method, but uses Montgomery arithmetic.
12058 *Bodo Moeller; point addition and point doubling
12059 implementation directly derived from source code provided by
12060 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12062 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12063 crypto/ec/ec_lib.c):
12065 Curves are EC_GROUP objects (with an optional group generator)
12066 based on EC_METHODs that are built into the library.
12068 Points are EC_POINT objects based on EC_GROUP objects.
12070 Most of the framework would be able to handle curves over arbitrary
12071 finite fields, but as there are no obvious types for fields other
12072 than GF(p), some functions are limited to that for now.
12076 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12077 that the file contains a complete HTTP response.
12081 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12082 change the def and num file printf format specifier from "%-40sXXX"
12083 to "%-39s XXX". The latter will always guarantee a space after the
12084 field while the former will cause them to run together if the field
12085 is 40 of more characters long.
12089 * Constify the cipher and digest 'method' functions and structures
12090 and modify related functions to take constant EVP_MD and EVP_CIPHER
12095 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12096 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12100 * Modify `EVP_Digest*()` routines so they now return values. Although the
12101 internal software routines can never fail additional hardware versions
12106 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12108 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12109 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12112 ERR_R_NESTED_ASN1_ERROR
12114 ERR_R_MISSING_ASN1_EOS
12115 were 4 .. 9, conflicting with
12116 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12118 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12119 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12121 Add new error code 'ERR_R_INTERNAL_ERROR'.
12125 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12130 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12131 sets the subject name for a new request or supersedes the
12132 subject name in a given request. Formats that can be parsed are
12133 'CN=Some Name, OU=myOU, C=IT'
12135 'CN=Some Name/OU=myOU/C=IT'.
12137 Add options '-batch' and '-verbose' to 'openssl req'.
12139 *Massimiliano Pala <madwolf@hackmasters.net>*
12141 * Introduce the possibility to access global variables through
12142 functions on platform were that's the best way to handle exporting
12143 global variables in shared libraries. To enable this functionality,
12144 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12145 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12146 is normally done by Configure or something similar).
12148 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12149 in the source file (foo.c) like this:
12151 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12152 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12154 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12155 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12157 OPENSSL_DECLARE_GLOBAL(int,foo);
12158 #define foo OPENSSL_GLOBAL_REF(foo)
12159 OPENSSL_DECLARE_GLOBAL(double,bar);
12160 #define bar OPENSSL_GLOBAL_REF(bar)
12162 The #defines are very important, and therefore so is including the
12163 header file everywhere where the defined globals are used.
12165 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12166 of ASN.1 items, but that structure is a bit different.
12168 The largest change is in util/mkdef.pl which has been enhanced with
12169 better and easier to understand logic to choose which symbols should
12170 go into the Windows .def files as well as a number of fixes and code
12171 cleanup (among others, algorithm keywords are now sorted
12172 lexicographically to avoid constant rewrites).
12176 * In BN_div() keep a copy of the sign of 'num' before writing the
12177 result to 'rm' because if rm==num the value will be overwritten
12178 and produce the wrong result if 'num' is negative: this caused
12179 problems with BN_mod() and BN_nnmod().
12183 * Function OCSP_request_verify(). This checks the signature on an
12184 OCSP request and verifies the signer certificate. The signer
12185 certificate is just checked for a generic purpose and OCSP request
12190 * Add OCSP_check_validity() function to check the validity of OCSP
12191 responses. OCSP responses are prepared in real time and may only
12192 be a few seconds old. Simply checking that the current time lies
12193 between thisUpdate and nextUpdate max reject otherwise valid responses
12194 caused by either OCSP responder or client clock inaccuracy. Instead
12195 we allow thisUpdate and nextUpdate to fall within a certain period of
12196 the current time. The age of the response can also optionally be
12197 checked. Two new options -validity_period and -status_age added to
12202 * If signature or public key algorithm is unrecognized print out its
12203 OID rather that just UNKNOWN.
12207 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12208 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12209 ID to be generated from the issuer certificate alone which can then be
12210 passed to OCSP_id_issuer_cmp().
12214 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12215 ASN1 modules to export functions returning ASN1_ITEM pointers
12216 instead of the ASN1_ITEM structures themselves. This adds several
12217 new macros which allow the underlying ASN1 function/structure to
12218 be accessed transparently. As a result code should not use ASN1_ITEM
12219 references directly (such as &X509_it) but instead use the relevant
12220 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12221 use of the new ASN1 code on platforms where exporting structures
12222 is problematical (for example in shared libraries) but exporting
12223 functions returning pointers to structures is not.
12227 * Add support for overriding the generation of SSL/TLS session IDs.
12228 These callbacks can be registered either in an SSL_CTX or per SSL.
12229 The purpose of this is to allow applications to control, if they wish,
12230 the arbitrary values chosen for use as session IDs, particularly as it
12231 can be useful for session caching in multiple-server environments. A
12232 command-line switch for testing this (and any client code that wishes
12233 to use such a feature) has been added to "s_server".
12235 *Geoff Thorpe, Lutz Jaenicke*
12237 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12238 of the form '#if defined(...) || defined(...) || ...' and
12239 '#if !defined(...) && !defined(...) && ...'. This also avoids
12240 the growing number of special cases it was previously handling.
12244 * Make all configuration macros available for application by making
12245 sure they are available in opensslconf.h, by giving them names starting
12246 with `OPENSSL_` to avoid conflicts with other packages and by making
12247 sure e_os2.h will cover all platform-specific cases together with
12249 Additionally, it is now possible to define configuration/platform-
12250 specific names (called "system identities"). In the C code, these
12251 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12252 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12253 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12258 * New option -set_serial to 'req' and 'x509' this allows the serial
12259 number to use to be specified on the command line. Previously self
12260 signed certificates were hard coded with serial number 0 and the
12261 CA options of 'x509' had to use a serial number in a file which was
12266 * New options to 'ca' utility to support V2 CRL entry extensions.
12267 Currently CRL reason, invalidity date and hold instruction are
12268 supported. Add new CRL extensions to V3 code and some new objects.
12272 * New function EVP_CIPHER_CTX_set_padding() this is used to
12273 disable standard block padding (aka PKCS#5 padding) in the EVP
12274 API, which was previously mandatory. This means that the data is
12275 not padded in any way and so the total length much be a multiple
12276 of the block size, otherwise an error occurs.
12280 * Initial (incomplete) OCSP SSL support.
12284 * New function OCSP_parse_url(). This splits up a URL into its host,
12285 port and path components: primarily to parse OCSP URLs. New -url
12286 option to ocsp utility.
12290 * New nonce behavior. The return value of OCSP_check_nonce() now
12291 reflects the various checks performed. Applications can decide
12292 whether to tolerate certain situations such as an absent nonce
12293 in a response when one was present in a request: the ocsp application
12294 just prints out a warning. New function OCSP_add1_basic_nonce()
12295 this is to allow responders to include a nonce in a response even if
12296 the request is nonce-less.
12300 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12301 skipped when using openssl x509 multiple times on a single input file,
12302 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12306 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12307 set string type: to handle setting ASN1_TIME structures. Fix ca
12308 utility to correctly initialize revocation date of CRLs.
12312 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12313 the clients preferred ciphersuites and rather use its own preferences.
12314 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12315 Internet Explorer by ensuring unchanged hash method during stepup.
12316 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12320 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12321 to aes and add a new 'exist' option to print out symbols that don't
12326 * Additional options to ocsp utility to allow flags to be set and
12327 additional certificates supplied.
12331 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12332 OCSP client a number of certificate to only verify the response
12337 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12338 handle the new API. Currently only ECB, CBC modes supported. Add new
12341 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12342 Encryption Standard (AES) Ciphersuites for Transport Layer
12343 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12344 not enabled by default and were not part of the "ALL" ciphersuite
12345 alias because they were not yet official; they could be
12346 explicitly requested by specifying the "AESdraft" ciphersuite
12347 group alias. In the final release of OpenSSL 0.9.7, the group
12348 alias is called "AES" and is part of "ALL".)
12350 *Ben Laurie, Steve Henson, Bodo Moeller*
12352 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12353 request to response.
12357 * Functions for OCSP responders. OCSP_request_onereq_count(),
12358 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12359 extract information from a certificate request. OCSP_response_create()
12360 creates a response and optionally adds a basic response structure.
12361 OCSP_basic_add1_status() adds a complete single response to a basic
12362 response and returns the OCSP_SINGLERESP structure just added (to allow
12363 extensions to be included for example). OCSP_basic_add1_cert() adds a
12364 certificate to a basic response and OCSP_basic_sign() signs a basic
12365 response with various flags. New helper functions ASN1_TIME_check()
12366 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12367 (converts ASN1_TIME to GeneralizedTime).
12371 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12372 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12373 structure from a certificate. X509_pubkey_digest() digests the public_key
12374 contents: this is used in various key identifiers.
12378 * Make sk_sort() tolerate a NULL argument.
12380 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12382 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12383 passed by the function are trusted implicitly. If any of them signed the
12384 response then it is assumed to be valid and is not verified.
12388 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12389 to data. This was previously part of the PKCS7 ASN1 code. This
12390 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12391 *Steve Henson, reported by Kenneth R. Robinette
12392 <support@securenetterm.com>*
12394 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12395 routines: without these tracing memory leaks is very painful.
12396 Fix leaks in PKCS12 and PKCS7 routines.
12400 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12401 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12402 effectively meant GeneralizedTime would never be used. Now it
12403 is initialised to -1 but X509_time_adj() now has to check the value
12404 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12405 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12406 *Steve Henson, reported by Kenneth R. Robinette
12407 <support@securenetterm.com>*
12409 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12410 result in a zero length in the ASN1_INTEGER structure which was
12411 not consistent with the structure when d2i_ASN1_INTEGER() was used
12412 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12413 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12414 where it did not print out a minus for negative ASN1_INTEGER.
12418 * Add summary printout to ocsp utility. The various functions which
12419 convert status values to strings have been renamed to:
12420 OCSP_response_status_str(), OCSP_cert_status_str() and
12421 OCSP_crl_reason_str() and are no longer static. New options
12422 to verify nonce values and to disable verification. OCSP response
12423 printout format cleaned up.
12427 * Add additional OCSP certificate checks. These are those specified
12428 in RFC2560. This consists of two separate checks: the CA of the
12429 certificate being checked must either be the OCSP signer certificate
12430 or the issuer of the OCSP signer certificate. In the latter case the
12431 OCSP signer certificate must contain the OCSP signing extended key
12432 usage. This check is performed by attempting to match the OCSP
12433 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12434 in the OCSP_CERTID structures of the response.
12438 * Initial OCSP certificate verification added to OCSP_basic_verify()
12439 and related routines. This uses the standard OpenSSL certificate
12440 verify routines to perform initial checks (just CA validity) and
12441 to obtain the certificate chain. Then additional checks will be
12442 performed on the chain. Currently the root CA is checked to see
12443 if it is explicitly trusted for OCSP signing. This is used to set
12444 a root CA as a global signing root: that is any certificate that
12445 chains to that CA is an acceptable OCSP signing certificate.
12449 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12450 extensions from a separate configuration file.
12451 As when reading extensions from the main configuration file,
12452 the '-extensions ...' option may be used for specifying the
12455 *Massimiliano Pala <madwolf@comune.modena.it>*
12457 * New OCSP utility. Allows OCSP requests to be generated or
12458 read. The request can be sent to a responder and the output
12459 parsed, outputted or printed in text form. Not complete yet:
12460 still needs to check the OCSP response validity.
12464 * New subcommands for 'openssl ca':
12465 `openssl ca -status <serial>` prints the status of the cert with
12466 the given serial number (according to the index file).
12467 `openssl ca -updatedb` updates the expiry status of certificates
12470 *Massimiliano Pala <madwolf@comune.modena.it>*
12472 * New '-newreq-nodes' command option to CA.pl. This is like
12473 '-newreq', but calls 'openssl req' with the '-nodes' option
12474 so that the resulting key is not encrypted.
12476 *Damien Miller <djm@mindrot.org>*
12478 * New configuration for the GNU Hurd.
12480 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12482 * Initial code to implement OCSP basic response verify. This
12483 is currently incomplete. Currently just finds the signer's
12484 certificate and verifies the signature on the response.
12488 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12489 value of OPENSSLDIR. This is available via the new '-d' option
12490 to 'openssl version', and is also included in 'openssl version -a'.
12494 * Allowing defining memory allocation callbacks that will be given
12495 file name and line number information in additional arguments
12496 (a `const char*` and an int). The basic functionality remains, as
12497 well as the original possibility to just replace malloc(),
12498 realloc() and free() by functions that do not know about these
12499 additional arguments. To register and find out the current
12500 settings for extended allocation functions, the following
12501 functions are provided:
12503 CRYPTO_set_mem_ex_functions
12504 CRYPTO_set_locked_mem_ex_functions
12505 CRYPTO_get_mem_ex_functions
12506 CRYPTO_get_locked_mem_ex_functions
12508 These work the same way as CRYPTO_set_mem_functions and friends.
12509 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12510 extended allocation function is enabled.
12511 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12512 a conventional allocation function is enabled.
12514 *Richard Levitte, Bodo Moeller*
12516 * Finish off removing the remaining LHASH function pointer casts.
12517 There should no longer be any prototype-casting required when using
12518 the LHASH abstraction, and any casts that remain are "bugs". See
12519 the callback types and macros at the head of lhash.h for details
12520 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12524 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12525 If /dev/[u]random devices are not available or do not return enough
12526 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12528 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12529 /etc/entropy will be queried once each in this sequence, querying stops
12530 when enough entropy was collected without querying more sockets.
12534 * Change the Unix RAND_poll() variant to be able to poll several
12535 random devices, as specified by DEVRANDOM, until a sufficient amount
12536 of data has been collected. We spend at most 10 ms on each file
12537 (select timeout) and read in non-blocking mode. DEVRANDOM now
12538 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12539 (previously it was just the string "/dev/urandom"), so on typical
12540 platforms the 10 ms delay will never occur.
12541 Also separate out the Unix variant to its own file, rand_unix.c.
12542 For VMS, there's a currently-empty rand_vms.c.
12546 * Move OCSP client related routines to ocsp_cl.c. These
12547 provide utility functions which an application needing
12548 to issue a request to an OCSP responder and analyse the
12549 response will typically need: as opposed to those which an
12550 OCSP responder itself would need which will be added later.
12552 OCSP_request_sign() signs an OCSP request with an API similar
12553 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12554 response. OCSP_response_get1_basic() extracts basic response
12555 from response. OCSP_resp_find_status(): finds and extracts status
12556 information from an OCSP_CERTID structure (which will be created
12557 when the request structure is built). These are built from lower
12558 level functions which work on OCSP_SINGLERESP structures but
12559 won't normally be used unless the application wishes to examine
12560 extensions in the OCSP response for example.
12562 Replace nonce routines with a pair of functions.
12563 OCSP_request_add1_nonce() adds a nonce value and optionally
12564 generates a random value. OCSP_check_nonce() checks the
12565 validity of the nonce in an OCSP response.
12569 * Change function OCSP_request_add() to OCSP_request_add0_id().
12570 This doesn't copy the supplied OCSP_CERTID and avoids the
12571 need to free up the newly created id. Change return type
12572 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12573 This can then be used to add extensions to the request.
12574 Deleted OCSP_request_new(), since most of its functionality
12575 is now in OCSP_REQUEST_new() (and the case insensitive name
12576 clash) apart from the ability to set the request name which
12577 will be added elsewhere.
12581 * Update OCSP API. Remove obsolete extensions argument from
12582 various functions. Extensions are now handled using the new
12583 OCSP extension code. New simple OCSP HTTP function which
12584 can be used to send requests and parse the response.
12588 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12589 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12590 uses the special reorder version of SET OF to sort the attributes
12591 and reorder them to match the encoded order. This resolves a long
12592 standing problem: a verify on a PKCS7 structure just after signing
12593 it used to fail because the attribute order did not match the
12594 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12595 it uses the received order. This is necessary to tolerate some broken
12596 software that does not order SET OF. This is handled by encoding
12597 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12598 to produce the required SET OF.
12602 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12603 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12604 files to get correct declarations of the ASN.1 item variables.
12608 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12609 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12610 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12611 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12612 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12613 ASN1_ITEM and no wrapper functions.
12617 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12618 replace the old function pointer based I/O routines. Change most of
12619 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12623 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12624 lines, recognize more "algorithms" that can be deselected, and make
12625 it complain about algorithm deselection that isn't recognised.
12629 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12630 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12631 to use new functions. Add NO_ASN1_OLD which can be set to remove
12632 some old style ASN1 functions: this can be used to determine if old
12633 code will still work when these eventually go away.
12637 * New extension functions for OCSP structures, these follow the
12638 same conventions as certificates and CRLs.
12642 * New function X509V3_add1_i2d(). This automatically encodes and
12643 adds an extension. Its behaviour can be customised with various
12644 flags to append, replace or delete. Various wrappers added for
12645 certificates and CRLs.
12649 * Fix to avoid calling the underlying ASN1 print routine when
12650 an extension cannot be parsed. Correct a typo in the
12651 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12655 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12656 entries for variables.
12660 * Add functionality to apps/openssl.c for detecting locking
12661 problems: As the program is single-threaded, all we have
12662 to do is register a locking callback using an array for
12663 storing which locks are currently held by the program.
12667 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12668 SSL_get_ex_data_X509_STORE_idx(), which is used in
12669 ssl_verify_cert_chain() and thus can be called at any time
12670 during TLS/SSL handshakes so that thread-safety is essential.
12671 Unfortunately, the ex_data design is not at all suited
12672 for multi-threaded use, so it probably should be abolished.
12676 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12678 *Broadcom, tweaked and integrated by Geoff Thorpe*
12680 * Move common extension printing code to new function
12681 X509V3_print_extensions(). Reorganise OCSP print routines and
12682 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12686 * New function X509_signature_print() to remove duplication in some
12691 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12692 set (this was treated exactly the same as SET OF previously). This
12693 is used to reorder the STACK representing the structure to match the
12694 encoding. This will be used to get round a problem where a PKCS7
12695 structure which was signed could not be verified because the STACK
12696 order did not reflect the encoded order.
12700 * Reimplement the OCSP ASN1 module using the new code.
12704 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12705 for its ASN1 operations. The old style function pointers still exist
12706 for now but they will eventually go away.
12710 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12711 completely replaces the old ASN1 functionality with a table driven
12712 encoder and decoder which interprets an ASN1_ITEM structure describing
12713 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12714 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12715 has also been converted to the new form.
12719 * Change BN_mod_exp_recp so that negative moduli are tolerated
12720 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12721 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12722 for negative moduli.
12726 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12727 of not touching the result's sign bit.
12731 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12736 * Changed the LHASH code to use prototypes for callbacks, and created
12737 macros to declare and implement thin (optionally static) functions
12738 that provide type-safety and avoid function pointer casting for the
12739 type-specific callbacks.
12743 * Added Kerberos Cipher Suites to be used with TLS, as written in
12745 *Veers Staats <staatsvr@asc.hpc.mil>,
12746 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12748 * Reformat the FAQ so the different questions and answers can be divided
12749 in sections depending on the subject.
