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 * Handshake now fails if Extended Master Secret extension is dropped
31 * Dropped interactive mode from the 'openssl' program. From now on,
32 the `openssl` command without arguments is equivalent to `openssl
37 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
38 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
39 While the old function names have been retained for backward compatibility
40 they should not be used in new developments
41 because their return values are confusing: Unlike other `_cmp()` functions
42 they do not return 0 in case their arguments are equal.
46 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
47 EC_GROUP_get_field_type().
51 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
52 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
53 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
54 Applications should rely on the library automatically assigning a suitable
55 EC_METHOD internally upon EC_GROUP construction.
59 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
60 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
61 assigned internally without application intervention.
62 Users of EC_GROUP_new() should switch to a different suitable constructor.
66 * Add CAdES-BES signature verification support, mostly derived
67 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
69 *Filipe Raimundo da Silva*
71 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
75 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
76 functions are not widely used and now OpenSSL automatically perform this
77 conversion when needed.
81 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
82 EC_KEY_precompute_mult(). These functions are not widely used and
83 applications should instead switch to named curves which OpenSSL has
84 hardcoded lookup tables for.
88 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
89 should instead use the L<EC_POINT_mul(3)> function.
93 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
94 that are not applicable to the new provider model. Applications should
95 instead use EVP_default_properties_is_fips_enabled() and
96 EVP_default_properties_enable_fips().
100 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
101 is set, an unexpected EOF is ignored, it pretends a close notify was received
102 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
106 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
107 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
108 used and applications should instead use the
109 L<EC_POINT_set_affine_coordinates(3)> and
110 L<EC_POINT_get_affine_coordinates(3)> functions.
114 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
115 arrays to be more easily constructed via a series of utility functions.
116 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
117 the various push functions and finally convert to a passable OSSL_PARAM
118 array using OSSL_PARAM_BLD_to_param().
122 * The security strength of SHA1 and MD5 based signatures in TLS has been
123 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
124 working at the default security level of 1 and instead requires security
125 level 0. The security level can be changed either using the cipher string
126 with @SECLEVEL, or calling SSL_CTX_set_security_level().
130 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
131 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
132 internal keys, if they correspond to one of those built in types.
136 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
137 contain a provider side internal key.
141 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
142 They are old functions that we don't use, and that you could disable with
143 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
147 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
148 have been converted to Markdown with the goal to produce documents
149 which not only look pretty when viewed online in the browser, but
150 remain well readable inside a plain text editor.
152 To achieve this goal, a 'minimalistic' Markdown style has been applied
153 which avoids formatting elements that interfere too much with the
154 reading flow in the text file. For example, it
156 * avoids [ATX headings][] and uses [setext headings][] instead
157 (which works for `<h1>` and `<h2>` headings only).
158 * avoids [inline links][] and uses [reference links][] instead.
159 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
161 [ATX headings]: https://github.github.com/gfm/#atx-headings
162 [setext headings]: https://github.github.com/gfm/#setext-headings
163 [inline links]: https://github.github.com/gfm/#inline-link
164 [reference links]: https://github.github.com/gfm/#reference-link
165 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
166 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
168 *Matthias St. Pierre*
170 * The test suite is changed to preserve results of each test recipe.
171 A new directory test-runs/ with subdirectories named like the
172 test recipes are created in the build tree for this purpose.
176 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
177 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
178 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
180 *David von Oheimb, Martin Peylo*
182 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
183 The legacy OCSP-focused and only partly documented API is retained.
184 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
188 * BIO_do_connect and BIO_do_handshake have been extended:
189 If domain name resolution yields multiple IP addresses all of them are tried
190 after connect() failures.
194 * All of the low level RSA functions have been deprecated including:
196 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
197 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
198 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
199 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
200 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
201 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
202 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
203 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
204 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
205 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
206 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
207 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
208 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
209 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
210 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
211 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
212 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
213 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
214 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
215 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
216 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
217 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
218 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
219 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
220 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
221 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
222 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
223 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
225 Use of these low level functions has been informally discouraged for a long
226 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
227 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
228 L<EVP_PKEY_decrypt(3)>.
232 * X509 certificates signed using SHA1 are no longer allowed at security
234 In TLS/SSL the default security level is 1. It can be set either
235 using the cipher string with @SECLEVEL, or calling
236 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
237 a call to SSL_CTX_use_certificate() will fail if the security level is not
239 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
240 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
245 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
246 modified to use PKEY APIs. These commands are now in maintenance mode
247 and no new features will be added to them.
251 * The command line utility rsautl has been deprecated.
252 Instead use the pkeyutl program.
256 * The command line utilities genrsa and rsa have been modified to use PKEY
257 APIs These commands are now in maintenance mode and no new features will
262 * All of the low level DH functions have been deprecated including:
264 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
265 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
266 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
267 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
268 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
269 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
270 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
271 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
272 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
273 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
274 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
275 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
276 DH_meth_get_generate_params and DH_meth_set_generate_params.
278 Use of these low level functions has been informally discouraged for a long
279 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
280 and L<EVP_PKEY_derive(3)>.
284 * All of the low level DSA functions have been deprecated including:
286 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
287 DSA_get_default_method, DSA_set_method, DSA_get_method,
288 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
289 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
290 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
291 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
292 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
293 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
294 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
295 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
296 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
297 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
298 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
301 Use of these low level functions has been informally discouraged for a long
302 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
303 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
307 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
308 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
309 This means that applications don't have to look at the curve NID and
310 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
311 However, they still can, that EVP_PKEY_set_alias_type() call acts as
312 a no-op when the EVP_PKEY is already of the given type.
314 Parameter and key generation is also reworked to make it possible
315 to generate EVP_PKEY_SM2 parameters and keys without having to go
316 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
317 However, code that does the latter will still work as before.
321 * Deprecated low level ECDH and ECDSA functions. These include:
323 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
324 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
327 Use of these low level functions has been informally discouraged for a long
328 time. Instead applications should use the EVP_PKEY_derive(3),
329 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
333 * Deprecated the EC_KEY_METHOD functions. These include:
335 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
336 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
337 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
338 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
339 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
340 EC_KEY_METHOD_get_verify.
342 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
346 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
347 and EVP_PKEY_decrypt() instead.
348 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
349 and EVP_PKEY_encrypt() instead.
353 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
354 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
355 a new formulation to include all the things it can be used for,
356 as well as words of caution.
360 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
361 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
365 * All of the low level HMAC functions have been deprecated including:
367 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
368 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
371 Use of these low level functions has been informally discouraged for a long
372 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
373 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
374 and L<EVP_MAC_final(3)>.
378 * Over two thousand fixes were made to the documentation, including:
379 - Common options (such as -rand/-writerand, TLS version control, etc)
380 were refactored and point to newly-enhanced descriptions in openssl.pod.
381 - Added style conformance for all options (with help from Richard Levitte),
382 documented all reported missing options, added a CI build to check
383 that all options are documented and that no unimplemented options
385 - Documented some internals, such as all use of environment variables.
386 - Addressed all internal broken L<> references.
390 * All of the low level CMAC functions have been deprecated including:
392 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
393 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
395 Use of these low level functions has been informally discouraged for a long
396 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
397 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
398 and L<EVP_MAC_final(3)>.
402 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
403 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
406 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
407 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
408 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
409 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
410 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
411 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
412 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
413 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
414 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
415 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
417 Use of these low level functions has been informally discouraged
418 for a long time. Applications should use the EVP_DigestInit_ex(3),
419 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
423 * Corrected the documentation of the return values from the `EVP_DigestSign*`
424 set of functions. The documentation mentioned negative values for some
425 errors, but this was never the case, so the mention of negative values
428 Code that followed the documentation and thereby check with something
429 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
433 * All of the low level cipher functions have been deprecated including:
435 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
436 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
437 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
438 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
439 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
440 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
441 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
442 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
443 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
444 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
445 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
446 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
447 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
448 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
449 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
450 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
451 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
452 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
453 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
454 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
455 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
456 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
457 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
458 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
459 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
460 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
461 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
462 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
463 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
465 Use of these low level functions has been informally discouraged for
466 a long time. Applications should use the high level EVP APIs, e.g.
467 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
468 equivalently named decrypt functions instead.