12753 * Have the zlib compression code load ZLIB.DLL dynamically under
12758 * New function BN_mod_sqrt for computing square roots modulo a prime
12759 (using the probabilistic Tonelli-Shanks algorithm unless
12760 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12761 be handled deterministically).
12763 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12765 * Make BN_mod_inverse faster by explicitly handling small quotients
12766 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12767 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12771 * New function BN_kronecker.
12775 * Fix BN_gcd so that it works on negative inputs; the result is
12776 positive unless both parameters are zero.
12777 Previously something reasonably close to an infinite loop was
12778 possible because numbers could be growing instead of shrinking
12779 in the implementation of Euclid's algorithm.
12783 * Fix BN_is_word() and BN_is_one() macros to take into account the
12784 sign of the number in question.
12786 Fix BN_is_word(a,w) to work correctly for w == 0.
12788 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12789 because its test if the absolute value of 'a' equals 'w'.
12790 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12791 it exists mostly for use in the implementations of BN_is_zero(),
12792 BN_is_one(), and BN_is_word().
12796 * New function BN_swap.
12800 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12801 the exponentiation functions are more likely to produce reasonable
12802 results on negative inputs.
12806 * Change BN_mod_mul so that the result is always non-negative.
12807 Previously, it could be negative if one of the factors was negative;
12808 I don't think anyone really wanted that behaviour.
12812 * Move `BN_mod_...` functions into new file crypto/bn/bn_mod.c
12813 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12814 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12815 and add new functions:
12824 BN_mod_lshift1_quick
12826 BN_mod_lshift_quick
12828 These functions always generate non-negative results.
12830 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12831 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12833 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12834 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12835 be reduced modulo m.
12837 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12840 The following entry accidentally appeared in the CHANGES file
12841 distributed with OpenSSL 0.9.7. The modifications described in
12842 it do *not* apply to OpenSSL 0.9.7.
12844 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12845 was actually never needed) and in BN_mul(). The removal in BN_mul()
12846 required a small change in bn_mul_part_recursive() and the addition
12847 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12848 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12849 bn_sub_words() and bn_add_words() except they take arrays with
12855 * In 'openssl passwd', verify passwords read from the terminal
12856 unless the '-salt' option is used (which usually means that
12857 verification would just waste user's time since the resulting
12858 hash is going to be compared with some given password hash)
12859 or the new '-noverify' option is used.
12861 This is an incompatible change, but it does not affect
12862 non-interactive use of 'openssl passwd' (passwords on the command
12863 line, '-stdin' option, '-in ...' option) and thus should not
12864 cause any problems.
12868 * Remove all references to RSAref, since there's no more need for it.
12872 * Make DSO load along a path given through an environment variable
12873 (SHLIB_PATH) with shl_load().
12877 * Constify the ENGINE code as a result of BIGNUM constification.
12878 Also constify the RSA code and most things related to it. In a
12879 few places, most notable in the depth of the ASN.1 code, ugly
12880 casts back to non-const were required (to be solved at a later
12885 * Make it so the openssl application has all engines loaded by default.
12889 * Constify the BIGNUM routines a little more.
12893 * Add the following functions:
12895 ENGINE_load_cswift()
12897 ENGINE_load_atalla()
12898 ENGINE_load_nuron()
12899 ENGINE_load_builtin_engines()
12901 That way, an application can itself choose if external engines that
12902 are built-in in OpenSSL shall ever be used or not. The benefit is
12903 that applications won't have to be linked with libdl or other dso
12904 libraries unless it's really needed.
12906 Changed 'openssl engine' to load all engines on demand.
12907 Changed the engine header files to avoid the duplication of some
12908 declarations (they differed!).
12912 * 'openssl engine' can now list capabilities.
12916 * Better error reporting in 'openssl engine'.
12920 * Never call load_dh_param(NULL) in s_server.
12924 * Add engine application. It can currently list engines by name and
12925 identity, and test if they are actually available.
12929 * Improve RPM specification file by forcing symbolic linking and making
12930 sure the installed documentation is also owned by root.root.
12932 *Damien Miller <djm@mindrot.org>*
12934 * Give the OpenSSL applications more possibilities to make use of
12935 keys (public as well as private) handled by engines.
12939 * Add OCSP code that comes from CertCo.
12943 * Add VMS support for the Rijndael code.
12947 * Added untested support for Nuron crypto accelerator.
12951 * Add support for external cryptographic devices. This code was
12952 previously distributed separately as the "engine" branch.
12954 *Geoff Thorpe, Richard Levitte*
12956 * Rework the filename-translation in the DSO code. It is now possible to
12957 have far greater control over how a "name" is turned into a filename
12958 depending on the operating environment and any oddities about the
12959 different shared library filenames on each system.
12963 * Support threads on FreeBSD-elf in Configure.
12967 * Fix for SHA1 assembly problem with MASM: it produces
12968 warnings about corrupt line number information when assembling
12969 with debugging information. This is caused by the overlapping
12972 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12975 NCONF_get_number() has no error checking at all. As a replacement,
12976 NCONF_get_number_e() is defined (`_e` for "error checking") and is
12977 promoted strongly. The old NCONF_get_number is kept around for
12978 binary backward compatibility.
12979 Make it possible for methods to load from something other than a BIO,
12980 by providing a function pointer that is given a name instead of a BIO.
12981 For example, this could be used to load configuration data from an
12986 * Fix for non blocking accept BIOs. Added new I/O special reason
12987 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12988 with non blocking I/O was not possible because no retry code was
12989 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12994 * Added the beginnings of Rijndael support.
12998 * Fix for bug in DirectoryString mask setting. Add support for
12999 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13000 to allow certificate printing to more controllable, additional
13001 'certopt' option to 'x509' to allow new printing options to be
13006 * Clean old EAY MD5 hack from e_os.h.
13010 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13012 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13013 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13015 *Joe Orton, Steve Henson*
13017 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13019 * Fix additional bug revealed by the NISCC test suite:
13021 Stop bug triggering large recursion when presented with
13022 certain ASN.1 tags [CVE-2003-0851][]
13026 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13028 * Fix various bugs revealed by running the NISCC test suite:
13030 Stop out of bounds reads in the ASN1 code when presented with
13031 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13033 If verify callback ignores invalid public key errors don't try to check
13034 certificate signature with the NULL public key.
13038 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13039 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13044 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13045 extra data after the compression methods not only for TLS 1.0
13046 but also for SSL 3.0 (as required by the specification).
13048 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13050 * Change X509_certificate_type() to mark the key as exported/exportable
13051 when it's 512 *bits* long, not 512 bytes.
13055 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13057 * Countermeasure against the Klima-Pokorny-Rosa extension of
13058 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13059 a protocol version number mismatch like a decryption error
13060 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13064 * Turn on RSA blinding by default in the default implementation
13065 to avoid a timing attack. Applications that don't want it can call
13066 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13067 They would be ill-advised to do so in most cases.
13069 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13071 * Change RSA blinding code so that it works when the PRNG is not
13072 seeded (in this case, the secret RSA exponent is abused as
13073 an unpredictable seed -- if it is not unpredictable, there
13074 is no point in blinding anyway). Make RSA blinding thread-safe
13075 by remembering the creator's thread ID in rsa->blinding and
13076 having all other threads use local one-time blinding factors
13077 (this requires more computation than sharing rsa->blinding, but
13078 avoids excessive locking; and if an RSA object is not shared
13079 between threads, blinding will still be very fast).
13083 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13085 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13086 via timing by performing a MAC computation even if incorrect
13087 block cipher padding has been found. This is a countermeasure
13088 against active attacks where the attacker has to distinguish
13089 between bad padding and a MAC verification error. [CVE-2003-0078][]
13091 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13092 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13093 Martin Vuagnoux (EPFL, Ilion)*
13095 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13097 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13098 memory from its contents. This is done with a counter that will
13099 place alternating values in each byte. This can be used to solve
13100 two issues: 1) the removal of calls to memset() by highly optimizing
13101 compilers, and 2) cleansing with other values than 0, since those can
13102 be read through on certain media, for example a swap space on disk.
13106 * Bugfix: client side session caching did not work with external caching,
13107 because the session->cipher setting was not restored when reloading
13108 from the external cache. This problem was masked, when
13109 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13110 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13114 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13115 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13117 *Zeev Lieber <zeev-l@yahoo.com>*
13119 * Undo an undocumented change introduced in 0.9.6e which caused
13120 repeated calls to OpenSSL_add_all_ciphers() and
13121 OpenSSL_add_all_digests() to be ignored, even after calling
13126 * Change the default configuration reader to deal with last line not
13127 being properly terminated.
13131 * Change X509_NAME_cmp() so it applies the special rules on handling
13132 DN values that are of type PrintableString, as well as RDNs of type
13133 emailAddress where the value has the type ia5String.
13135 *stefank@valicert.com via Richard Levitte*
13137 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13138 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13139 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13140 the bitwise-OR of the two for use by the majority of applications
13141 wanting this behaviour, and update the docs. The documented
13142 behaviour and actual behaviour were inconsistent and had been
13143 changing anyway, so this is more a bug-fix than a behavioural
13146 *Geoff Thorpe, diagnosed by Nadav Har'El*
13148 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13149 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13153 * Fix initialization code race conditions in
13154 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13155 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13156 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13157 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13158 ssl2_get_cipher_by_char(),
13159 ssl3_get_cipher_by_char().
13161 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13163 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13164 the cached sessions are flushed, as the remove_cb() might use ex_data
13165 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13166 (see [openssl.org #212]).
13168 *Geoff Thorpe, Lutz Jaenicke*
13170 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13171 length, instead of the encoding length to d2i_ASN1_OBJECT.
13175 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13177 * [In 0.9.6g-engine release:]
13178 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13180 *Lynn Gazis <lgazis@rainbow.com>*
13182 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13184 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13185 and get fix the header length calculation.
13186 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13187 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13189 * Use proper error handling instead of 'assertions' in buffer
13190 overflow checks added in 0.9.6e. This prevents DoS (the
13191 assertions could call abort()).
13193 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13195 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13197 * Add various sanity checks to asn1_get_length() to reject
13198 the ASN1 length bytes if they exceed sizeof(long), will appear
13199 negative or the content length exceeds the length of the
13202 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13204 * Fix cipher selection routines: ciphers without encryption had no flags
13205 for the cipher strength set and where therefore not handled correctly
13206 by the selection routines (PR #130).
13210 * Fix EVP_dsa_sha macro.
13215 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13216 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13217 that was added in OpenSSL 0.9.6d.
13219 As the countermeasure turned out to be incompatible with some
13220 broken SSL implementations, the new option is part of SSL_OP_ALL.
13221 SSL_OP_ALL is usually employed when compatibility with weird SSL
13222 implementations is desired (e.g. '-bugs' option to 's_client' and
13223 's_server'), so the new option is automatically set in many
13228 * Changes in security patch:
13230 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13231 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13232 Air Force Materiel Command, USAF, under agreement number
13235 * Add various sanity checks to asn1_get_length() to reject
13236 the ASN1 length bytes if they exceed sizeof(long), will appear
13237 negative or the content length exceeds the length of the
13238 supplied buffer. [CVE-2002-0659][]
13240 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13242 * Assertions for various potential buffer overflows, not known to
13243 happen in practice.
13245 *Ben Laurie (CHATS)*
13247 * Various temporary buffers to hold ASCII versions of integers were
13248 too small for 64 bit platforms. [CVE-2002-0655][]
13249 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13251 * Remote buffer overflow in SSL3 protocol - an attacker could
13252 supply an oversized session ID to a client. [CVE-2002-0656][]
13254 *Ben Laurie (CHATS)*
13256 * Remote buffer overflow in SSL2 protocol - an attacker could
13257 supply an oversized client master key. [CVE-2002-0656][]
13259 *Ben Laurie (CHATS)*
13261 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13263 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13264 encoded as NULL) with id-dsa-with-sha1.
13266 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13268 * Check various `X509_...()` return values in apps/req.c.
13270 *Nils Larsch <nla@trustcenter.de>*
13272 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13273 an end-of-file condition would erroneously be flagged, when the CRLF
13274 was just at the end of a processed block. The bug was discovered when
13275 processing data through a buffering memory BIO handing the data to a
13276 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13277 <ptsekov@syntrex.com> and Nedelcho Stanev.
13281 * Implement a countermeasure against a vulnerability recently found
13282 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13283 before application data chunks to avoid the use of known IVs
13284 with data potentially chosen by the attacker.
13288 * Fix length checks in ssl3_get_client_hello().
13292 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13293 to prevent ssl3_read_internal() from incorrectly assuming that
13294 ssl3_read_bytes() found application data while handshake
13295 processing was enabled when in fact s->s3->in_read_app_data was
13296 merely automatically cleared during the initial handshake.
13298 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13300 * Fix object definitions for Private and Enterprise: they were not
13301 recognized in their shortname (=lowercase) representation. Extend
13302 obj_dat.pl to issue an error when using undefined keywords instead
13303 of silently ignoring the problem (Svenning Sorensen
13304 <sss@sss.dnsalias.net>).
13308 * Fix DH_generate_parameters() so that it works for 'non-standard'
13309 generators, i.e. generators other than 2 and 5. (Previously, the
13310 code did not properly initialise the 'add' and 'rem' values to
13311 BN_generate_prime().)
13313 In the new general case, we do not insist that 'generator' is
13314 actually a primitive root: This requirement is rather pointless;
13315 a generator of the order-q subgroup is just as good, if not
13320 * Map new X509 verification errors to alerts. Discovered and submitted by
13321 Tom Wu <tom@arcot.com>.
13325 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13326 returning non-zero before the data has been completely received
13327 when using non-blocking I/O.
13329 *Bodo Moeller; problem pointed out by John Hughes*
13331 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13333 *Ben Laurie, Lutz Jaenicke*
13335 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13336 Yoram Zahavi <YoramZ@gilian.com>).
13340 * Add information about CygWin 1.3 and on, and preserve proper
13341 configuration for the versions before that.
13343 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13345 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13346 check whether we deal with a copy of a session and do not delete from
13347 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13348 <izhar@checkpoint.com>.
13352 * Do not store session data into the internal session cache, if it
13353 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13354 flag is set). Proposed by Aslam <aslam@funk.com>.
13358 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13363 * [In 0.9.6d-engine release:]
13364 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13366 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13368 * Add the configuration target linux-s390x.
13370 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13372 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13373 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13374 variable as an indication that a ClientHello message has been
13375 received. As the flag value will be lost between multiple
13376 invocations of ssl3_accept when using non-blocking I/O, the
13377 function may not be aware that a handshake has actually taken
13378 place, thus preventing a new session from being added to the
13381 To avoid this problem, we now set s->new_session to 2 instead of
13382 using a local variable.
13384 *Lutz Jaenicke, Bodo Moeller*
13386 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13387 if the SSL_R_LENGTH_MISMATCH error is detected.
13389 *Geoff Thorpe, Bodo Moeller*
13391 * New 'shared_ldflag' column in Configure platform table.
13395 * Fix EVP_CIPHER_mode macro.
13397 *"Dan S. Camper" <dan@bti.net>*
13399 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13400 type, we must throw them away by setting rr->length to 0.
13402 *D P Chang <dpc@qualys.com>*
13404 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13406 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13407 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13408 worked incorrectly for those cases where range = `10..._2` and
13409 `3*range` is two bits longer than range.)
13413 * Only add signing time to PKCS7 structures if it is not already
13418 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13419 OBJ_ld_ce should be OBJ_id_ce.
13420 Also some ip-pda OIDs in crypto/objects/objects.txt were
13421 incorrect (cf. RFC 3039).
13423 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13425 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13426 returns early because it has nothing to do.
13428 *Andy Schneider <andy.schneider@bjss.co.uk>*
13430 * [In 0.9.6c-engine release:]
13431 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13433 *Andy Schneider <andy.schneider@bjss.co.uk>*
13435 * [In 0.9.6c-engine release:]
13436 Add support for Cryptographic Appliance's keyserver technology.
13437 (Use engine 'keyclient')
13439 *Cryptographic Appliances and Geoff Thorpe*
13441 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13442 is called via tools/c89.sh because arguments have to be
13443 rearranged (all '-L' options must appear before the first object
13446 *Richard Shapiro <rshapiro@abinitio.com>*
13448 * [In 0.9.6c-engine release:]
13449 Add support for Broadcom crypto accelerator cards, backported
13452 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13454 * [In 0.9.6c-engine release:]
13455 Add support for SureWare crypto accelerator cards from
13456 Baltimore Technologies. (Use engine 'sureware')
13458 *Baltimore Technologies and Mark Cox*
13460 * [In 0.9.6c-engine release:]
13461 Add support for crypto accelerator cards from Accelerated
13462 Encryption Processing, www.aep.ie. (Use engine 'aep')
13464 *AEP Inc. and Mark Cox*
13466 * Add a configuration entry for gcc on UnixWare.
13468 *Gary Benson <gbenson@redhat.com>*
13470 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13471 messages are stored in a single piece (fixed-length part and
13472 variable-length part combined) and fix various bugs found on the way.
13476 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13477 instead. BIO_gethostbyname() does not know what timeouts are
13478 appropriate, so entries would stay in cache even when they have
13480 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13482 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13483 faced with a pathologically small ClientHello fragment that does
13484 not contain client_version: Instead of aborting with an error,
13485 simply choose the highest available protocol version (i.e.,
13486 TLS 1.0 unless it is disabled). In practice, ClientHello
13487 messages are never sent like this, but this change gives us
13488 strictly correct behaviour at least for TLS.
13492 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13493 never resets s->method to s->ctx->method when called from within
13494 one of the SSL handshake functions.
13496 *Bodo Moeller; problem pointed out by Niko Baric*
13498 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13499 (sent using the client's version number) if client_version is
13500 smaller than the protocol version in use. Also change
13501 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13502 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13503 the client will at least see that alert.
13507 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13512 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13513 client receives HelloRequest while in a handshake.
13515 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13517 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13518 should end in 'break', not 'goto end' which circumvents various
13519 cleanups done in state SSL_ST_OK. But session related stuff
13520 must be disabled for SSL_ST_OK in the case that we just sent a
13523 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13524 before just sending a HelloRequest.
13526 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13528 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13529 reveal whether illegal block cipher padding was found or a MAC
13530 verification error occurred. (Neither SSLerr() codes nor alerts
13531 are directly visible to potential attackers, but the information
13532 may leak via logfiles.)
13534 Similar changes are not required for the SSL 2.0 implementation
13535 because the number of padding bytes is sent in clear for SSL 2.0,
13536 and the extra bytes are just ignored. However ssl/s2_pkt.c
13537 failed to verify that the purported number of padding bytes is in
13542 * Add OpenUNIX-8 support including shared libraries
13543 (Boyd Lynn Gerber <gerberb@zenez.com>).
13547 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13548 'wristwatch attack' using huge encoding parameters (cf.
13549 James H. Manger's CRYPTO 2001 paper). Note that the
13550 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13551 encoding parameters and hence was not vulnerable.
13555 * BN_sqr() bug fix.
13557 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13559 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13560 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13561 followed by modular reduction.
13563 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13565 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13566 equivalent based on BN_pseudo_rand() instead of BN_rand().
13570 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13571 This function was broken, as the check for a new client hello message
13572 to handle SGC did not allow these large messages.