470 *Matt Caswell and Paul Dale*
472 * Removed include/openssl/opensslconf.h.in and replaced it with
473 include/openssl/configuration.h.in, which differs in not including
474 <openssl/macros.h>. A short header include/openssl/opensslconf.h
475 was added to include both.
477 This allows internal hacks where one might need to modify the set
478 of configured macros, for example this if deprecated symbols are
479 still supposed to be available internally:
481 #include <openssl/configuration.h>
483 #undef OPENSSL_NO_DEPRECATED
484 #define OPENSSL_SUPPRESS_DEPRECATED
486 #include <openssl/macros.h>
488 This should not be used by applications that use the exported
489 symbols, as that will lead to linking errors.
493 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
494 used in exponentiation with 512-bit moduli. No EC algorithms are
495 affected. Analysis suggests that attacks against 2-prime RSA1024,
496 3-prime RSA1536, and DSA1024 as a result of this defect would be very
497 difficult to perform and are not believed likely. Attacks against DH512
498 are considered just feasible. However, for an attack the target would
499 have to re-use the DH512 private key, which is not recommended anyway.
500 Also applications directly using the low level API BN_mod_exp may be
501 affected if they use BN_FLG_CONSTTIME.
506 * Most memory-debug features have been deprecated, and the functionality
507 replaced with no-ops.
511 * Added documentation for the STACK API. OpenSSL only defines the STACK
512 functions where they are used.
516 * Introduced a new method type and API, OSSL_SERIALIZER, to
517 represent generic serializers. An implementation is expected to
518 be able to serialize an object associated with a given name (such
519 as an algorithm name for an asymmetric key) into forms given by
520 implementation properties.
522 Serializers are primarily used from inside libcrypto, through
523 calls to functions like EVP_PKEY_print_private(),
524 PEM_write_bio_PrivateKey() and similar.
526 Serializers are specified in such a way that they can be made to
527 directly handle the provider side portion of an object, if this
528 provider side part comes from the same provider as the serializer
529 itself, but can also be made to handle objects in parametrized
530 form (as an OSSL_PARAM array of data). This allows a provider to
531 offer generic serializers as a service for any other provider.
535 * Added a .pragma directive to the syntax of configuration files, to
536 allow varying behavior in a supported and predictable manner.
537 Currently added pragma:
541 This allows dollar signs to be a keyword character unless it's
542 followed by a opening brace or parenthesis. This is useful for
543 platforms where dollar signs are commonly used in names, such as
544 volume names and system directory names on VMS.
548 * Added functionality to create an EVP_PKEY from user data. This
549 is effectively the same as creating a RSA, DH or DSA object and
550 then assigning them to an EVP_PKEY, but directly using algorithm
551 agnostic EVP functions. A benefit is that this should be future
552 proof for public key algorithms to come.
556 * Change the interpretation of the '--api' configuration option to
557 mean that this is a desired API compatibility level with no
558 further meaning. The previous interpretation, that this would
559 also mean to remove all deprecated symbols up to and including
560 the given version, no requires that 'no-deprecated' is also used
561 in the configuration.
563 When building applications, the desired API compatibility level
564 can be set with the OPENSSL_API_COMPAT macro like before. For
565 API compatibility version below 3.0, the old style numerical
566 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
567 For version 3.0 and on, the value is expected to be the decimal
568 value calculated from the major and minor version like this:
570 MAJOR * 10000 + MINOR * 100
574 -DOPENSSL_API_COMPAT=30000 For 3.0
575 -DOPENSSL_API_COMPAT=30200 For 3.2
577 To hide declarations that are deprecated up to and including the
578 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
579 given when building the application as well.
583 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
584 access to certificate and CRL stores via URIs and OSSL_STORE
587 This adds the following functions:
589 - X509_LOOKUP_store()
590 - X509_STORE_load_file()
591 - X509_STORE_load_path()
592 - X509_STORE_load_store()
593 - SSL_add_store_cert_subjects_to_stack()
594 - SSL_CTX_set_default_verify_store()
595 - SSL_CTX_load_verify_file()
596 - SSL_CTX_load_verify_dir()
597 - SSL_CTX_load_verify_store()
601 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
602 The presence of this system service is determined at run-time.
606 * Added functionality to create an EVP_PKEY context based on data
607 for methods from providers. This takes an algorithm name and a
608 property query string and simply stores them, with the intent
609 that any operation that uses this context will use those strings
610 to fetch the needed methods implicitly, thereby making the port
611 of application written for pre-3.0 OpenSSL easier.
615 * The undocumented function NCONF_WIN32() has been deprecated; for
616 conversion details see the HISTORY section of doc/man5/config.pod
620 * Introduced the new functions EVP_DigestSignInit_ex() and
621 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
622 EVP_DigestVerifyUpdate() have been converted to functions. See the man
623 pages for further details.
627 * Over two thousand fixes were made to the documentation, including:
628 adding missing command flags, better style conformance, documentation
631 *Rich Salz, Richard Levitte*
633 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
634 X25519, X448, Ed25519 and Ed448.
638 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
643 * Deprecated the public definition of ERR_STATE as well as the function
644 ERR_get_state(). This is done in preparation of making ERR_STATE an
649 * Added ERR functionality to give callers access to the stored function
650 names that have replaced the older function code based functions.
652 New functions are ERR_get_error_func(), ERR_peek_error_func(),
653 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
654 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
655 and ERR_peek_last_error_all().
657 These functions have become deprecated: ERR_get_error_line_data(),
658 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
659 ERR_func_error_string().
663 * Extended testing to be verbose for failing tests only. The make variables
664 VERBOSE_FAILURE or VF can be used to enable this:
666 $ make VF=1 test # Unix
667 $ mms /macro=(VF=1) test ! OpenVMS
668 $ nmake VF=1 test # Windows
672 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
673 used even when parsing explicit parameters, when loading a serialized key
674 or calling `EC_GROUP_new_from_ecpkparameters()`/
675 `EC_GROUP_new_from_ecparameters()`.
676 This prevents bypass of security hardening and performance gains,
677 especially for curves with specialized EC_METHODs.
678 By default, if a key encoded with explicit parameters is loaded and later
679 serialized, the output is still encoded with explicit parameters, even if
680 internally a "named" EC_GROUP is used for computation.
684 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
685 this change, EC_GROUP_set_generator would accept order and/or cofactor as
686 NULL. After this change, only the cofactor parameter can be NULL. It also
687 does some minimal sanity checks on the passed order.
692 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
693 An attack is simple, if the first CMS_recipientInfo is valid but the
694 second CMS_recipientInfo is chosen ciphertext. If the second
695 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
696 encryption key will be replaced by garbage, and the message cannot be
697 decoded, but if the RSA decryption fails, the correct encryption key is
698 used and the recipient will not notice the attack.
699 As a work around for this potential attack the length of the decrypted
700 key must be equal to the cipher default key length, in case the
701 certifiate is not given and all recipientInfo are tried out.
702 The old behaviour can be re-enabled in the CMS code by setting the
703 CMS_DEBUG_DECRYPT flag.
707 * Early start up entropy quality from the DEVRANDOM seed source has been
708 improved for older Linux systems. The RAND subsystem will wait for
709 /dev/random to be producing output before seeding from /dev/urandom.
710 The seeded state is stored for future library initialisations using
711 a system global shared memory segment. The shared memory identifier
712 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
713 the desired value. The default identifier is 114.
717 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
718 when primes for RSA keys are computed.
719 Since we previously always generated primes == 2 (mod 3) for RSA keys,
720 the 2-prime and 3-prime RSA modules were easy to distinguish, since
721 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
722 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
723 This avoids possible fingerprinting of newly generated RSA modules.
727 * Correct the extended master secret constant on EBCDIC systems. Without this
728 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
729 negotiate EMS will fail. Unfortunately this also means that TLS connections
730 between EBCDIC systems with this fix, and EBCDIC systems without this
731 fix will fail if they negotiate EMS.
735 * Changed the library initialisation so that the config file is now loaded
736 by default. This was already the case for libssl. It now occurs for both
737 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
738 OPENSSL_init_crypto() to suppress automatic loading of a config file.
742 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
743 where the former acts as a replacement for ERR_put_error(), and the
744 latter replaces the combination ERR_put_error()+ERR_add_error_data().