13573 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13577 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13581 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13582 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13586 * Rework the configuration and shared library support for Tru64 Unix.
13587 The configuration part makes use of modern compiler features and
13588 still retains old compiler behavior for those that run older versions
13589 of the OS. The shared library support part includes a variant that
13590 uses the RPATH feature, and is available through the special
13591 configuration target "alpha-cc-rpath", which will never be selected
13594 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13596 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13597 with the same message size as in ssl3_get_certificate_request().
13598 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13599 messages might inadvertently be reject as too long.
13601 *Petr Lampa <lampa@fee.vutbr.cz>*
13603 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13607 * Modified SSL library such that the verify_callback that has been set
13608 specifically for an SSL object with SSL_set_verify() is actually being
13609 used. Before the change, a verify_callback set with this function was
13610 ignored and the verify_callback() set in the SSL_CTX at the time of
13611 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13612 to allow the necessary settings.
13616 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13617 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13618 done automatically (in contradiction to the requirements of the C
13619 standard). This made problems when used from OpenSSH.
13623 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13624 dh->length and always used
13626 BN_rand_range(priv_key, dh->p).
13628 BN_rand_range() is not necessary for Diffie-Hellman, and this
13629 specific range makes Diffie-Hellman unnecessarily inefficient if
13630 dh->length (recommended exponent length) is much smaller than the
13631 length of dh->p. We could use BN_rand_range() if the order of
13632 the subgroup was stored in the DH structure, but we only have
13637 BN_rand(priv_key, l, ...)
13639 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13646 RSA_eay_public_encrypt
13647 RSA_eay_private_decrypt
13648 RSA_eay_private_encrypt (signing)
13649 RSA_eay_public_decrypt (signature verification)
13651 (default implementations for RSA_public_encrypt,
13652 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13653 always reject numbers >= n.
13657 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13658 to synchronize access to 'locking_thread'. This is necessary on
13659 systems where access to 'locking_thread' (an 'unsigned long'
13660 variable) is not atomic.
13664 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13665 *before* setting the 'crypto_lock_rand' flag. The previous code had
13666 a race condition if 0 is a valid thread ID.
13668 *Travis Vitek <vitek@roguewave.com>*
13670 * Add support for shared libraries under Irix.
13672 *Albert Chin-A-Young <china@thewrittenword.com>*
13674 * Add configuration option to build on Linux on both big-endian and
13675 little-endian MIPS.
13677 *Ralf Baechle <ralf@uni-koblenz.de>*
13679 * Add the possibility to create shared libraries on HP-UX.
13683 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13685 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13686 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13687 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13688 PRNG state recovery was possible based on the output of
13689 one PRNG request appropriately sized to gain knowledge on
13690 'md' followed by enough consecutive 1-byte PRNG requests
13691 to traverse all of 'state'.
13693 1. When updating 'md_local' (the current thread's copy of 'md')
13694 during PRNG output generation, hash all of the previous
13695 'md_local' value, not just the half used for PRNG output.
13697 2. Make the number of bytes from 'state' included into the hash
13698 independent from the number of PRNG bytes requested.
13700 The first measure alone would be sufficient to avoid
13701 Markku-Juhani's attack. (Actually it had never occurred
13702 to me that the half of 'md_local' used for chaining was the
13703 half from which PRNG output bytes were taken -- I had always
13704 assumed that the secret half would be used.) The second
13705 measure makes sure that additional data from 'state' is never
13706 mixed into 'md_local' in small portions; this heuristically
13707 further strengthens the PRNG.
13711 * Fix crypto/bn/asm/mips3.s.
13715 * When only the key is given to "enc", the IV is undefined. Print out
13716 an error message in this case.
13720 * Handle special case when X509_NAME is empty in X509 printing routines.
13724 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13725 positive and less than q.
13729 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13730 used: it isn't thread safe and the add_lock_callback should handle
13733 *Paul Rose <Paul.Rose@bridge.com>*
13735 * Verify that incoming data obeys the block size in
13736 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13742 *Ulf Möller, Bodo Möller*
13744 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13745 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13746 when fixing the server behaviour for backwards-compatible 'client
13747 hello' messages. (Note that the attack is impractical against
13748 SSL 3.0 and TLS 1.0 anyway because length and version checking
13749 means that the probability of guessing a valid ciphertext is
13750 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13753 Before 0.9.5, the countermeasure (hide the error by generating a
13754 random 'decryption result') did not work properly because
13755 ERR_clear_error() was missing, meaning that SSL_get_error() would
13756 detect the supposedly ignored error.
13758 Both problems are now fixed.
13762 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13763 (previously it was 1024).
13767 * Fix for compatibility mode trust settings: ignore trust settings
13768 unless some valid trust or reject settings are present.
13772 * Fix for blowfish EVP: its a variable length cipher.
13776 * Fix various bugs related to DSA S/MIME verification. Handle missing
13777 parameters in DSA public key structures and return an error in the
13778 DSA routines if parameters are absent.
13782 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13783 in the current directory if neither $RANDFILE nor $HOME was set.
13784 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13785 caused some confusion to Windows users who haven't defined $HOME.
13786 Thus RAND_file_name() is changed again: e_os.h can define a
13787 DEFAULT_HOME, which will be used if $HOME is not set.
13788 For Windows, we use "C:"; on other platforms, we still require
13789 environment variables.
13791 * Move 'if (!initialized) RAND_poll()' into regions protected by
13792 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13793 having multiple threads call RAND_poll() concurrently.
13797 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13798 combination of a flag and a thread ID variable.
13799 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13800 flag), *other* threads can enter ssleay_add_bytes without obeying
13801 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13802 that they do not hold after the first thread unsets add_do_not_lock).
13806 * Change bctest again: '-x' expressions are not available in all
13807 versions of 'test'.
13811 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13813 * Fix a couple of memory leaks in PKCS7_dataDecode()
13815 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13817 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13818 the default extension for executables, if any. Also, make the perl
13819 scripts that use symlink() to test if it really exists and use "cp"
13820 if it doesn't. All this made OpenSSL compilable and installable in
13825 * Fix for asn1_GetSequence() for indefinite length constructed data.
13826 If SEQUENCE is length is indefinite just set c->slen to the total
13827 amount of data available.
13829 *Steve Henson, reported by shige@FreeBSD.org*
13831 *This change does not apply to 0.9.7.*
13833 * Change bctest to avoid here-documents inside command substitution
13834 (workaround for FreeBSD /bin/sh bug).
13835 For compatibility with Ultrix, avoid shell functions (introduced
13836 in the bctest version that searches along $PATH).
13840 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13841 with des_encrypt() defined on some operating systems, like Solaris
13846 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13847 On the Importance of Eliminating Errors in Cryptographic
13848 Computations, J. Cryptology 14 (2001) 2, 101-119,
13849 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13853 * MIPS assembler BIGNUM division bug fix.
13857 * Disabled incorrect Alpha assembler code.
13861 * Fix PKCS#7 decode routines so they correctly update the length
13862 after reading an EOC for the EXPLICIT tag.
13866 *This change does not apply to 0.9.7.*
13868 * Fix bug in PKCS#12 key generation routines. This was triggered
13869 if a 3DES key was generated with a 0 initial byte. Include
13870 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13871 (but broken) behaviour.
13875 * Enhance bctest to search for a working bc along $PATH and print
13878 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13880 * Fix memory leaks in err.c: free err_data string if necessary;
13881 don't write to the wrong index in ERR_set_error_data.
13885 * Implement ssl23_peek (analogous to ssl23_read), which previously
13890 * Replace rdtsc with `_emit` statements for VC++ version 5.
13892 *Jeremy Cooper <jeremy@baymoo.org>*
13894 * Make it possible to reuse SSLv2 sessions.
13898 * In copy_email() check for >= 0 as a return value for
13899 X509_NAME_get_index_by_NID() since 0 is a valid index.
13901 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13903 * Avoid coredump with unsupported or invalid public keys by checking if
13904 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13905 PKCS7_verify() fails with non detached data.
13909 * Don't use getenv in library functions when run as setuid/setgid.
13910 New function OPENSSL_issetugid().
13914 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13915 due to incorrect handling of multi-threading:
13917 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13919 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13921 3. Count how many times MemCheck_off() has been called so that
13922 nested use can be treated correctly. This also avoids
13923 inband-signalling in the previous code (which relied on the
13924 assumption that thread ID 0 is impossible).
13928 * Add "-rand" option also to s_client and s_server.
13932 * Fix CPU detection on Irix 6.x.
13933 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13934 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13936 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13941 *This change does not apply to 0.9.7.*
13943 * Use the cached encoding of an X509_NAME structure rather than
13944 copying it. This is apparently the reason for the libsafe "errors"
13945 but the code is actually correct.
13949 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13950 Bleichenbacher's DSA attack.
13951 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13952 to be set and top=0 forces the highest bit to be set; top=-1 is new
13953 and leaves the highest bit random.
13955 *Ulf Moeller, Bodo Moeller*
13957 * In the `NCONF_...`-based implementations for `CONF_...` queries
13958 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13959 a temporary CONF structure with the data component set to NULL
13960 (which gives segmentation faults in lh_retrieve).
13961 Instead, use NULL for the CONF pointer in CONF_get_string and
13962 CONF_get_number (which may use environment variables) and directly
13963 return NULL from CONF_get_section.
13967 * Fix potential buffer overrun for EBCDIC.
13971 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13972 keyUsage if basicConstraints absent for a CA.
13976 * Make SMIME_write_PKCS7() write mail header values with a format that
13977 is more generally accepted (no spaces before the semicolon), since
13978 some programs can't parse those values properly otherwise. Also make
13979 sure BIO's that break lines after each write do not create invalid
13984 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13985 macros previously used would not encode an empty SEQUENCE OF
13986 and break the signature.
13990 *This change does not apply to 0.9.7.*
13992 * Zero the premaster secret after deriving the master secret in
13997 * Add some EVP_add_digest_alias registrations (as found in
13998 OpenSSL_add_all_digests()) to SSL_library_init()
13999 aka OpenSSL_add_ssl_algorithms(). This provides improved
14000 compatibility with peers using X.509 certificates
14001 with unconventional AlgorithmIdentifier OIDs.
14005 * Fix for Irix with NO_ASM.
14007 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14009 * ./config script fixes.
14011 *Ulf Moeller, Richard Levitte*
14013 * Fix 'openssl passwd -1'.
14017 * Change PKCS12_key_gen_asc() so it can cope with non null
14018 terminated strings whose length is passed in the passlen
14019 parameter, for example from PEM callbacks. This was done
14020 by adding an extra length parameter to asc2uni().
14022 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14024 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14025 call failed, free the DSA structure.
14029 * Fix to uni2asc() to cope with zero length Unicode strings.
14030 These are present in some PKCS#12 files.
14034 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14035 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14036 when writing a 32767 byte record.
14038 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14040 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14041 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14043 (RSA objects have a reference count access to which is protected
14044 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14045 so they are meant to be shared between threads.)
14046 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14047 "Reddie, Steven" <Steven.Reddie@ca.com>*
14049 * Fix a deadlock in CRYPTO_mem_leaks().
14053 * Use better test patterns in bntest.
14057 * rand_win.c fix for Borland C.
14061 * BN_rshift bugfix for n == 0.
14065 * Add a 'bctest' script that checks for some known 'bc' bugs
14066 so that 'make test' does not abort just because 'bc' is broken.
14070 * Store verify_result within SSL_SESSION also for client side to
14071 avoid potential security hole. (Re-used sessions on the client side
14072 always resulted in verify_result==X509_V_OK, not using the original
14073 result of the server certificate verification.)
14077 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14078 SSL3_RT_APPLICATION_DATA, return 0.
14079 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14084 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14085 releases, have been re-implemented by renaming the previous
14086 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14087 and ssl3_read_internal, respectively, and adding 'peek' parameters
14088 to them. The new ssl[23]_{read,peek} functions are calls to
14089 ssl[23]_read_internal with the 'peek' flag set appropriately.
14090 A 'peek' parameter has also been added to ssl3_read_bytes, which
14091 does the actual work for ssl3_read_internal.
14095 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14096 the method-specific "init()" handler. Also clean up ex_data after
14097 calling the method-specific "finish()" handler. Previously, this was
14098 happening the other way round.
14102 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14103 The previous value, 12, was not always sufficient for BN_mod_exp().
14107 * Make sure that shared libraries get the internal name engine with
14108 the full version number and not just 0. This should mark the
14109 shared libraries as not backward compatible. Of course, this should
14110 be changed again when we can guarantee backward binary compatibility.
14114 * Fix typo in get_cert_by_subject() in by_dir.c
14116 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14118 * Rework the system to generate shared libraries:
14120 - Make note of the expected extension for the shared libraries and
14121 if there is a need for symbolic links from for example libcrypto.so.0
14122 to libcrypto.so.0.9.7. There is extended info in Configure for
14125 - Make as few rebuilds of the shared libraries as possible.
14127 - Still avoid linking the OpenSSL programs with the shared libraries.
14129 - When installing, install the shared libraries separately from the
14134 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14136 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14137 and not in SSL_clear because the latter is also used by the
14138 accept/connect functions; previously, the settings made by
14139 SSL_set_read_ahead would be lost during the handshake.
14141 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14143 * Correct util/mkdef.pl to be selective about disabled algorithms.
14144 Previously, it would create entries for disabled algorithms no
14149 * Added several new manual pages for SSL_* function.
14153 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14155 * In ssl23_get_client_hello, generate an error message when faced
14156 with an initial SSL 3.0/TLS record that is too small to contain the
14157 first two bytes of the ClientHello message, i.e. client_version.
14158 (Note that this is a pathologic case that probably has never happened
14159 in real life.) The previous approach was to use the version number
14160 from the record header as a substitute; but our protocol choice
14161 should not depend on that one because it is not authenticated
14162 by the Finished messages.
14166 * More robust randomness gathering functions for Windows.
14168 *Jeffrey Altman <jaltman@columbia.edu>*
14170 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14171 not set then we don't setup the error code for issuer check errors
14172 to avoid possibly overwriting other errors which the callback does
14173 handle. If an application does set the flag then we assume it knows
14174 what it is doing and can handle the new informational codes
14179 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14180 a general "ANY" type, as such it should be able to decode anything
14181 including tagged types. However it didn't check the class so it would
14182 wrongly interpret tagged types in the same way as their universal
14183 counterpart and unknown types were just rejected. Changed so that the
14184 tagged and unknown types are handled in the same way as a SEQUENCE:
14185 that is the encoding is stored intact. There is also a new type
14186 "V_ASN1_OTHER" which is used when the class is not universal, in this
14187 case we have no idea what the actual type is so we just lump them all
14192 * On VMS, stdout may very well lead to a file that is written to
14193 in a record-oriented fashion. That means that every write() will
14194 write a separate record, which will be read separately by the
14195 programs trying to read from it. This can be very confusing.
14197 The solution is to put a BIO filter in the way that will buffer
14198 text until a linefeed is reached, and then write everything a
14199 line at a time, so every record written will be an actual line,
14200 not chunks of lines and not (usually doesn't happen, but I've
14201 seen it once) several lines in one record. BIO_f_linebuffer() is
14204 Currently, it's a VMS-only method, because that's where it has
14205 been tested well enough.
14209 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14210 it can return incorrect results.
14211 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14212 but it was in 0.9.6-beta[12].)
14216 * Disable the check for content being present when verifying detached
14217 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14218 include zero length content when signing messages.
14222 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14223 BIO_ctrl (for BIO pairs).
14227 * Add DSO method for VMS.
14231 * Bug fix: Montgomery multiplication could produce results with the
14236 * Add RPM specification openssl.spec and modify it to build three
14237 packages. The default package contains applications, application
14238 documentation and run-time libraries. The devel package contains
14239 include files, static libraries and function documentation. The
14240 doc package contains the contents of the doc directory. The original
14241 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14245 * Add a large number of documentation files for many SSL routines.
14247 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14249 * Add a configuration entry for Sony News 4.
14251 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14253 * Don't set the two most significant bits to one when generating a
14254 random number < q in the DSA library.
14258 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14259 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14260 the underlying transport is blocking) if a handshake took place.
14261 (The default behaviour is needed by applications such as s_client
14262 and s_server that use select() to determine when to use SSL_read;
14263 but for applications that know in advance when to expect data, it
14264 just makes things more complicated.)
14268 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14273 * Add a few more EBCDIC conditionals that make `req` and `x509`
14274 work better on such systems.
14276 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14278 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14279 Update PKCS12_parse() so it copies the friendlyName and the
14280 keyid to the certificates aux info.
14284 * Fix bug in PKCS7_verify() which caused an infinite loop
14285 if there was more than one signature.
14287 *Sven Uszpelkat <su@celocom.de>*
14289 * Major change in util/mkdef.pl to include extra information
14290 about each symbol, as well as presenting variables as well
14291 as functions. This change means that there's n more need
14292 to rebuild the .num files when some algorithms are excluded.
14296 * Allow the verify time to be set by an application,
14297 rather than always using the current time.
14301 * Phase 2 verify code reorganisation. The certificate
14302 verify code now looks up an issuer certificate by a
14303 number of criteria: subject name, authority key id
14304 and key usage. It also verifies self signed certificates
14305 by the same criteria. The main comparison function is
14306 X509_check_issued() which performs these checks.
14308 Lot of changes were necessary in order to support this
14309 without completely rewriting the lookup code.
14311 Authority and subject key identifier are now cached.
14313 The LHASH 'certs' is X509_STORE has now been replaced
14314 by a STACK_OF(X509_OBJECT). This is mainly because an
14315 LHASH can't store or retrieve multiple objects with
14316 the same hash value.
14318 As a result various functions (which were all internal
14319 use only) have changed to handle the new X509_STORE
14320 structure. This will break anything that messed round
14321 with X509_STORE internally.
14323 The functions X509_STORE_add_cert() now checks for an
14324 exact match, rather than just subject name.
14326 The X509_STORE API doesn't directly support the retrieval
14327 of multiple certificates matching a given criteria, however
14328 this can be worked round by performing a lookup first
14329 (which will fill the cache with candidate certificates)
14330 and then examining the cache for matches. This is probably
14331 the best we can do without throwing out X509_LOOKUP
14332 entirely (maybe later...).
14334 The X509_VERIFY_CTX structure has been enhanced considerably.
14336 All certificate lookup operations now go via a get_issuer()
14337 callback. Although this currently uses an X509_STORE it
14338 can be replaced by custom lookups. This is a simple way
14339 to bypass the X509_STORE hackery necessary to make this
14340 work and makes it possible to use more efficient techniques
14341 in future. A very simple version which uses a simple
14342 STACK for its trusted certificate store is also provided
14343 using X509_STORE_CTX_trusted_stack().
14345 The verify_cb() and verify() callbacks now have equivalents
14346 in the X509_STORE_CTX structure.
14348 X509_STORE_CTX also has a 'flags' field which can be used
14349 to customise the verify behaviour.
14353 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14354 excludes S/MIME capabilities.
14358 * When a certificate request is read in keep a copy of the
14359 original encoding of the signed data and use it when outputting
14360 again. Signatures then use the original encoding rather than
14361 a decoded, encoded version which may cause problems if the
14362 request is improperly encoded.
14366 * For consistency with other BIO_puts implementations, call
14367 buffer_write(b, ...) directly in buffer_puts instead of calling
14370 In BIO_puts, increment b->num_write as in BIO_write.
14372 *Peter.Sylvester@EdelWeb.fr*
14374 * Fix BN_mul_word for the case where the word is 0. (We have to use
14375 BN_zero, we may not return a BIGNUM with an array consisting of
14376 words set to zero.)
14380 * Avoid calling abort() from within the library when problems are
14381 detected, except if preprocessor symbols have been defined
14382 (such as REF_CHECK, BN_DEBUG etc.).
14386 * New openssl application 'rsautl'. This utility can be
14387 used for low level RSA operations. DER public key
14388 BIO/fp routines also added.
14392 * New Configure entry and patches for compiling on QNX 4.
14394 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14396 * A demo state-machine implementation was sponsored by
14397 Nuron (<http://www.nuron.com/>) and is now available in
14398 demos/state_machine.
14402 * New options added to the 'dgst' utility for signature
14403 generation and verification.
14407 * Unrecognized PKCS#7 content types are now handled via a
14408 catch all ASN1_TYPE structure. This allows unsupported
14409 types to be stored as a "blob" and an application can
14410 encode and decode it manually.
14414 * Fix various signed/unsigned issues to make a_strex.c
14415 compile under VC++.
14417 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14419 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14420 length if passed a buffer. ASN1_INTEGER_to_BN failed
14421 if passed a NULL BN and its argument was negative.
14423 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14425 * Modification to PKCS#7 encoding routines to output definite
14426 length encoding. Since currently the whole structures are in
14427 memory there's not real point in using indefinite length
14428 constructed encoding. However if OpenSSL is compiled with
14429 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14433 * Added BIO_vprintf() and BIO_vsnprintf().
14437 * Added more prefixes to parse for in the strings written
14438 through a logging bio, to cover all the levels that are available
14439 through syslog. The prefixes are now:
14441 PANIC, EMERG, EMR => LOG_EMERG
14442 ALERT, ALR => LOG_ALERT
14443 CRIT, CRI => LOG_CRIT
14444 ERROR, ERR => LOG_ERR
14445 WARNING, WARN, WAR => LOG_WARNING
14446 NOTICE, NOTE, NOT => LOG_NOTICE
14447 INFO, INF => LOG_INFO
14448 DEBUG, DBG => LOG_DEBUG
14450 and as before, if none of those prefixes are present at the
14451 beginning of the string, LOG_ERR is chosen.
14453 On Win32, the `LOG_*` levels are mapped according to this:
14455 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14456 LOG_WARNING => EVENTLOG_WARNING_TYPE
14457 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14461 * Made it possible to reconfigure with just the configuration
14462 argument "reconf" or "reconfigure". The command line arguments
14463 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14464 and are retrieved from there when reconfiguring.
14470 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14472 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14476 * The obj_dat.pl script was messing up the sorting of object
14477 names. The reason was that it compared the quoted version
14478 of strings as a result "OCSP" > "OCSP Signing" because
14479 " > SPACE. Changed script to store unquoted versions of
14480 names and add quotes on output. It was also omitting some
14481 names from the lookup table if they were given a default
14482 value (that is if SN is missing it is given the same
14483 value as LN and vice versa), these are now added on the
14484 grounds that if an object has a name we should be able to
14485 look it up. Finally added warning output when duplicate
14486 short or long names are found.
14490 * Changes needed for Tandem NSK.
14492 *Scott Uroff <scott@xypro.com>*
14494 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14495 RSA_padding_check_SSLv23(), special padding was never detected
14496 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14497 version rollback attacks was not effective.
14499 In s23_clnt.c, don't use special rollback-attack detection padding
14500 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14501 client; similarly, in s23_srvr.c, don't do the rollback check if
14502 SSL 2.0 is the only protocol enabled in the server.
14506 * Make it possible to get hexdumps of unprintable data with 'openssl
14507 asn1parse'. By implication, the functions ASN1_parse_dump() and
14508 BIO_dump_indent() are added.
14512 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14513 these print out strings and name structures based on various
14514 flags including RFC2253 support and proper handling of
14515 multibyte characters. Added options to the 'x509' utility
14516 to allow the various flags to be set.
14520 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14521 Also change the functions X509_cmp_current_time() and
14522 X509_gmtime_adj() work with an ASN1_TIME structure,
14523 this will enable certificates using GeneralizedTime in validity
14524 dates to be checked.
14528 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14529 negative public key encodings) on by default,
14530 NO_NEG_PUBKEY_BUG can be set to disable it.
14534 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14535 content octets. An i2c_ASN1_OBJECT is unnecessary because
14536 the encoding can be trivially obtained from the structure.
14540 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14541 not read locks (`CRYPTO_r_[un]lock`).
14545 * A first attempt at creating official support for shared
14546 libraries through configuration. I've kept it so the
14547 default is static libraries only, and the OpenSSL programs
14548 are always statically linked for now, but there are
14549 preparations for dynamic linking in place.
14550 This has been tested on Linux and Tru64.
14554 * Randomness polling function for Win9x, as described in:
14555 Peter Gutmann, Software Generation of Practically Strong
14560 * Fix so PRNG is seeded in req if using an already existing
14565 * New options to smime application. -inform and -outform
14566 allow alternative formats for the S/MIME message including
14567 PEM and DER. The -content option allows the content to be
14568 specified separately. This should allow things like Netscape
14569 form signing output easier to verify.
14573 * Fix the ASN1 encoding of tags using the 'long form'.
14577 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14578 STRING types. These convert content octets to and from the
14579 underlying type. The actual tag and length octets are
14580 already assumed to have been read in and checked. These
14581 are needed because all other string types have virtually
14582 identical handling apart from the tag. By having versions
14583 of the ASN1 functions that just operate on content octets
14584 IMPLICIT tagging can be handled properly. It also allows
14585 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14586 and ASN1_INTEGER are identical apart from the tag.
14590 * Change the handling of OID objects as follows:
14592 - New object identifiers are inserted in objects.txt, following
14593 the syntax given in objects.README.
14594 - objects.pl is used to process obj_mac.num and create a new
14596 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14599 This is currently kind of a hack, and the perl code in objects.pl
14600 isn't very elegant, but it works as I intended. The simplest way
14601 to check that it worked correctly is to look in obj_dat.h and
14602 check the array nid_objs and make sure the objects haven't moved
14603 around (this is important!). Additions are OK, as well as
14604 consistent name changes.
14608 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14612 * Addition of the command line parameter '-rand file' to 'openssl req'.
14613 The given file adds to whatever has already been seeded into the
14614 random pool through the RANDFILE configuration file option or
14615 environment variable, or the default random state file.
14619 * mkstack.pl now sorts each macro group into lexical order.
14620 Previously the output order depended on the order the files
14621 appeared in the directory, resulting in needless rewriting
14626 * Patches to make OpenSSL compile under Win32 again. Mostly
14627 work arounds for the VC++ problem that it treats func() as
14628 func(void). Also stripped out the parts of mkdef.pl that
14629 added extra typesafe functions: these no longer exist.
14633 * Reorganisation of the stack code. The macros are now all
14634 collected in safestack.h . Each macro is defined in terms of
14635 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14636 DEBUG_SAFESTACK is now handled in terms of function casts,
14637 this has the advantage of retaining type safety without the
14638 use of additional functions. If DEBUG_SAFESTACK is not defined
14639 then the non typesafe macros are used instead. Also modified the
14640 mkstack.pl script to handle the new form. Needs testing to see
14641 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14642 the default if no major problems. Similar behaviour for ASN1_SET_OF
14643 and PKCS12_STACK_OF.
14647 * When some versions of IIS use the 'NET' form of private key the
14648 key derivation algorithm is different. Normally MD5(password) is
14649 used as a 128 bit RC4 key. In the modified case
14650 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14651 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14652 as the old Netscape_RSA functions except they have an additional
14653 'sgckey' parameter which uses the modified algorithm. Also added
14654 an -sgckey command line option to the rsa utility. Thanks to
14655 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14656 algorithm to openssl-dev.
14660 * The evp_local.h macros were using 'c.##kname' which resulted in
14661 invalid expansion on some systems (SCO 5.0.5 for example).
14662 Corrected to 'c.kname'.
14664 *Phillip Porch <root@theporch.com>*
14666 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14667 a STACK of email addresses from a certificate or request, these look
14668 in the subject name and the subject alternative name extensions and
14669 omit any duplicate addresses.
14673 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14674 This makes DSA verification about 2 % faster.
14678 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14679 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14680 plus overhead for 1024 bit moduli).
14681 This makes exponentiations about 0.5 % faster for 1024 bit
14682 exponents (as measured by "openssl speed rsa2048").
14686 * Rename memory handling macros to avoid conflicts with other
14688 Malloc => OPENSSL_malloc
14689 Malloc_locked => OPENSSL_malloc_locked
14690 Realloc => OPENSSL_realloc
14691 Free => OPENSSL_free
14695 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14696 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14700 * CygWin32 support.
14702 *John Jarvie <jjarvie@newsguy.com>*
14704 * The type-safe stack code has been rejigged. It is now only compiled
14705 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14706 by default all type-specific stack functions are "#define"d back to
14707 standard stack functions. This results in more streamlined output
14708 but retains the type-safety checking possibilities of the original
14713 * The STACK code has been cleaned up, and certain type declarations
14714 that didn't make a lot of sense have been brought in line. This has
14715 also involved a cleanup of sorts in safestack.h to more correctly
14716 map type-safe stack functions onto their plain stack counterparts.
14717 This work has also resulted in a variety of "const"ifications of
14718 lots of the code, especially `_cmp` operations which should normally
14719 be prototyped with "const" parameters anyway.
14723 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14724 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14725 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14726 where all of 'md' is used each time the PRNG is used, but 'state'
14727 is used only indexed by a cyclic counter. As entropy may not be
14728 well distributed from the beginning, 'md' is important as a
14729 chaining variable. However, the output function chains only half
14730 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14731 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14732 in all of 'state' being rewritten, with the new values depending
14733 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14737 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14738 the handshake is continued after ssl_verify_cert_chain();
14739 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14740 can lead to 'unexplainable' connection aborts later.
14742 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14744 * Major EVP API cipher revision.
14745 Add hooks for extra EVP features. This allows various cipher
14746 parameters to be set in the EVP interface. Support added for variable
14747 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14748 setting of RC2 and RC5 parameters.
14750 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14753 Remove lots of duplicated code from the EVP library. For example *every*
14754 cipher init() function handles the 'iv' in the same way according to the
14755 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14756 for CFB and OFB modes they zero ctx->num.
14758 New functionality allows removal of S/MIME code RC2 hack.
14760 Most of the routines have the same form and so can be declared in terms
14763 By shifting this to the top level EVP_CipherInit() it can be removed from
14764 all individual ciphers. If the cipher wants to handle IVs or keys
14765 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14768 Change lots of functions like EVP_EncryptUpdate() to now return a
14769 value: although software versions of the algorithms cannot fail
14770 any installed hardware versions can.
14774 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14775 this option is set, tolerate broken clients that send the negotiated
14776 protocol version number instead of the requested protocol version
14781 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14782 i.e. non-zero for export ciphersuites, zero otherwise.
14783 Previous versions had this flag inverted, inconsistent with
14784 rsa_tmp_cb (..._TMP_RSA_CB).
14786 *Bodo Moeller; problem reported by Amit Chopra*
14788 * Add missing DSA library text string. Work around for some IIS
14789 key files with invalid SEQUENCE encoding.
14793 * Add a document (doc/standards.txt) that list all kinds of standards
14794 and so on that are implemented in OpenSSL.
14798 * Enhance c_rehash script. Old version would mishandle certificates
14799 with the same subject name hash and wouldn't handle CRLs at all.
14800 Added -fingerprint option to crl utility, to support new c_rehash
14805 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14809 * Fix for SSL server purpose checking. Server checking was
14810 rejecting certificates which had extended key usage present
14811 but no ssl client purpose.
14813 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14815 * Make PKCS#12 code work with no password. The PKCS#12 spec
14816 is a little unclear about how a blank password is handled.
14817 Since the password in encoded as a BMPString with terminating
14818 double NULL a zero length password would end up as just the
14819 double NULL. However no password at all is different and is
14820 handled differently in the PKCS#12 key generation code. NS
14821 treats a blank password as zero length. MSIE treats it as no
14822 password on export: but it will try both on import. We now do
14823 the same: PKCS12_parse() tries zero length and no password if
14824 the password is set to "" or NULL (NULL is now a valid password:
14825 it wasn't before) as does the pkcs12 application.
14829 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14830 perror when PEM_read_bio_X509_REQ fails, the error message must
14831 be obtained from the error queue.
14835 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14836 it in ERR_remove_state if appropriate, and change ERR_get_state
14837 accordingly to avoid race conditions (this is necessary because
14838 thread_hash is no longer constant once set).
14842 * Bugfix for linux-elf makefile.one.
14846 * RSA_get_default_method() will now cause a default
14847 RSA_METHOD to be chosen if one doesn't exist already.
14848 Previously this was only set during a call to RSA_new()
14849 or RSA_new_method(NULL) meaning it was possible for
14850 RSA_get_default_method() to return NULL.
14854 * Added native name translation to the existing DSO code
14855 that will convert (if the flag to do so is set) filenames
14856 that are sufficiently small and have no path information
14857 into a canonical native form. Eg. "blah" converted to
14858 "libblah.so" or "blah.dll" etc.
14862 * New function ERR_error_string_n(e, buf, len) which is like
14863 ERR_error_string(e, buf), but writes at most 'len' bytes
14864 including the 0 terminator. For ERR_error_string_n, 'buf'
14867 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14869 * CONF library reworked to become more general. A new CONF
14870 configuration file reader "class" is implemented as well as a
14871 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14872 old `CONF_*` functions are still there, but are reimplemented to
14873 work in terms of the new functions. Also, a set of functions
14874 to handle the internal storage of the configuration data is
14875 provided to make it easier to write new configuration file
14876 reader "classes" (I can definitely see something reading a
14877 configuration file in XML format, for example), called `_CONF_*`,
14878 or "the configuration storage API"...
14880 The new configuration file reading functions are:
14882 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14883 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14885 NCONF_default, NCONF_WIN32
14887 NCONF_dump_fp, NCONF_dump_bio
14889 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14890 NCONF_new creates a new CONF object. This works in the same way
14891 as other interfaces in OpenSSL, like the BIO interface.
14892 `NCONF_dump_*` dump the internal storage of the configuration file,
14893 which is useful for debugging. All other functions take the same
14894 arguments as the old `CONF_*` functions with the exception of the
14895 first that must be a `CONF *` instead of a `LHASH *`.
14897 To make it easier to use the new classes with the old `CONF_*` functions,
14898 the function CONF_set_default_method is provided.
14902 * Add '-tls1' option to 'openssl ciphers', which was already
14903 mentioned in the documentation but had not been implemented.
14904 (This option is not yet really useful because even the additional
14905 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14909 * Initial DSO code added into libcrypto for letting OpenSSL (and
14910 OpenSSL-based applications) load shared libraries and bind to
14911 them in a portable way.
14913 *Geoff Thorpe, with contributions from Richard Levitte*
14915 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
14917 * Make sure _lrotl and _lrotr are only used with MSVC.
14919 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14920 (the default implementation of RAND_status).
14922 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14923 to '-clrext' (= clear extensions), as intended and documented.
14924 *Bodo Moeller; inconsistency pointed out by Michael Attili
14925 <attili@amaxo.com>*
14927 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14928 was larger than the MD block size.
14930 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14932 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14933 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14934 using the passed key: if the passed key was a private key the result
14935 of X509_print(), for example, would be to print out all the private key
14940 * des_quad_cksum() byte order bug fix.
14941 *Ulf Möller, using the problem description in krb4-0.9.7, where
14942 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14944 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14947 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14949 * For easily testing in shell scripts whether some command
14950 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14951 returns with exit code 0 iff no command of the given name is available.
14952 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14953 the output goes to stdout and nothing is printed to stderr.
14954 Additional arguments are always ignored.
14956 Since for each cipher there is a command of the same name,
14957 the 'no-cipher' compilation switches can be tested this way.
14959 ('openssl no-XXX' is not able to detect pseudo-commands such
14960 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14964 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14968 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14969 is set; it will be thrown away anyway because each handshake creates
14971 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14972 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14973 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14974 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14978 * New s_client option -ign_eof: EOF at stdin is ignored, and
14979 'Q' and 'R' lose their special meanings (quit/renegotiate).
14980 This is part of what -quiet does; unlike -quiet, -ign_eof
14981 does not suppress any output.
14985 * Add compatibility options to the purpose and trust code. The
14986 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14987 accepts a certificate or CA, this was the previous behaviour,
14988 with all the associated security issues.
14990 X509_TRUST_COMPAT is the old trust behaviour: only and
14991 automatically trust self signed roots in certificate store. A
14992 new trust setting X509_TRUST_DEFAULT is used to specify that
14993 a purpose has no associated trust setting and it should instead
14994 use the value in the default purpose.
14998 * Fix the PKCS#8 DSA private key code so it decodes keys again
14999 and fix a memory leak.
15003 * In util/mkerr.pl (which implements 'make errors'), preserve
15004 reason strings from the previous version of the .c file, as
15005 the default to have only downcase letters (and digits) in
15006 automatically generated reasons codes is not always appropriate.
15010 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15011 using strerror. Previously, ERR_reason_error_string() returned
15012 library names as reason strings for SYSerr; but SYSerr is a special
15013 case where small numbers are errno values, not library numbers.
15017 * Add '-dsaparam' option to 'openssl dhparam' application. This
15018 converts DSA parameters into DH parameters. (When creating parameters,
15019 DSA_generate_parameters is used.)
15023 * Include 'length' (recommended exponent length) in C code generated
15024 by 'openssl dhparam -C'.
15028 * The second argument to set_label in perlasm was already being used
15029 so couldn't be used as a "file scope" flag. Moved to third argument
15034 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15035 instead of RAND_bytes for encryption IVs and salts.
15039 * Include RAND_status() into RAND_METHOD instead of implementing
15040 it only for md_rand.c Otherwise replacing the PRNG by calling
15041 RAND_set_rand_method would be impossible.
15045 * Don't let DSA_generate_key() enter an infinite loop if the random
15046 number generation fails.
15050 * New 'rand' application for creating pseudo-random output.
15054 * Added configuration support for Linux/IA64
15056 *Rolf Haberrecker <rolf@suse.de>*
15058 * Assembler module support for Mingw32.
15062 * Shared library support for HPUX (in shlib/).
15064 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15066 * Shared library support for Solaris gcc.
15068 *Lutz Behnke <behnke@trustcenter.de>*
15070 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15072 * PKCS7_encrypt() was adding text MIME headers twice because they
15073 were added manually and by SMIME_crlf_copy().
15077 * In bntest.c don't call BN_rand with zero bits argument.
15079 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15081 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15082 case was implemented. This caused BN_div_recp() to fail occasionally.
15086 * Add an optional second argument to the set_label() in the perl
15087 assembly language builder. If this argument exists and is set
15088 to 1 it signals that the assembler should use a symbol whose
15089 scope is the entire file, not just the current function. This
15090 is needed with MASM which uses the format label:: for this scope.