745 ERR_raise_data() adds more flexibility by taking a format string and
746 an arbitrary number of arguments following it, to be processed with
751 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
752 to check if a named provider is loaded and available. When called, it
753 will also activate all fallback providers if such are still present.
757 * Enforce a minimum DH modulus size of 512 bits.
761 * Changed DH parameters to generate the order q subgroup instead of 2q.
762 Previously generated DH parameters are still accepted by DH_check
763 but DH_generate_key works around that by clearing bit 0 of the
764 private key for those. This avoids leaking bit 0 of the private key.
768 * Significantly reduce secure memory usage by the randomness pools.
772 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
777 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
778 algorithms. An implementation of a key exchange algorithm can be obtained
779 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
780 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
781 the older EVP_PKEY_derive_init() function. See the man pages for the new
782 functions for further details.
786 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
790 * Removed the function names from error messages and deprecated the
793 * Removed NextStep support and the macro OPENSSL_UNISTD
797 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
798 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
799 Also removed "export var as function" capability; we do not export
800 variables, only functions.
804 * RC5_32_set_key has been changed to return an int type, with 0 indicating
805 an error and 1 indicating success. In previous versions of OpenSSL this
806 was a void type. If a key was set longer than the maximum possible this
811 * Support SM2 signing and verification schemes with X509 certificate.
815 * Use SHA256 as the default digest for TS query in the ts app.
819 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
820 This checks that the salt length is at least 128 bits, the derived key
821 length is at least 112 bits, and that the iteration count is at least 1000.
822 For backwards compatibility these checks are disabled by default in the
823 default provider, but are enabled by default in the fips provider.
824 To enable or disable these checks use the control
825 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
829 * Default cipher lists/suites are now available via a function, the
830 #defines are deprecated.
834 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
835 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
836 for Windows Store apps easier. Also, the "no-uplink" option has been added.
840 * Join the directories crypto/x509 and crypto/x509v3
844 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
845 This changes the size when using the genpkey app when no size is given. It
846 fixes an omission in earlier changes that changed all RSA, DSA and DH
847 generation apps to use 2048 bits by default.
851 * Added command 'openssl kdf' that uses the EVP_KDF API.
855 * Added command 'openssl mac' that uses the EVP_MAC API.
859 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
860 as default directories. Also added the command 'openssl info'
861 for scripting purposes.
865 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
866 deprecated. These undocumented functions were never integrated into the EVP
867 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
868 Bi-directional IGE mode. These modes were never formally standardised and
869 usage of these functions is believed to be very small. In particular
870 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
871 is ever used. The security implications are believed to be minimal, but
872 this issue was never fixed for backwards compatibility reasons. New code
873 should not use these modes.
877 * Add prediction resistance to the DRBG reseeding process.
881 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
882 mandated by IEEE Std 1619-2018.
886 * Added newline escaping functionality to a filename when using openssl dgst.
887 This output format is to replicate the output format found in the `*sum`
888 checksum programs. This aims to preserve backward compatibility.
890 *Matt Eaton, Richard Levitte, and Paul Dale*
892 * Removed the heartbeat message in DTLS feature, as it has very
893 little usage and doesn't seem to fulfill a valuable purpose.
894 The configuration option is now deprecated.
898 * Changed the output of 'openssl {digestname} < file' to display the
899 digest name in its output.
903 * Added a new generic trace API which provides support for enabling
904 instrumentation through trace output. This feature is mainly intended
905 as an aid for developers and is disabled by default. To utilize it,
906 OpenSSL needs to be configured with the `enable-trace` option.
908 If the tracing API is enabled, the application can activate trace output
909 by registering BIOs as trace channels for a number of tracing and debugging
912 The 'openssl' application has been expanded to enable any of the types
913 available via environment variables defined by the user, and serves as
914 one possible example on how to use this functionality.
916 *Richard Levitte & Matthias St. Pierre*
918 * Added build tests for C++. These are generated files that only do one
919 thing, to include one public OpenSSL head file each. This tests that
920 the public header files can be usefully included in a C++ application.