15092 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15094 * Change the ASN1 types so they are typedefs by default. Before
15095 almost all types were #define'd to ASN1_STRING which was causing
15096 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15101 * Change names of new functions to the new get1/get0 naming
15102 convention: After 'get1', the caller owns a reference count
15103 and has to call `..._free`; 'get0' returns a pointer to some
15104 data structure without incrementing reference counters.
15105 (Some of the existing 'get' functions increment a reference
15106 counter, some don't.)
15107 Similarly, 'set1' and 'add1' functions increase reference
15108 counters or duplicate objects.
15112 * Allow for the possibility of temp RSA key generation failure:
15113 the code used to assume it always worked and crashed on failure.
15117 * Fix potential buffer overrun problem in BIO_printf().
15118 *Ulf Möller, using public domain code by Patrick Powell; problem
15119 pointed out by David Sacerdote <das33@cornell.edu>*
15121 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15122 RAND_egd() and RAND_status(). In the command line application,
15123 the EGD socket can be specified like a seed file using RANDFILE
15128 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15129 Some CAs (e.g. Verisign) distribute certificates in this form.
15133 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15134 list to exclude them. This means that no special compilation option
15135 is needed to use anonymous DH: it just needs to be included in the
15140 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15141 EVP_MD_type. The old functionality is available in a new macro called
15142 EVP_MD_md(). Change code that uses it and update docs.
15146 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15147 where the `void *` argument is replaced by a function pointer argument.
15148 Previously `void *` was abused to point to functions, which works on
15149 many platforms, but is not correct. As these functions are usually
15150 called by macros defined in OpenSSL header files, most source code
15151 should work without changes.
15155 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15156 sections with information on -D... compiler switches used for
15157 compiling the library so that applications can see them. To enable
15158 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15159 must be defined. E.g.,
15160 #define OPENSSL_ALGORITHM_DEFINES
15161 #include <openssl/opensslconf.h>
15162 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15164 *Richard Levitte, Ulf and Bodo Möller*
15166 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15171 * Change the 'other' type in certificate aux info to a STACK_OF
15172 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15173 the required ASN1 format: arbitrary types determined by an OID.
15177 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15178 argument to 'req'. This is not because the function is newer or
15179 better than others it just uses the work 'NEW' in the certificate
15180 request header lines. Some software needs this.
15184 * Reorganise password command line arguments: now passwords can be
15185 obtained from various sources. Delete the PEM_cb function and make
15186 it the default behaviour: i.e. if the callback is NULL and the
15187 usrdata argument is not NULL interpret it as a null terminated pass
15188 phrase. If usrdata and the callback are NULL then the pass phrase
15189 is prompted for as usual.
15193 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15194 the support is automatically enabled. The resulting binaries will
15195 autodetect the card and use it if present.
15197 *Ben Laurie and Compaq Inc.*
15199 * Work around for Netscape hang bug. This sends certificate request
15200 and server done in one record. Since this is perfectly legal in the
15201 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15202 the bugs/SSLv3 entry for more info.
15206 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15210 * Add -rand argument to smime and pkcs12 applications and read/write
15215 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15219 * Add command line password options to the remaining applications.
15223 * Bug fix for BN_div_recp() for numerators with an even number of
15228 * More tests in bntest.c, and changed test_bn output.
15232 * ./config recognizes MacOS X now.
15236 * Bug fix for BN_div() when the first words of num and divisor are
15237 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15241 * Add support for various broken PKCS#8 formats, and command line
15242 options to produce them.
15246 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15247 get temporary BIGNUMs from a BN_CTX.
15251 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15256 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15257 include a #define from the old name to the new. The original intent
15258 was that statically linked binaries could for example just call
15259 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15260 link with digests. This never worked because SSLeay_add_all_digests()
15261 and SSLeay_add_all_ciphers() were in the same source file so calling
15262 one would link with the other. They are now in separate source files.
15266 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15270 * Use a less unusual form of the Miller-Rabin primality test (it used
15271 a binary algorithm for exponentiation integrated into the Miller-Rabin
15272 loop, our standard modexp algorithms are faster).
15276 * Support for the EBCDIC character set completed.
15278 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15280 * Source code cleanups: use const where appropriate, eliminate casts,
15281 use `void *` instead of `char *` in lhash.
15285 * Bugfix: ssl3_send_server_key_exchange was not restartable
15286 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15287 this the server could overwrite ephemeral keys that the client
15292 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15293 using 50 iterations of the Rabin-Miller test.
15295 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15296 iterations of the Rabin-Miller test as required by the appendix
15297 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15298 As BN_is_prime_fasttest includes trial division, DSA parameter
15299 generation becomes much faster.
15301 This implies a change for the callback functions in DSA_is_prime
15302 and DSA_generate_parameters: The callback function is called once
15303 for each positive witness in the Rabin-Miller test, not just
15304 occasionally in the inner loop; and the parameters to the
15305 callback function now provide an iteration count for the outer
15306 loop rather than for the current invocation of the inner loop.
15307 DSA_generate_parameters additionally can call the callback
15308 function with an 'iteration count' of -1, meaning that a
15309 candidate has passed the trial division test (when q is generated
15310 from an application-provided seed, trial division is skipped).
15314 * New function BN_is_prime_fasttest that optionally does trial
15315 division before starting the Rabin-Miller test and has
15316 an additional BN_CTX * argument (whereas BN_is_prime always
15317 has to allocate at least one BN_CTX).
15318 'callback(1, -1, cb_arg)' is called when a number has passed the
15319 trial division stage.
15323 * Fix for bug in CRL encoding. The validity dates weren't being handled
15328 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15332 * New function BN_pseudo_rand().
15336 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15337 bignum version of BN_from_montgomery() with the working code from
15338 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15343 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15344 made it impossible to use the same SSL_SESSION data structure in
15345 SSL2 clients in multiple threads.
15349 * The return value of RAND_load_file() no longer counts bytes obtained
15350 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15351 to seed the PRNG (previously an explicit byte count was required).
15353 *Ulf Möller, Bodo Möller*
15355 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15356 used `char *` instead of `void *` and had casts all over the place.
15360 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15364 * Retain source code compatibility for BN_prime_checks macro:
15365 BN_is_prime(..., BN_prime_checks, ...) now uses
15366 BN_prime_checks_for_size to determine the appropriate number of
15367 Rabin-Miller iterations.
15371 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15372 DH_CHECK_P_NOT_SAFE_PRIME.
15373 (Check if this is true? OpenPGP calls them "strong".)
15377 * Merge the functionality of "dh" and "gendh" programs into a new program
15378 "dhparam". The old programs are retained for now but will handle DH keys
15379 (instead of parameters) in future.
15383 * Make the ciphers, s_server and s_client programs check the return values
15384 when a new cipher list is set.
15388 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15389 ciphers. Before when the 56bit ciphers were enabled the sorting was
15392 The syntax for the cipher sorting has been extended to support sorting by
15393 cipher-strength (using the strength_bits hard coded in the tables).
15394 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15396 Fix a bug in the cipher-command parser: when supplying a cipher command
15397 string with an "undefined" symbol (neither command nor alphanumeric
15398 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15399 an error is flagged.
15401 Due to the strength-sorting extension, the code of the
15402 ssl_create_cipher_list() function was completely rearranged. I hope that
15403 the readability was also increased :-)
15405 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15407 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15408 for the first serial number and places 2 in the serial number file. This
15409 avoids problems when the root CA is created with serial number zero and
15410 the first user certificate has the same issuer name and serial number
15415 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15416 the new code. Add documentation for this stuff.
15420 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15421 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15422 structures and behave in an analogous way to the X509v3 functions:
15423 they shouldn't be called directly but wrapper functions should be used
15426 So we also now have some wrapper functions that call the X509at functions
15427 when passed certificate requests. (TO DO: similar things can be done with
15428 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15429 things. Some of these need some d2i or i2d and print functionality
15430 because they handle more complex structures.)
15434 * Add missing #ifndefs that caused missing symbols when building libssl
15435 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15436 NO_RSA in `ssl/s2*.c`.
15438 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15440 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15441 has a return value which indicates the quality of the random data
15442 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15443 error queue. New function RAND_pseudo_bytes() generates output that is
15444 guaranteed to be unique but not unpredictable. RAND_add is like
15445 RAND_seed, but takes an extra argument for an entropy estimate
15446 (RAND_seed always assumes full entropy).
15450 * Do more iterations of Rabin-Miller probable prime test (specifically,
15451 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15452 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15453 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15454 false-positive rate of at most 2^-80 for random input.
15458 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15462 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15463 in the 0.9.5 release), this returns the chain
15464 from an X509_CTX structure with a dup of the stack and all
15465 the X509 reference counts upped: so the stack will exist
15466 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15469 Also make SSL_SESSION_print() print out the verify return
15474 * Add manpage for the pkcs12 command. Also change the default
15475 behaviour so MAC iteration counts are used unless the new
15476 -nomaciter option is used. This improves file security and
15477 only older versions of MSIE (4.0 for example) need it.
15481 * Honor the no-xxx Configure options when creating .DEF files.
15485 * Add PKCS#10 attributes to field table: challengePassword,
15486 unstructuredName and unstructuredAddress. These are taken from
15487 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15488 international characters are used.
15490 More changes to X509_ATTRIBUTE code: allow the setting of types
15491 based on strings. Remove the 'loc' parameter when adding
15492 attributes because these will be a SET OF encoding which is sorted
15497 * Initial changes to the 'req' utility to allow request generation
15498 automation. This will allow an application to just generate a template
15499 file containing all the field values and have req construct the
15502 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15503 used all over the place including certificate requests and PKCS#7
15504 structures. They are currently handled manually where necessary with
15505 some primitive wrappers for PKCS#7. The new functions behave in a
15506 manner analogous to the X509 extension functions: they allow
15507 attributes to be looked up by NID and added.
15509 Later something similar to the X509V3 code would be desirable to
15510 automatically handle the encoding, decoding and printing of the
15511 more complex types. The string types like challengePassword can
15512 be handled by the string table functions.
15514 Also modified the multi byte string table handling. Now there is
15515 a 'global mask' which masks out certain types. The table itself
15516 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15517 is useful when for example there is only one permissible type
15518 (as in countryName) and using the mask might result in no valid
15523 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15524 SSL_get_peer_finished to allow applications to obtain the latest
15525 Finished messages sent to the peer or expected from the peer,
15526 respectively. (SSL_get_peer_finished is usually the Finished message
15527 actually received from the peer, otherwise the protocol will be aborted.)
15529 As the Finished message are message digests of the complete handshake
15530 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15531 be used for external authentication procedures when the authentication
15532 provided by SSL/TLS is not desired or is not enough.
15536 * Enhanced support for Alpha Linux is added. Now ./config checks if
15537 the host supports BWX extension and if Compaq C is present on the
15538 $PATH. Just exploiting of the BWX extension results in 20-30%
15539 performance kick for some algorithms, e.g. DES and RC4 to mention
15540 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15545 * Add support for MS "fast SGC". This is arguably a violation of the
15546 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15547 weak crypto and after checking the certificate is SGC a second one
15548 with strong crypto. MS SGC stops the first handshake after receiving
15549 the server certificate message and sends a second client hello. Since
15550 a server will typically do all the time consuming operations before
15551 expecting any further messages from the client (server key exchange
15552 is the most expensive) there is little difference between the two.
15554 To get OpenSSL to support MS SGC we have to permit a second client
15555 hello message after we have sent server done. In addition we have to
15556 reset the MAC if we do get this second client hello.
15560 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15561 if a DER encoded private key is RSA or DSA traditional format. Changed
15562 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15563 format DER encoded private key. Newer code should use PKCS#8 format which
15564 has the key type encoded in the ASN1 structure. Added DER private key
15565 support to pkcs8 application.
15569 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15570 ciphersuites has been selected (as required by the SSL 3/TLS 1
15571 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15572 is set, we interpret this as a request to violate the specification
15573 (the worst that can happen is a handshake failure, and 'correct'
15574 behaviour would result in a handshake failure anyway).
15578 * In SSL_CTX_add_session, take into account that there might be multiple
15579 SSL_SESSION structures with the same session ID (e.g. when two threads
15580 concurrently obtain them from an external cache).
15581 The internal cache can handle only one SSL_SESSION with a given ID,
15582 so if there's a conflict, we now throw out the old one to achieve
15587 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15588 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15589 some routines that use cipher OIDs: some ciphers do not have OIDs
15590 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15595 * Simplify the trust setting structure and code. Now we just have
15596 two sequences of OIDs for trusted and rejected settings. These will
15597 typically have values the same as the extended key usage extension
15598 and any application specific purposes.
15600 The trust checking code now has a default behaviour: it will just
15601 check for an object with the same NID as the passed id. Functions can
15602 be provided to override either the default behaviour or the behaviour
15603 for a given id. SSL client, server and email already have functions
15604 in place for compatibility: they check the NID and also return "trusted"
15605 if the certificate is self signed.
15609 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15610 traditional format into an EVP_PKEY structure.
15614 * Add a password callback function PEM_cb() which either prompts for
15615 a password if usr_data is NULL or otherwise assumes it is a null
15616 terminated password. Allow passwords to be passed on command line
15617 environment or config files in a few more utilities.
15621 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15622 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15623 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15624 Update documentation.
15628 * Support for ASN1 "NULL" type. This could be handled before by using
15629 ASN1_TYPE but there wasn't any function that would try to read a NULL
15630 and produce an error if it couldn't. For compatibility we also have
15631 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15632 don't allocate anything because they don't need to.
15636 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15639 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15641 * Rebuild of the memory allocation routines used by OpenSSL code and
15642 possibly others as well. The purpose is to make an interface that
15643 provide hooks so anyone can build a separate set of allocation and
15644 deallocation routines to be used by OpenSSL, for example memory
15645 pool implementations, or something else, which was previously hard
15646 since Malloc(), Realloc() and Free() were defined as macros having
15647 the values malloc, realloc and free, respectively (except for Win32
15648 compilations). The same is provided for memory debugging code.
15649 OpenSSL already comes with functionality to find memory leaks, but
15650 this gives people a chance to debug other memory problems.
15652 With these changes, a new set of functions and macros have appeared:
15654 CRYPTO_set_mem_debug_functions() [F]
15655 CRYPTO_get_mem_debug_functions() [F]
15656 CRYPTO_dbg_set_options() [F]
15657 CRYPTO_dbg_get_options() [F]
15658 CRYPTO_malloc_debug_init() [M]
15660 The memory debug functions are NULL by default, unless the library
15661 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15662 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15663 gives the standard debugging functions that come with OpenSSL) or
15664 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15665 provided by the library user) must be used. When the standard
15666 debugging functions are used, CRYPTO_dbg_set_options can be used to
15667 request additional information:
15668 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15669 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15671 Also, things like CRYPTO_set_mem_functions will always give the
15672 expected result (the new set of functions is used for allocation
15673 and deallocation) at all times, regardless of platform and compiler
15676 To finish it up, some functions that were never use in any other
15677 way than through macros have a new API and new semantic:
15679 CRYPTO_dbg_malloc()
15680 CRYPTO_dbg_realloc()
15683 All macros of value have retained their old syntax.
15685 *Richard Levitte and Bodo Moeller*
15687 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15688 ordering of SMIMECapabilities wasn't in "strength order" and there
15689 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15694 * Some ASN1 types with illegal zero length encoding (INTEGER,
15695 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15697 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15699 * Merge in my S/MIME library for OpenSSL. This provides a simple
15700 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15701 functionality to handle multipart/signed properly) and a utility
15702 called 'smime' to call all this stuff. This is based on code I
15703 originally wrote for Celo who have kindly allowed it to be
15704 included in OpenSSL.
15708 * Add variants des_set_key_checked and des_set_key_unchecked of
15709 des_set_key (aka des_key_sched). Global variable des_check_key
15710 decides which of these is called by des_set_key; this way
15711 des_check_key behaves as it always did, but applications and
15712 the library itself, which was buggy for des_check_key == 1,
15713 have a cleaner way to pick the version they need.
15717 * New function PKCS12_newpass() which changes the password of a
15722 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15723 dynamic mix. In both cases the ids can be used as an index into the
15724 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15725 functions so they accept a list of the field values and the
15726 application doesn't need to directly manipulate the X509_TRUST
15731 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15736 * Modify the way the V3 extension code looks up extensions. This now
15737 works in a similar way to the object code: we have some "standard"
15738 extensions in a static table which is searched with OBJ_bsearch()
15739 and the application can add dynamic ones if needed. The file
15740 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15741 updated whenever a new extension is added to the core code and kept
15742 in ext_nid order. There is a simple program 'tabtest.c' which checks
15743 this. New extensions are not added too often so this file can readily
15744 be maintained manually.
15746 There are two big advantages in doing things this way. The extensions
15747 can be looked up immediately and no longer need to be "added" using
15748 X509V3_add_standard_extensions(): this function now does nothing.
15749 Side note: I get *lots* of email saying the extension code doesn't
15750 work because people forget to call this function.
15751 Also no dynamic allocation is done unless new extensions are added:
15752 so if we don't add custom extensions there is no need to call
15753 X509V3_EXT_cleanup().
15757 * Modify enc utility's salting as follows: make salting the default. Add a
15758 magic header, so unsalted files fail gracefully instead of just decrypting
15759 to garbage. This is because not salting is a big security hole, so people
15760 should be discouraged from doing it.
15764 * Fixes and enhancements to the 'x509' utility. It allowed a message
15765 digest to be passed on the command line but it only used this
15766 parameter when signing a certificate. Modified so all relevant
15767 operations are affected by the digest parameter including the
15768 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15769 DSA key was used because it didn't fix the digest.
15773 * Initial certificate chain verify code. Currently tests the untrusted
15774 certificates for consistency with the verify purpose (which is set
15775 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15777 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15778 this is because it will reject chains with invalid extensions whereas
15779 every previous version of OpenSSL and SSLeay made no checks at all.
15781 Trust code: checks the root CA for the relevant trust settings. Trust
15782 settings have an initial value consistent with the verify purpose: e.g.
15783 if the verify purpose is for SSL client use it expects the CA to be
15784 trusted for SSL client use. However the default value can be changed to
15785 permit custom trust settings: one example of this would be to only trust
15786 certificates from a specific "secure" set of CAs.
15788 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15789 which should be used for version portability: especially since the
15790 verify structure is likely to change more often now.
15792 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15793 to set them. If not set then assume SSL clients will verify SSL servers
15796 Two new options to the verify program: -untrusted allows a set of
15797 untrusted certificates to be passed in and -purpose which sets the
15798 intended purpose of the certificate. If a purpose is set then the
15799 new chain verify code is used to check extension consistency.
15803 * Support for the authority information access extension.
15807 * Modify RSA and DSA PEM read routines to transparently handle
15808 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15809 public keys in a format compatible with certificate
15810 SubjectPublicKeyInfo structures. Unfortunately there were already
15811 functions called *_PublicKey_* which used various odd formats so
15812 these are retained for compatibility: however the DSA variants were
15813 never in a public release so they have been deleted. Changed dsa/rsa
15814 utilities to handle the new format: note no releases ever handled public
15815 keys so we should be OK.
15817 The primary motivation for this change is to avoid the same fiasco
15818 that dogs private keys: there are several incompatible private key
15819 formats some of which are standard and some OpenSSL specific and
15820 require various evil hacks to allow partial transparent handling and
15821 even then it doesn't work with DER formats. Given the option anything
15822 other than PKCS#8 should be dumped: but the other formats have to
15823 stay in the name of compatibility.
15825 With public keys and the benefit of hindsight one standard format
15826 is used which works with EVP_PKEY, RSA or DSA structures: though
15827 it clearly returns an error if you try to read the wrong kind of key.
15829 Added a -pubkey option to the 'x509' utility to output the public key.
15830 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15831 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15832 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15833 that do the same as the `EVP_PKEY_assign_*()` except they up the
15834 reference count of the added key (they don't "swallow" the
15839 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15840 CRLs would fail if the file contained no certificates or no CRLs:
15841 added a new function to read in both types and return the number
15842 read: this means that if none are read it will be an error. The
15843 DER versions of the certificate and CRL reader would always fail
15844 because it isn't possible to mix certificates and CRLs in DER format
15845 without choking one or the other routine. Changed this to just read
15846 a certificate: this is the best we can do. Also modified the code
15847 in apps/verify.c to take notice of return codes: it was previously
15848 attempting to read in certificates from NULL pointers and ignoring
15849 any errors: this is one reason why the cert and CRL reader seemed
15850 to work. It doesn't check return codes from the default certificate
15851 routines: these may well fail if the certificates aren't installed.
15855 * Code to support otherName option in GeneralName.
15859 * First update to verify code. Change the verify utility
15860 so it warns if it is passed a self signed certificate:
15861 for consistency with the normal behaviour. X509_verify
15862 has been modified to it will now verify a self signed
15863 certificate if *exactly* the same certificate appears
15864 in the store: it was previously impossible to trust a
15865 single self signed certificate. This means that:
15866 openssl verify ss.pem
15867 now gives a warning about a self signed certificate but
15868 openssl verify -CAfile ss.pem ss.pem
15873 * For servers, store verify_result in SSL_SESSION data structure
15874 (and add it to external session representation).
15875 This is needed when client certificate verifications fails,
15876 but an application-provided verification callback (set by
15877 SSL_CTX_set_cert_verify_callback) allows accepting the session
15878 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15879 but returns 1): When the session is reused, we have to set
15880 ssl->verify_result to the appropriate error code to avoid
15883 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15885 * Fix a bug in the new PKCS#7 code: it didn't consider the
15886 case in PKCS7_dataInit() where the signed PKCS7 structure
15887 didn't contain any existing data because it was being created.
15889 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15891 * Add a salt to the key derivation routines in enc.c. This
15892 forms the first 8 bytes of the encrypted file. Also add a
15893 -S option to allow a salt to be input on the command line.
15897 * New function X509_cmp(). Oddly enough there wasn't a function
15898 to compare two certificates. We do this by working out the SHA1
15899 hash and comparing that. X509_cmp() will be needed by the trust
15904 * SSL_get1_session() is like SSL_get_session(), but increments
15905 the reference count in the SSL_SESSION returned.
15907 *Geoff Thorpe <geoff@eu.c2.net>*
15909 * Fix for 'req': it was adding a null to request attributes.
15910 Also change the X509_LOOKUP and X509_INFO code to handle
15911 certificate auxiliary information.
15915 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15920 * Add the possibility to add extra information to the memory leak
15921 detecting output, to form tracebacks, showing from where each
15922 allocation was originated: CRYPTO_push_info("constant string") adds
15923 the string plus current file name and line number to a per-thread
15924 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15925 is like calling CYRPTO_pop_info() until the stack is empty.
15926 Also updated memory leak detection code to be multi-thread-safe.
15930 * Add options -text and -noout to pkcs7 utility and delete the
15931 encryption options which never did anything. Update docs.
15935 * Add options to some of the utilities to allow the pass phrase
15936 to be included on either the command line (not recommended on
15937 OSes like Unix) or read from the environment. Update the
15938 manpages and fix a few bugs.
15942 * Add a few manpages for some of the openssl commands.
15946 * Fix the -revoke option in ca. It was freeing up memory twice,
15947 leaking and not finding already revoked certificates.
15951 * Extensive changes to support certificate auxiliary information.
15952 This involves the use of X509_CERT_AUX structure and X509_AUX
15953 functions. An X509_AUX function such as PEM_read_X509_AUX()
15954 can still read in a certificate file in the usual way but it
15955 will also read in any additional "auxiliary information". By
15956 doing things this way a fair degree of compatibility can be
15957 retained: existing certificates can have this information added
15958 using the new 'x509' options.
15960 Current auxiliary information includes an "alias" and some trust
15961 settings. The trust settings will ultimately be used in enhanced
15962 certificate chain verification routines: currently a certificate
15963 can only be trusted if it is self signed and then it is trusted
15968 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
15969 The problem was that one of the replacement routines had not been working
15970 since SSLeay releases. For now the offending routine has been replaced
15971 with non-optimised assembler. Even so, this now gives around 95%
15972 performance improvement for 1024 bit RSA signs.
15976 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15977 handling. Most clients have the effective key size in bits equal to
15978 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15979 A few however don't do this and instead use the size of the decrypted key
15980 to determine the RC2 key length and the AlgorithmIdentifier to determine
15981 the effective key length. In this case the effective key length can still
15982 be 40 bits but the key length can be 168 bits for example. This is fixed
15983 by manually forcing an RC2 key into the EVP_PKEY structure because the
15984 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15985 the key length and effective key length are equal.
15989 * Add a bunch of functions that should simplify the creation of
15990 X509_NAME structures. Now you should be able to do:
15991 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15992 and have it automatically work out the correct field type and fill in
15993 the structures. The more adventurous can try:
15994 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15995 and it will (hopefully) work out the correct multibyte encoding.
15999 * Change the 'req' utility to use the new field handling and multibyte
16000 copy routines. Before the DN field creation was handled in an ad hoc
16001 way in req, ca, and x509 which was rather broken and didn't support
16002 BMPStrings or UTF8Strings. Since some software doesn't implement
16003 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16004 using the dirstring_type option. See the new comment in the default
16005 openssl.cnf for more info.
16009 * Make crypto/rand/md_rand.c more robust:
16010 - Assure unique random numbers after fork().
16011 - Make sure that concurrent threads access the global counter and
16012 md serializably so that we never lose entropy in them
16013 or use exactly the same state in multiple threads.
16014 Access to the large state is not always serializable because
16015 the additional locking could be a performance killer, and
16016 md should be large enough anyway.
16020 * New file apps/app_rand.c with commonly needed functionality
16021 for handling the random seed file.
16023 Use the random seed file in some applications that previously did not:
16025 dsaparam -genkey (which also ignored its '-rand' option),
16028 x509 (when signing).
16029 Except on systems with /dev/urandom, it is crucial to have a random
16030 seed file at least for key creation, DSA signing, and for DH exchanges;
16031 for RSA signatures we could do without one.
16033 gendh and gendsa (unlike genrsa) used to read only the first byte
16034 of each file listed in the '-rand' option. The function as previously
16035 found in genrsa is now in app_rand.c and is used by all programs
16036 that support '-rand'.
16040 * In RAND_write_file, use mode 0600 for creating files;
16041 don't just chmod when it may be too late.
16045 * Report an error from X509_STORE_load_locations
16046 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16050 * New function ASN1_mbstring_copy() this copies a string in either
16051 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16052 into an ASN1_STRING type. A mask of permissible types is passed
16053 and it chooses the "minimal" type to use or an error if not type
16058 * Add function equivalents to the various macros in asn1.h. The old
16059 macros are retained with an `M_` prefix. Code inside the library can
16060 use the `M_` macros. External code (including the openssl utility)
16061 should *NOT* in order to be "shared library friendly".
16065 * Add various functions that can check a certificate's extensions
16066 to see if it usable for various purposes such as SSL client,
16067 server or S/MIME and CAs of these types. This is currently
16068 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16069 verification. Also added a -purpose flag to x509 utility to
16070 print out all the purposes.
16074 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16079 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16080 for, obtain and decode and extension and obtain its critical flag.
16081 This allows all the necessary extension code to be handled in a
16082 single function call.
16086 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16087 platforms. See crypto/rc4/rc4_enc.c for further details.
16091 * New -noout option to asn1parse. This causes no output to be produced
16092 its main use is when combined with -strparse and -out to extract data
16093 from a file (which may not be in ASN.1 format).
16097 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16098 when producing the local key id.
16100 *Richard Levitte <levitte@stacken.kth.se>*
16102 * New option -dhparam in s_server. This allows a DH parameter file to be
16103 stated explicitly. If it is not stated then it tries the first server
16104 certificate file. The previous behaviour hard coded the filename
16109 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16110 a public key to be input or output. For example:
16111 openssl rsa -in key.pem -pubout -out pubkey.pem
16112 Also added necessary DSA public key functions to handle this.
16116 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16117 in the message. This was handled by allowing
16118 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16120 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16122 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16123 to the end of the strings whereas this didn't. This would cause problems
16124 if strings read with d2i_ASN1_bytes() were later modified.
16126 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16128 * Fix for base64 decode bug. When a base64 bio reads only one line of
16129 data and it contains EOF it will end up returning an error. This is
16130 caused by input 46 bytes long. The cause is due to the way base64
16131 BIOs find the start of base64 encoded data. They do this by trying a
16132 trial decode on each line until they find one that works. When they
16133 do a flag is set and it starts again knowing it can pass all the
16134 data directly through the decoder. Unfortunately it doesn't reset
16135 the context it uses. This means that if EOF is reached an attempt
16136 is made to pass two EOFs through the context and this causes the
16137 resulting error. This can also cause other problems as well. As is
16138 usual with these problems it takes *ages* to find and the fix is
16139 trivial: move one line.
16141 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16143 * Ugly workaround to get s_client and s_server working under Windows. The
16144 old code wouldn't work because it needed to select() on sockets and the
16145 tty (for keypresses and to see if data could be written). Win32 only
16146 supports select() on sockets so we select() with a 1s timeout on the
16147 sockets and then see if any characters are waiting to be read, if none
16148 are present then we retry, we also assume we can always write data to
16149 the tty. This isn't nice because the code then blocks until we've
16150 received a complete line of data and it is effectively polling the
16151 keyboard at 1s intervals: however it's quite a bit better than not
16152 working at all :-) A dedicated Windows application might handle this
16153 with an event loop for example.
16157 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16158 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16159 will be called when RSA_sign() and RSA_verify() are used. This is useful
16160 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16161 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16162 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16163 This necessitated the support of an extra signature type NID_md5_sha1
16164 for SSL signatures and modifications to the SSL library to use it instead
16165 of calling RSA_public_decrypt() and RSA_private_encrypt().
16169 * Add new -verify -CAfile and -CApath options to the crl program, these
16170 will lookup a CRL issuers certificate and verify the signature in a
16171 similar way to the verify program. Tidy up the crl program so it
16172 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16173 less strict. It will now permit CRL extensions even if it is not
16174 a V2 CRL: this will allow it to tolerate some broken CRLs.
16178 * Initialize all non-automatic variables each time one of the openssl
16179 sub-programs is started (this is necessary as they may be started
16180 multiple times from the "OpenSSL>" prompt).
16182 *Lennart Bang, Bodo Moeller*
16184 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16185 removing all other RSA functionality (this is what NO_RSA does). This
16186 is so (for example) those in the US can disable those operations covered
16187 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16192 * Non-copying interface to BIO pairs.
16193 (still largely untested)
16197 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16198 ASCII string. This was handled independently in various places before.
16202 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16203 UTF8 strings a character at a time.
16207 * Use client_version from client hello to select the protocol
16208 (s23_srvr.c) and for RSA client key exchange verification
16209 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16213 * Add various utility functions to handle SPKACs, these were previously
16214 handled by poking round in the structure internals. Added new function
16215 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16216 print, verify and generate SPKACs. Based on an original idea from
16217 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16221 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16225 * Allow the config file extension section to be overwritten on the
16226 command line. Based on an original idea from Massimiliano Pala
16227 <madwolf@comune.modena.it>. The new option is called -extensions
16228 and can be applied to ca, req and x509. Also -reqexts to override
16229 the request extensions in req and -crlexts to override the crl extensions
16234 * Add new feature to the SPKAC handling in ca. Now you can include
16235 the same field multiple times by preceding it by "XXXX." for example:
16238 this is the same syntax as used in the req config file.
16242 * Allow certificate extensions to be added to certificate requests. These
16243 are specified in a 'req_extensions' option of the req section of the
16244 config file. They can be printed out with the -text option to req but
16245 are otherwise ignored at present.
16249 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16250 data read consists of only the final block it would not decrypted because
16251 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16252 A misplaced 'break' also meant the decrypted final block might not be
16253 copied until the next read.
16257 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16258 a few extra parameters to the DH structure: these will be useful if
16259 for example we want the value of 'q' or implement X9.42 DH.
16263 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16264 provides hooks that allow the default DSA functions or functions on a
16265 "per key" basis to be replaced. This allows hardware acceleration and
16266 hardware key storage to be handled without major modification to the
16267 library. Also added low level modexp hooks and CRYPTO_EX structure and
16268 associated functions.
16272 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16273 as "read only": it can't be written to and the buffer it points to will
16274 not be freed. Reading from a read only BIO is much more efficient than
16275 a normal memory BIO. This was added because there are several times when
16276 an area of memory needs to be read from a BIO. The previous method was
16277 to create a memory BIO and write the data to it, this results in two
16278 copies of the data and an O(n^2) reading algorithm. There is a new
16279 function BIO_new_mem_buf() which creates a read only memory BIO from
16280 an area of memory. Also modified the PKCS#7 routines to use read only
16285 * Bugfix: ssl23_get_client_hello did not work properly when called in
16286 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16287 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16288 but a retry condition occurred while trying to read the rest.
16292 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16293 NID_pkcs7_encrypted by default: this was wrong since this should almost
16294 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16295 the encrypted data type: this is a more sensible place to put it and it
16296 allows the PKCS#12 code to be tidied up that duplicated this
16301 * Changed obj_dat.pl script so it takes its input and output files on
16302 the command line. This should avoid shell escape redirection problems
16307 * Initial support for certificate extension requests, these are included
16308 in things like Xenroll certificate requests. Included functions to allow
16309 extensions to be obtained and added.
16313 * -crlf option to s_client and s_server for sending newlines as
16314 CRLF (as required by many protocols).
16318 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16320 * Install libRSAglue.a when OpenSSL is built with RSAref.
16322 *Ralf S. Engelschall*
16324 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16326 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16328 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16333 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16334 DH parameters/keys (q is lost during that conversion, but the resulting
16335 DH parameters contain its length).
16337 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16338 much faster than DH_generate_parameters (which creates parameters
16339 where `p = 2*q + 1`), and also the smaller q makes DH computations
16340 much more efficient (160-bit exponentiation instead of 1024-bit
16341 exponentiation); so this provides a convenient way to support DHE
16342 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16343 utter importance to use
16344 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16346 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16347 when such DH parameters are used, because otherwise small subgroup
16348 attacks may become possible!
16352 * Avoid memory leak in i2d_DHparams.
16356 * Allow the -k option to be used more than once in the enc program:
16357 this allows the same encrypted message to be read by multiple recipients.
16361 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16362 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16363 it will always use the numerical form of the OID, even if it has a short
16368 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16369 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16370 otherwise bn_mod_exp was called. In the case of hardware keys for example
16371 no private key components need be present and it might store extra data
16372 in the RSA structure, which cannot be accessed from bn_mod_exp.
16373 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16374 private key operations.
16378 * Added support for SPARC Linux.
16382 * pem_password_cb function type incompatibly changed from
16383 typedef int pem_password_cb(char *buf, int size, int rwflag);
16385 ....(char *buf, int size, int rwflag, void *userdata);
16386 so that applications can pass data to their callbacks:
16387 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16388 additional void * argument, which is just handed through whenever
16389 the password callback is called.
16391 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16393 New function SSL_CTX_set_default_passwd_cb_userdata.
16395 Compatibility note: As many C implementations push function arguments
16396 onto the stack in reverse order, the new library version is likely to
16397 interoperate with programs that have been compiled with the old
16398 pem_password_cb definition (PEM_whatever takes some data that
16399 happens to be on the stack as its last argument, and the callback
16400 just ignores this garbage); but there is no guarantee whatsoever that
16403 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16404 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16405 problems not only on Windows, but also on some Unix platforms.
16406 To avoid problematic command lines, these definitions are now in an
16407 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16408 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16412 * MIPS III/IV assembler module is reimplemented.
16416 * More DES library cleanups: remove references to srand/rand and
16417 delete an unused file.
16421 * Add support for the free Netwide assembler (NASM) under Win32,
16422 since not many people have MASM (ml) and it can be hard to obtain.
16423 This is currently experimental but it seems to work OK and pass all
16424 the tests. Check out INSTALL.W32 for info.
16428 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16429 without temporary keys kept an extra copy of the server key,
16430 and connections with temporary keys did not free everything in case
16435 * New function RSA_check_key and new openssl rsa option -check
16436 for verifying the consistency of RSA keys.
16438 *Ulf Moeller, Bodo Moeller*
16440 * Various changes to make Win32 compile work:
16441 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16442 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16443 comparison" warnings.
16444 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16448 * Add a debugging option to PKCS#5 v2 key generation function: when
16449 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16450 derived keys are printed to stderr.
16454 * Copy the flags in ASN1_STRING_dup().
16456 *Roman E. Pavlov <pre@mo.msk.ru>*
16458 * The x509 application mishandled signing requests containing DSA
16459 keys when the signing key was also DSA and the parameters didn't match.
16461 It was supposed to omit the parameters when they matched the signing key:
16462 the verifying software was then supposed to automatically use the CA's
16463 parameters if they were absent from the end user certificate.
16465 Omitting parameters is no longer recommended. The test was also
16466 the wrong way round! This was probably due to unusual behaviour in
16467 EVP_cmp_parameters() which returns 1 if the parameters match.
16468 This meant that parameters were omitted when they *didn't* match and
16469 the certificate was useless. Certificates signed with 'ca' didn't have
16472 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16474 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16475 The interface is as follows:
16476 Applications can use
16477 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16478 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16479 "off" is now the default.
16480 The library internally uses
16481 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16482 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16483 to disable memory-checking temporarily.
16485 Some inconsistent states that previously were possible (and were
16486 even the default) are now avoided.
16488 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16489 with each memory chunk allocated; this is occasionally more helpful
16490 than just having a counter.
16492 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16494 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16499 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16500 which largely parallels "options", but is for changing API behaviour,
16501 whereas "options" are about protocol behaviour.
16502 Initial "mode" flags are:
16504 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16505 a single record has been written.
16506 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16507 retries use the same buffer location.
16508 (But all of the contents must be
16513 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16516 * Fix problems with no-hmac etc.
16518 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16520 * New functions RSA_get_default_method(), RSA_set_method() and
16521 RSA_get_method(). These allows replacement of RSA_METHODs without having
16522 to mess around with the internals of an RSA structure.
16526 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16527 Also really enable memory leak checks in openssl.c and in some
16530 *Chad C. Mulligan, Bodo Moeller*
16532 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16533 up the length of negative integers. This has now been simplified to just
16534 store the length when it is first determined and use it later, rather
16535 than trying to keep track of where data is copied and updating it to
16537 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16539 * Add a new function PKCS7_signatureVerify. This allows the verification
16540 of a PKCS#7 signature but with the signing certificate passed to the
16541 function itself. This contrasts with PKCS7_dataVerify which assumes the
16542 certificate is present in the PKCS#7 structure. This isn't always the
16543 case: certificates can be omitted from a PKCS#7 structure and be
16544 distributed by "out of band" means (such as a certificate database).
16548 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16549 function prototypes in pem.h, also change util/mkdef.pl to add the
16550 necessary function names.
16554 * mk1mf.pl (used by Windows builds) did not properly read the
16555 options set by Configure in the top level Makefile, and Configure
16556 was not even able to write more than one option correctly.
16557 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16561 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16562 file to be loaded from a BIO or FILE pointer. The BIO version will
16563 for example allow memory BIOs to contain config info.
16567 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16568 Whoever hopes to achieve shared-library compatibility across versions
16569 must use this, not the compile-time macro.
16570 (Exercise 0.9.4: Which is the minimum library version required by
16572 Note: All this applies only to multi-threaded programs, others don't
16577 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16578 through a BIO pair triggered the default case, i.e.
16579 SSLerr(...,SSL_R_UNKNOWN_STATE).
16583 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16584 can use the SSL library even if none of the specific BIOs is
16589 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16590 for the encoded length.
16592 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16594 * Add initial documentation of the X509V3 functions.
16598 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16599 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16600 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16601 secure PKCS#8 private key format with a high iteration count.
16605 * Fix determination of Perl interpreter: A perl or perl5
16606 *directory* in $PATH was also accepted as the interpreter.
16608 *Ralf S. Engelschall*
16610 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16611 wrong with it but it was very old and did things like calling
16612 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16613 unusual formatting.
16617 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16618 to use the new extension code.
16622 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16623 with macros. This should make it easier to change their form, add extra
16624 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16629 * Add to configuration table a new entry that can specify an alternative
16630 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16631 according to Mark Crispin <MRC@Panda.COM>.
16635 * DES CBC did not update the IV. Weird.
16639 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16640 Changing the behaviour of the former might break existing programs --
16641 where IV updating is needed, des_ncbc_encrypt can be used.
16644 * When bntest is run from "make test" it drives bc to check its
16645 calculations, as well as internally checking them. If an internal check
16646 fails, it needs to cause bc to give a non-zero result or make test carries
16647 on without noticing the failure. Fixed.
16651 * DES library cleanups.
16655 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16656 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16657 ciphers. NOTE: although the key derivation function has been verified
16658 against some published test vectors it has not been extensively tested
16659 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16664 * Instead of "mkdir -p", which is not fully portable, use new
16665 Perl script "util/mkdir-p.pl".
16669 * Rewrite the way password based encryption (PBE) is handled. It used to
16670 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16671 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16672 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16673 the 'parameter' field of the AlgorithmIdentifier is passed to the
16674 underlying key generation function so it must do its own ASN1 parsing.
16675 This has also changed the EVP_PBE_CipherInit() function which now has a
16676 'parameter' argument instead of literal salt and iteration count values
16677 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16681 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16682 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16683 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16684 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16685 value was just used as a "magic string" and not used directly its
16686 value doesn't matter.
16690 * Introduce some semblance of const correctness to BN. Shame C doesn't
16695 * "linux-sparc64" configuration (ultrapenguin).
16697 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16698 "linux-sparc" configuration.
16700 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16702 * config now generates no-xxx options for missing ciphers.
16706 * Support the EBCDIC character set (work in progress).
16707 File ebcdic.c not yet included because it has a different license.
16709 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16711 * Support BS2000/OSD-POSIX.
16713 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16715 * Make callbacks for key generation use `void *` instead of `char *`.
16719 * Make S/MIME samples compile (not yet tested).
16723 * Additional typesafe stacks.
16727 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16731 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16733 * New configuration variant "sco5-gcc".
16735 * Updated some demos.
16737 *Sean O Riordain, Wade Scholine*
16739 * Add missing BIO_free at exit of pkcs12 application.
16743 * Fix memory leak in conf.c.
16747 * Updates for Win32 to assembler version of MD5.
16751 * Set #! path to perl in apps/der_chop to where we found it
16752 instead of using a fixed path.
16756 * SHA library changes for irix64-mips4-cc.
16760 * Improvements for VMS support.
16764 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16766 * Bignum library bug fix. IRIX 6 passes "make test" now!
16767 This also avoids the problems with SC4.2 and unpatched SC5.
16769 *Andy Polyakov <appro@fy.chalmers.se>*
16771 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16772 These are required because of the typesafe stack would otherwise break
16773 existing code. If old code used a structure member which used to be STACK
16774 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16775 sk_num or sk_value it would produce an error because the num, data members
16776 are not present in STACK_OF. Now it just produces a warning. sk_set
16777 replaces the old method of assigning a value to sk_value
16778 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16779 that does this will no longer work (and should use sk_set instead) but
16780 this could be regarded as a "questionable" behaviour anyway.
16784 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16785 correctly handle encrypted S/MIME data.
16789 * Change type of various DES function arguments from des_cblock
16790 (which means, in function argument declarations, pointer to char)
16791 to des_cblock * (meaning pointer to array with 8 char elements),
16792 which allows the compiler to do more typechecking; it was like
16793 that back in SSLeay, but with lots of ugly casts.
16795 Introduce new type const_des_cblock.
16799 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16800 problems: find RecipientInfo structure that matches recipient certificate
16801 and initialise the ASN1 structures properly based on passed cipher.
16805 * Belatedly make the BN tests actually check the results.
16809 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16810 to and from BNs: it was completely broken. New compilation option
16811 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16812 key elements as negative integers.
16816 * Reorganize and speed up MD5.
16818 *Andy Polyakov <appro@fy.chalmers.se>*
16822 *Richard Levitte <richard@levitte.org>*
16824 * New option -out to asn1parse to allow the parsed structure to be
16825 output to a file. This is most useful when combined with the -strparse
16826 option to examine the output of things like OCTET STRINGS.
16830 * Make SSL library a little more fool-proof by not requiring any longer
16831 that `SSL_set_{accept,connect}_state` be called before
16832 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16833 in many applications because usually everything *appeared* to work as
16834 intended anyway -- now it really works as intended).
16838 * Move openssl.cnf out of lib/.
16842 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16843 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16844 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16846 *Ralf S. Engelschall*
16848 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16849 handle PKCS#7 enveloped data properly.
16851 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16853 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16854 copying pointers. The cert_st handling is changed by this in
16855 various ways (and thus what used to be known as ctx->default_cert
16856 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16857 any longer when s->cert does not give us what we need).
16858 ssl_cert_instantiate becomes obsolete by this change.
16859 As soon as we've got the new code right (possibly it already is?),
16860 we have solved a couple of bugs of the earlier code where s->cert
16861 was used as if it could not have been shared with other SSL structures.
16863 Note that using the SSL API in certain dirty ways now will result
16864 in different behaviour than observed with earlier library versions:
16865 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16866 does not influence s as it used to.
16868 In order to clean up things more thoroughly, inside SSL_SESSION
16869 we don't use CERT any longer, but a new structure SESS_CERT
16870 that holds per-session data (if available); currently, this is
16871 the peer's certificate chain and, for clients, the server's certificate
16872 and temporary key. CERT holds only those values that can have
16873 meaningful defaults in an SSL_CTX.
16877 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16878 from the internal representation. Various PKCS#7 fixes: remove some
16879 evil casts and set the enc_dig_alg field properly based on the signing
16884 * Allow PKCS#12 password to be set from the command line or the
16885 environment. Let 'ca' get its config file name from the environment
16886 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16891 * Allow certificate policies extension to use an IA5STRING for the
16892 organization field. This is contrary to the PKIX definition but
16893 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16898 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16899 without disallowing inline assembler and the like for non-pedantic builds.
16903 * Support Borland C++ builder.
16905 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16911 * SHA-1 cleanups and performance enhancements.
16913 *Andy Polyakov <appro@fy.chalmers.se>*
16915 * Sparc v8plus assembler for the bignum library.
16917 *Andy Polyakov <appro@fy.chalmers.se>*
16919 * Accept any -xxx and +xxx compiler options in Configure.
16923 * Update HPUX configuration.
16927 * Add missing `sk_<type>_unshift()` function to safestack.h
16929 *Ralf S. Engelschall*
16931 * New function SSL_CTX_use_certificate_chain_file that sets the
16932 "extra_cert"s in addition to the certificate. (This makes sense
16933 only for "PEM" format files, as chains as a whole are not
16938 * Support verify_depth from the SSL API.
16939 x509_vfy.c had what can be considered an off-by-one-error:
16940 Its depth (which was not part of the external interface)
16941 was actually counting the number of certificates in a chain;
16942 now it really counts the depth.
16946 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16947 instead of X509err, which often resulted in confusing error
16948 messages since the error codes are not globally unique
16949 (e.g. an alleged error in ssl3_accept when a certificate
16950 didn't match the private key).
16952 * New function SSL_CTX_set_session_id_context that allows to set a default
16953 value (so that you don't need SSL_set_session_id_context for each
16954 connection using the SSL_CTX).
16958 * OAEP decoding bug fix.
16962 * Support INSTALL_PREFIX for package builders, as proposed by
16967 * New Configure options "threads" and "no-threads". For systems
16968 where the proper compiler options are known (currently Solaris
16969 and Linux), "threads" is the default.
16973 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16977 * Install various scripts to $(OPENSSLDIR)/misc, not to
16978 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16979 such as /usr/local/bin.
16983 * "make linux-shared" to build shared libraries.
16985 *Niels Poppe <niels@netbox.org>*
16987 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
16991 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16992 extension adding in x509 utility.
16996 * Remove NOPROTO sections and error code comments.
17000 * Partial rewrite of the DEF file generator to now parse the ANSI
17005 * New Configure options --prefix=DIR and --openssldir=DIR.
17009 * Complete rewrite of the error code script(s). It is all now handled
17010 by one script at the top level which handles error code gathering,
17011 header rewriting and C source file generation. It should be much better
17012 than the old method: it now uses a modified version of Ulf's parser to
17013 read the ANSI prototypes in all header files (thus the old K&R definitions
17014 aren't needed for error creation any more) and do a better job of
17015 translating function codes into names. The old 'ASN1 error code embedded
17016 in a comment' is no longer necessary and it doesn't use .err files which
17017 have now been deleted. Also the error code call doesn't have to appear all
17018 on one line (which resulted in some large lines...).
17022 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17026 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17027 0 (which usually indicates a closed connection), but continue reading.
17031 * Fix some race conditions.
17035 * Add support for CRL distribution points extension. Add Certificate
17036 Policies and CRL distribution points documentation.
17040 * Move the autogenerated header file parts to crypto/opensslconf.h.
17044 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17045 8 of keying material. Merlin has also confirmed interop with this fix
17046 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17048 *Merlin Hughes <merlin@baltimore.ie>*
17050 * Fix lots of warnings.
17052 *Richard Levitte <levitte@stacken.kth.se>*
17054 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17055 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17057 *Richard Levitte <levitte@stacken.kth.se>*
17059 * Fix problems with sizeof(long) == 8.
17061 *Andy Polyakov <appro@fy.chalmers.se>*
17063 * Change functions to ANSI C.
17067 * Fix typos in error codes.
17069 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17071 * Remove defunct assembler files from Configure.
17075 * SPARC v8 assembler BIGNUM implementation.
17077 *Andy Polyakov <appro@fy.chalmers.se>*
17079 * Support for Certificate Policies extension: both print and set.
17080 Various additions to support the r2i method this uses.
17084 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17085 return a const string when you are expecting an allocated buffer.
17089 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17090 types DirectoryString and DisplayText.
17094 * Add code to allow r2i extensions to access the configuration database,
17095 add an LHASH database driver and add several ctx helper functions.
17099 * Fix an evil bug in bn_expand2() which caused various BN functions to
17100 fail when they extended the size of a BIGNUM.
17104 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17105 support typesafe stack.
17109 * Fix typo in SSL_[gs]et_options().
17111 *Nils Frostberg <nils@medcom.se>*
17113 * Delete various functions and files that belonged to the (now obsolete)
17114 old X509V3 handling code.
17118 * New Configure option "rsaref".
17122 * Don't auto-generate pem.h.
17126 * Introduce type-safe ASN.1 SETs.
17130 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17132 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17134 * Introduce type-safe STACKs. This will almost certainly break lots of code
17135 that links with OpenSSL (well at least cause lots of warnings), but fear
17136 not: the conversion is trivial, and it eliminates loads of evil casts. A
17137 few STACKed things have been converted already. Feel free to convert more.
17138 In the fullness of time, I'll do away with the STACK type altogether.
17142 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17143 specified in `<certfile>` by updating the entry in the index.txt file.
17144 This way one no longer has to edit the index.txt file manually for
17145 revoking a certificate. The -revoke option does the gory details now.
17147 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17149 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17150 `-text` option at all and this way the `-noout -text` combination was
17151 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17153 *Ralf S. Engelschall*
17155 * Make sure a corresponding plain text error message exists for the
17156 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17157 verify callback function determined that a certificate was revoked.
17159 *Ralf S. Engelschall*
17161 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17162 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17163 all available ciphers including rc5, which was forgotten until now.
17164 In order to let the testing shell script know which algorithms
17165 are available, a new (up to now undocumented) command
17166 `openssl list-cipher-commands` is used.
17170 * Bugfix: s_client occasionally would sleep in select() when
17171 it should have checked SSL_pending() first.
17175 * New functions DSA_do_sign and DSA_do_verify to provide access to
17176 the raw DSA values prior to ASN.1 encoding.
17180 * Tweaks to Configure
17182 *Niels Poppe <niels@netbox.org>*
17184 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17189 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17193 * New config option to avoid instructions that are illegal on the 80386.
17194 The default code is faster, but requires at least a 486.
17198 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17199 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17200 same as SSL2_VERSION anyway.
17204 * New "-showcerts" option for s_client.
17208 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17209 application. Various cleanups and fixes.
17213 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17214 modify error routines to work internally. Add error codes and PBE init
17215 to library startup routines.
17219 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17220 packing functions to asn1 and evp. Changed function names and error
17221 codes along the way.
17225 * PKCS12 integration: and so it begins... First of several patches to
17226 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17227 objects to objects.h
17231 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17232 and display support for Thawte strong extranet extension.
17236 * Add LinuxPPC support.
17238 *Jeff Dubrule <igor@pobox.org>*
17240 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17241 bn_div_words in alpha.s.
17243 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17245 * Make sure the RSA OAEP test is skipped under -DRSAref because
17246 OAEP isn't supported when OpenSSL is built with RSAref.
17248 *Ulf Moeller <ulf@fitug.de>*
17250 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17251 so they no longer are missing under -DNOPROTO.
17253 *Soren S. Jorvang <soren@t.dk>*
17255 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17257 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17258 doesn't work when the session is reused. Coming soon!
17262 * Fix a security hole, that allows sessions to be reused in the wrong
17263 context thus bypassing client cert protection! All software that uses
17264 client certs and session caches in multiple contexts NEEDS PATCHING to
17265 allow session reuse! A fuller solution is in the works.
17267 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17269 * Some more source tree cleanups (removed obsolete files
17270 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17271 permission on "config" script to be executable) and a fix for the INSTALL
17274 *Ulf Moeller <ulf@fitug.de>*
17276 * Remove some legacy and erroneous uses of malloc, free instead of
17279 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17281 * Make rsa_oaep_test return non-zero on error.
17283 *Ulf Moeller <ulf@fitug.de>*
17285 * Add support for native Solaris shared libraries. Configure
17286 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17287 if someone would make that last step automatic.
17289 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17291 * ctx_size was not built with the right compiler during "make links". Fixed.
17295 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17296 except NULL ciphers". This means the default cipher list will no longer
17297 enable NULL ciphers. They need to be specifically enabled e.g. with
17298 the string "DEFAULT:eNULL".
17302 * Fix to RSA private encryption routines: if p < q then it would
17303 occasionally produce an invalid result. This will only happen with
17304 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17308 * Be less restrictive and allow also `perl util/perlpath.pl
17309 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17310 because this way one can also use an interpreter named `perl5' (which is
17311 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17312 installed as `perl').
17314 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17316 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17318 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17320 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17321 advapi32.lib to Win32 build and change the pem test comparison
17322 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17323 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17324 and crypto/des/ede_cbcm_enc.c.
17328 * DES quad checksum was broken on big-endian architectures. Fixed.
17332 * Comment out two functions in bio.h that aren't implemented. Fix up the
17333 Win32 test batch file so it (might) work again. The Win32 test batch file
17334 is horrible: I feel ill....
17338 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17339 in e_os.h. Audit of header files to check ANSI and non ANSI
17340 sections: 10 functions were absent from non ANSI section and not exported
17341 from Windows DLLs. Fixed up libeay.num for new functions.
17345 * Make `openssl version' output lines consistent.
17347 *Ralf S. Engelschall*
17349 * Fix Win32 symbol export lists for BIO functions: Added
17350 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17351 to ms/libeay{16,32}.def.
17353 *Ralf S. Engelschall*
17355 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17356 fine under Unix and passes some trivial tests I've now added. But the
17357 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17358 added to make sure no one expects that this stuff really works in the
17359 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17360 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17363 *Ralf S. Engelschall*
17365 * Fix the generation of two part addresses in perl.
17367 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17369 * Add config entry for Linux on MIPS.
17371 *John Tobey <jtobey@channel1.com>*
17373 * Make links whenever Configure is run, unless we are on Windoze.
17377 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17378 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17383 * Add a useful kludge to allow package maintainers to specify compiler and
17384 other platforms details on the command line without having to patch the
17385 Configure script every time: One now can use
17386 `perl Configure <id>:<details>`,
17387 i.e. platform ids are allowed to have details appended
17388 to them (separated by colons). This is treated as there would be a static
17389 pre-configured entry in Configure's %table under key `<id>` with value
17390 `<details>` and `perl Configure <id>` is called. So, when you want to
17391 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17392 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17393 now, which overrides the FreeBSD-elf entry on-the-fly.
17395 *Ralf S. Engelschall*
17397 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17401 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17402 on the `perl Configure ...' command line. This way one can compile
17403 OpenSSL libraries with Position Independent Code (PIC) which is needed
17404 for linking it into DSOs.
17406 *Ralf S. Engelschall*
17408 * Remarkably, export ciphers were totally broken and no-one had noticed!
17413 * Cleaned up the LICENSE document: The official contact for any license
17414 questions now is the OpenSSL core team under openssl-core@openssl.org.
17415 And add a paragraph about the dual-license situation to make sure people
17416 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17417 to the OpenSSL toolkit.
17419 *Ralf S. Engelschall*
17421 * General source tree makefile cleanups: Made `making xxx in yyy...'
17422 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17423 Additionally cleaned up the `make links' target: Remove unnecessary
17424 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17425 to speed processing and no longer clutter the display with confusing
17426 stuff. Instead only the actually done links are displayed.
17428 *Ralf S. Engelschall*
17430 * Permit null encryption ciphersuites, used for authentication only. It used
17431 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17432 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17437 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17438 signed attributes when verifying signatures (this would break them),
17439 the detached data encoding was wrong and public keys obtained using
17440 X509_get_pubkey() weren't freed.
17444 * Add text documentation for the BUFFER functions. Also added a work around
17445 to a Win95 console bug. This was triggered by the password read stuff: the
17446 last character typed gets carried over to the next fread(). If you were
17447 generating a new cert request using 'req' for example then the last
17448 character of the passphrase would be CR which would then enter the first
17453 * Added the new 'Includes OpenSSL Cryptography Software' button as
17454 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17455 button and can be used by applications based on OpenSSL to show the
17456 relationship to the OpenSSL project.
17458 *Ralf S. Engelschall*
17460 * Remove confusing variables in function signatures in files
17461 ssl/ssl_lib.c and ssl/ssl.h.
17463 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17465 * Don't install bss_file.c under PREFIX/include/
17467 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17469 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17470 functions that return function pointers and has support for NT specific
17471 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17472 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17473 unsigned to signed types: this was killing the Win32 compile.
17477 * Add new certificate file to stack functions,
17478 SSL_add_dir_cert_subjects_to_stack() and
17479 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17480 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17481 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17482 This means that Apache-SSL and similar packages don't have to mess around
17483 to add as many CAs as they want to the preferred list.
17487 * Experiment with doxygen documentation. Currently only partially applied to
17489 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17490 openssl.doxy as the configuration file.
17494 * Get rid of remaining C++-style comments which strict C compilers hate.
17496 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17498 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17499 compiled in by default: it has problems with large keys.
17503 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17504 DH private keys and/or callback functions which directly correspond to
17505 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17506 is needed for applications which have to configure certificates on a
17507 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17509 For the RSA certificate situation is makes no difference, but
17510 for the DSA certificate situation this fixes the "no shared cipher"
17511 problem where the OpenSSL cipher selection procedure failed because the
17512 temporary keys were not overtaken from the context and the API provided
17513 no way to reconfigure them.
17514 The new functions now let applications reconfigure the stuff and they
17515 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17516 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17517 non-public-API function ssl_cert_instantiate() is used as a helper
17518 function and also to reduce code redundancy inside ssl_rsa.c.
17520 *Ralf S. Engelschall*
17522 * Move s_server -dcert and -dkey options out of the undocumented feature
17523 area because they are useful for the DSA situation and should be
17524 recognized by the users.
17526 *Ralf S. Engelschall*
17528 * Fix the cipher decision scheme for export ciphers: the export bits are
17529 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17530 SSL_EXP_MASK. So, the original variable has to be used instead of the
17531 already masked variable.
17533 *Richard Levitte <levitte@stacken.kth.se>*
17535 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17537 *Richard Levitte <levitte@stacken.kth.se>*
17539 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17540 from `int` to `unsigned int` because it is a length and initialized by
17541 EVP_DigestFinal() which expects an `unsigned int *`.
17543 *Richard Levitte <levitte@stacken.kth.se>*
17545 * Don't hard-code path to Perl interpreter on shebang line of Configure
17546 script. Instead use the usual Shell->Perl transition trick.
17548 *Ralf S. Engelschall*
17550 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17551 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17552 -noout -modulus' as it's already the case for `openssl rsa -noout
17553 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17554 currently the public key is printed (a decision which was already done by
17555 `openssl dsa -modulus' in the past) which serves a similar purpose.
17556 Additionally the NO_RSA no longer completely removes the whole -modulus
17557 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17560 *Ralf S. Engelschall*
17562 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17563 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17565 *Arne Ansper <arne@ats.cyber.ee>*
17567 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17568 to be added. Now both 'req' and 'ca' can use new objects defined in the
17573 * Add cool BIO that does syslog (or event log on NT).
17575 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17577 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17578 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17579 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17580 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17584 * Add preliminary config info for new extension code.
17588 * Make RSA_NO_PADDING really use no padding.
17590 *Ulf Moeller <ulf@fitug.de>*
17592 * Generate errors when private/public key check is done.
17596 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17597 for some CRL extensions and new objects added.
17601 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17602 key usage extension and fuller support for authority key id.
17606 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17607 padding method for RSA, which is recommended for new applications in PKCS
17608 #1 v2.0 (RFC 2437, October 1998).
17609 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17610 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17611 against Bleichbacher's attack on RSA.
17612 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17615 * Updates to the new SSL compression code
17617 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17619 * Fix so that the version number in the master secret, when passed
17620 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17621 (because the server will not accept higher), that the version number
17622 is 0x03,0x01, not 0x03,0x00
17624 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17626 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17627 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17628 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17632 * Support for RAW extensions where an arbitrary extension can be
17633 created by including its DER encoding. See apps/openssl.cnf for
17638 * Make sure latest Perl versions don't interpret some generated C array
17639 code as Perl array code in the crypto/err/err_genc.pl script.
17641 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17643 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17644 not many people have the assembler. Various Win32 compilation fixes and
17645 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17646 build instructions.
17650 * Modify configure script 'Configure' to automatically create crypto/date.h
17651 file under Win32 and also build pem.h from pem.org. New script
17652 util/mkfiles.pl to create the MINFO file on environments that can't do a
17653 'make files': perl util/mkfiles.pl >MINFO should work.
17657 * Major rework of DES function declarations, in the pursuit of correctness
17658 and purity. As a result, many evil casts evaporated, and some weirdness,
17659 too. You may find this causes warnings in your code. Zapping your evil
17660 casts will probably fix them. Mostly.
17664 * Fix for a typo in asn1.h. Bug fix to object creation script
17665 obj_dat.pl. It considered a zero in an object definition to mean
17666 "end of object": none of the objects in objects.h have any zeros
17667 so it wasn't spotted.
17669 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17671 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17672 Masking (CBCM). In the absence of test vectors, the best I have been able
17673 to do is check that the decrypt undoes the encrypt, so far. Send me test
17674 vectors if you have them.
17678 * Correct calculation of key length for export ciphers (too much space was
17679 allocated for null ciphers). This has not been tested!
17683 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17684 message is now correct (it understands "crypto" and "ssl" on its
17685 command line). There is also now an "update" option. This will update
17686 the util/ssleay.num and util/libeay.num files with any new functions.
17688 perl util/mkdef.pl crypto ssl update
17689 it will update them.
17693 * Overhauled the Perl interface:
17694 - ported BN stuff to OpenSSL's different BN library
17695 - made the perl/ source tree CVS-aware
17696 - renamed the package from SSLeay to OpenSSL (the files still contain
17697 their history because I've copied them in the repository)
17698 - removed obsolete files (the test scripts will be replaced
17699 by better Test::Harness variants in the future)
17701 *Ralf S. Engelschall*
17703 * First cut for a very conservative source tree cleanup:
17704 1. merge various obsolete readme texts into doc/ssleay.txt
17705 where we collect the old documents and readme texts.
17706 2. remove the first part of files where I'm already sure that we no
17707 longer need them because of three reasons: either they are just temporary
17708 files which were left by Eric or they are preserved original files where
17709 I've verified that the diff is also available in the CVS via "cvs diff
17710 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17711 the crypto/md/ stuff).
17713 *Ralf S. Engelschall*
17715 * More extension code. Incomplete support for subject and issuer alt
17716 name, issuer and authority key id. Change the i2v function parameters
17717 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17718 what that's for :-) Fix to ASN1 macro which messed up
17719 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17723 * Preliminary support for ENUMERATED type. This is largely copied from the
17728 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17730 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17732 * Make sure `make rehash` target really finds the `openssl` program.
17734 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17736 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17737 like to hear about it if this slows down other processors.
17741 * Add CygWin32 platform information to Configure script.
17743 *Alan Batie <batie@aahz.jf.intel.com>*
17745 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17747 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17749 * New program nseq to manipulate netscape certificate sequences
17753 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17758 * Fixes to BN code. Previously the default was to define BN_RECURSION
17759 but the BN code had some problems that would cause failures when
17760 doing certificate verification and some other functions.
17762 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17764 * Add ASN1 and PEM code to support netscape certificate sequences.
17768 * Add ASN1 and PEM code to support netscape certificate sequences.
17772 * Add several PKIX and private extended key usage OIDs.
17776 * Modify the 'ca' program to handle the new extension code. Modify
17777 openssl.cnf for new extension format, add comments.
17781 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17782 and add a sample to openssl.cnf so req -x509 now adds appropriate
17787 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17788 error code, add initial support to X509_print() and x509 application.
17792 * Takes a deep breath and start adding X509 V3 extension support code. Add
17793 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17794 stuff is currently isolated and isn't even compiled yet.
17798 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17799 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17800 Removed the versions check from X509 routines when loading extensions:
17801 this allows certain broken certificates that don't set the version
17802 properly to be processed.
17806 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17807 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17808 can still be regenerated with "make depend".
17812 * Spelling mistake in C version of CAST-128.
17814 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17816 * Changes to the error generation code. The perl script err-code.pl
17817 now reads in the old error codes and retains the old numbers, only
17818 adding new ones if necessary. It also only changes the .err files if new
17819 codes are added. The makefiles have been modified to only insert errors
17820 when needed (to avoid needlessly modifying header files). This is done
17821 by only inserting errors if the .err file is newer than the auto generated
17822 C file. To rebuild all the error codes from scratch (the old behaviour)
17823 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17824 or delete all the .err files.
17828 * CAST-128 was incorrectly implemented for short keys. The C version has
17829 been fixed, but is untested. The assembler versions are also fixed, but
17830 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17831 to regenerate it if needed.
17832 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17833 Hagino <itojun@kame.net>*
17835 * File was opened incorrectly in randfile.c.
17837 *Ulf Möller <ulf@fitug.de>*
17839 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17840 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17841 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17842 al: it's just almost always a UTCTime. Note this patch adds new error
17843 codes so do a "make errors" if there are problems.
17847 * Correct Linux 1 recognition in config.
17849 *Ulf Möller <ulf@fitug.de>*
17851 * Remove pointless MD5 hash when using DSA keys in ca.
17853 *Anonymous <nobody@replay.com>*
17855 * Generate an error if given an empty string as a cert directory. Also
17856 generate an error if handed NULL (previously returned 0 to indicate an
17857 error, but didn't set one).
17859 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17861 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17865 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17866 parameters. This was causing a warning which killed off the Win32 compile.
17870 * Remove C++ style comments from crypto/bn/bn_local.h.
17872 *Neil Costigan <neil.costigan@celocom.com>*
17874 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17875 based on a text string, looking up short and long names and finally
17876 "dot" format. The "dot" format stuff didn't work. Added new function
17877 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17878 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17879 OID is not part of the table.
17883 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17884 X509_LOOKUP_by_alias().
17888 * Sort openssl functions by name.
17892 * Get the gendsa program working (hopefully) and add it to app list. Remove
17893 encryption from sample DSA keys (in case anyone is interested the password
17898 * Make *all* `*_free` functions accept a NULL pointer.
17900 *Frans Heymans <fheymans@isaserver.be>*
17902 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17905 *Anonymous <nobody@replay.com>*
17907 * s_server should send the CAfile as acceptable CAs, not its own cert.
17909 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17911 * Don't blow it for numeric -newkey arguments to apps/req.
17913 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17915 * Temp key "for export" tests were wrong in s3_srvr.c.
17917 *Anonymous <nobody@replay.com>*
17919 * Add prototype for temp key callback functions
17920 SSL_CTX_set_tmp_{rsa,dh}_callback().
17924 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17925 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17929 * X509_name_add_entry() freed the wrong thing after an error.
17931 *Arne Ansper <arne@ats.cyber.ee>*
17933 * rsa_eay.c would attempt to free a NULL context.
17935 *Arne Ansper <arne@ats.cyber.ee>*
17937 * BIO_s_socket() had a broken should_retry() on Windoze.
17939 *Arne Ansper <arne@ats.cyber.ee>*
17941 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17943 *Arne Ansper <arne@ats.cyber.ee>*
17945 * Make sure the already existing X509_STORE->depth variable is initialized
17946 in X509_STORE_new(), but document the fact that this variable is still
17947 unused in the certificate verification process.
17949 *Ralf S. Engelschall*
17951 * Fix the various library and apps files to free up pkeys obtained from
17952 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17956 * Fix reference counting in X509_PUBKEY_get(). This makes
17957 demos/maurice/example2.c work, amongst others, probably.
17959 *Steve Henson and Ben Laurie*
17961 * First cut of a cleanup for apps/. First the `ssleay` program is now named
17962 `openssl` and second, the shortcut symlinks for the `openssl <command>`
17963 are no longer created. This way we have a single and consistent command
17964 line interface `openssl <command>`, similar to `cvs <command>`.
17966 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17968 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17969 BIT STRING wrapper always have zero unused bits.
17973 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17977 * Make the top-level INSTALL documentation easier to understand.
17981 * Makefiles updated to exit if an error occurs in a sub-directory
17982 make (including if user presses ^C) [Paul Sutton]
17984 * Make Montgomery context stuff explicit in RSA data structure.
17988 * Fix build order of pem and err to allow for generated pem.h.
17992 * Fix renumbering bug in X509_NAME_delete_entry().
17996 * Enhanced the err-ins.pl script so it makes the error library number
17997 global and can add a library name. This is needed for external ASN1 and
17998 other error libraries.
18002 * Fixed sk_insert which never worked properly.
18006 * Fix ASN1 macros so they can handle indefinite length constructed
18007 EXPLICIT tags. Some non standard certificates use these: they can now
18012 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18013 into a single doc/ssleay.txt bundle. This way the information is still
18014 preserved but no longer messes up this directory. Now it's new room for
18015 the new set of documentation files.
18017 *Ralf S. Engelschall*
18019 * SETs were incorrectly DER encoded. This was a major pain, because they
18020 shared code with SEQUENCEs, which aren't coded the same. This means that
18021 almost everything to do with SETs or SEQUENCEs has either changed name or
18022 number of arguments.
18024 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18026 * Fix test data to work with the above.
18030 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18031 was already fixed by Eric for 0.9.1 it seems.
18033 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18035 * Autodetect FreeBSD3.
18039 * Fix various bugs in Configure. This affects the following platforms:
18043 unixware-2.0-pentium
18048 * Eliminate generated files from CVS. Reorder tests to regenerate files
18049 before they are needed.
18053 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18057 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18059 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18060 changed SSLeay to OpenSSL in version strings.
18062 *Ralf S. Engelschall*
18064 * Some fixups to the top-level documents.
18068 * Fixed the nasty bug where rsaref.h was not found under compile-time
18069 because the symlink to include/ was missing.
18071 *Ralf S. Engelschall*
18073 * Incorporated the popular no-RSA/DSA-only patches
18074 which allow to compile a RSA-free SSLeay.
18076 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18078 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18079 when "ssleay" is still not found.
18081 *Ralf S. Engelschall*
18083 * Added more platforms to Configure: Cray T3E, HPUX 11,
18085 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18087 * Updated the README file.
18089 *Ralf S. Engelschall*
18091 * Added various .cvsignore files in the CVS repository subdirs
18092 to make a "cvs update" really silent.
18094 *Ralf S. Engelschall*
18096 * Recompiled the error-definition header files and added
18097 missing symbols to the Win32 linker tables.
18099 *Ralf S. Engelschall*
18101 * Cleaned up the top-level documents;
18102 o new files: CHANGES and LICENSE
18103 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18104 o merged COPYRIGHT into LICENSE
18105 o removed obsolete TODO file
18106 o renamed MICROSOFT to INSTALL.W32
18108 *Ralf S. Engelschall*
18110 * Removed dummy files from the 0.9.1b source tree:
18111 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18112 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18113 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18114 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18115 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18117 *Ralf S. Engelschall*
18119 * Added various platform portability fixes.
18123 * The Genesis of the OpenSSL rpject:
18124 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18125 Young and Tim J. Hudson created while they were working for C2Net until
18128 *The OpenSSL Project*
18130 ### Changes between 0.9.0b and 0.9.1b [not released]
18132 * Updated a few CA certificates under certs/
18136 * Changed some BIGNUM api stuff.
18140 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18141 DGUX x86, Linux Alpha, etc.
18145 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18146 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18151 * Add -strparse option to asn1pars program which parses nested
18154 *Dr Stephen Henson <shenson@bigfoot.com>*
18156 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18160 * DSA fix for "ca" program.
18164 * Added "-genkey" option to "dsaparam" program.
18168 * Added RIPE MD160 (rmd160) message digest.
18172 * Added -a (all) option to "ssleay version" command.
18176 * Added PLATFORM define which is the id given to Configure.
18180 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18184 * Extended the ASN.1 parser routines.
18188 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18192 * Added a BN_CTX to the BN library.
18196 * Fixed the weak key values in DES library
18200 * Changed API in EVP library for cipher aliases.
18204 * Added support for RC2/64bit cipher.
18208 * Converted the lhash library to the crypto/mem.c functions.
18212 * Added more recognized ASN.1 object ids.
18216 * Added more RSA padding checks for SSL/TLS.
18220 * Added BIO proxy/filter functionality.
18224 * Added extra_certs to SSL_CTX which can be used
18225 send extra CA certificates to the client in the CA cert chain sending
18226 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18230 * Now Fortezza is denied in the authentication phase because
18231 this is key exchange mechanism is not supported by SSLeay at all.
18235 * Additional PKCS1 checks.
18239 * Support the string "TLSv1" for all TLS v1 ciphers.
18243 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18244 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18248 * Fixed a few memory leaks.
18252 * Fixed various code and comment typos.
18256 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18257 bytes sent in the client random.
18259 *Edward Bishop <ebishop@spyglass.com>*
18263 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18264 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18265 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18266 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18267 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18268 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18269 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18270 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18271 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18272 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18273 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18274 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18275 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18276 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18277 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18278 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18279 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18280 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18281 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18282 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18283 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18284 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18285 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18286 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18287 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18288 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18289 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18290 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18291 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18292 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18293 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18294 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18295 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18296 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18297 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18298 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18299 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18300 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18301 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18302 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18303 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18304 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18305 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18306 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18307 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18308 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18309 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18310 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18311 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18312 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18313 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18314 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18315 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18316 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18317 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18318 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18319 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18320 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18321 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18322 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18323 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18324 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18325 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18326 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18327 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18328 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18329 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18330 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18331 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18332 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18333 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18334 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18335 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18336 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18337 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18338 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18339 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18340 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18341 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18342 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18343 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18344 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18345 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18346 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18347 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18348 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18349 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18350 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18351 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18352 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18353 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18354 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18355 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18356 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18357 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18358 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18359 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18360 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18361 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18362 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18363 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18364 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18365 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18366 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18367 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18368 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18369 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18370 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18371 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18372 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18373 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18374 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18375 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18376 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18377 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18378 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18379 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18380 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18381 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18382 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18383 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18384 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18385 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18386 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18387 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18388 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18389 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18390 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18391 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18392 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18393 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18394 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18395 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18396 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18397 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18398 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18399 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18400 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18401 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18402 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18403 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18404 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18405 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18406 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18407 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18408 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18409 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18410 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18411 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18412 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18413 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18414 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18415 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18416 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18417 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18418 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18419 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18420 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18421 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18422 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655