922 This test isn't enabled by default. It can be enabled with the option
923 'enable-buildtest-c++'.
927 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
931 * Add KMAC to EVP_MAC.
935 * Added property based algorithm implementation selection framework to
940 * Added SCA hardening for modular field inversion in EC_GROUP through
941 a new dedicated field_inv() pointer in EC_METHOD.
942 This also addresses a leakage affecting conversions from projective
943 to affine coordinates.
945 *Billy Bob Brumley, Nicola Tuveri*
947 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
948 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
949 those algorithms that were already supported through the EVP_PKEY API
950 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
951 and scrypt are now wrappers that call EVP_KDF.
955 * Build devcrypto engine as a dynamic engine.
959 * Add keyed BLAKE2 to EVP_MAC.
963 * Fix a bug in the computation of the endpoint-pair shared secret used
964 by DTLS over SCTP. This breaks interoperability with older versions
965 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
966 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
967 interoperability with such broken implementations. However, enabling
968 this switch breaks interoperability with correct implementations.
970 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
971 re-used X509_PUBKEY object if the second PUBKEY is malformed.
975 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
979 * Change the license to the Apache License v2.0.
983 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
985 - Major releases (indicated by incrementing the MAJOR release number)
986 may introduce incompatible API/ABI changes.
987 - Minor releases (indicated by incrementing the MINOR release number)
988 may introduce new features but retain API/ABI compatibility.
989 - Patch releases (indicated by incrementing the PATCH number)
990 are intended for bug fixes and other improvements of existing
991 features only (like improving performance or adding documentation)
992 and retain API/ABI compatibility.
996 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1000 * Remove the 'dist' target and add a tarball building script. The
1001 'dist' target has fallen out of use, and it shouldn't be
1002 necessary to configure just to create a source distribution.
1006 * Recreate the OS390-Unix config target. It no longer relies on a
1007 special script like it did for OpenSSL pre-1.1.0.
1011 * Instead of having the source directories listed in Configure, add
1012 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1017 * Add GMAC to EVP_MAC.
1021 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1025 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1026 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1027 to facilitate the continued use of MACs through raw private keys in
1028 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
1032 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1033 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1037 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1038 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1039 are retained for backwards compatibility.
1043 * AES-XTS mode now enforces that its two keys are different to mitigate
1044 the attacked described in "Efficient Instantiations of Tweakable
1045 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1046 Details of this attack can be obtained from:
1047 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1051 * Rename the object files, i.e. give them other names than in previous
1052 versions. Their names now include the name of the final product, as
1053 well as its type mnemonic (bin, lib, shlib).
1057 * Added new option for 'openssl list', '-objects', which will display the
1058 list of built in objects, i.e. OIDs with names.
1062 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1063 improves application performance by removing data copies and providing
1064 applications with zero-copy system calls such as sendfile and splice.
1071 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1073 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1075 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1076 while reading in libssl then we would report an error back to the
1077 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1078 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1079 therefore give a hint as to what went wrong.
1083 * Check that ed25519 and ed448 are allowed by the security level. Previously
1084 signature algorithms not using an MD were not being checked that they were
1085 allowed by the security level.
1089 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1090 was not quite right. The behaviour was not consistent between resumption
1091 and normal handshakes, and also not quite consistent with historical
1092 behaviour. The behaviour in various scenarios has been clarified and
1093 it has been updated to make it match historical behaviour as closely as
1098 * *[VMS only]* The header files that the VMS compilers include automatically,
1099 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1100 that the C++ compiler doesn't understand. This is a shortcoming in the
1101 compiler, but can be worked around with `__cplusplus` guards.
1103 C++ applications that use OpenSSL libraries must be compiled using the
1104 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1105 functions. Otherwise, only functions with symbols of less than 31
1106 characters can be used, as the linker will not be able to successfully
1107 resolve symbols with longer names.
1111 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1112 The presence of this system service is determined at run-time.
1116 * Added newline escaping functionality to a filename when using openssl dgst.
1117 This output format is to replicate the output format found in the `*sum`
1118 checksum programs. This aims to preserve backward compatibility.
1120 *Matt Eaton, Richard Levitte, and Paul Dale*
1122 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1127 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1129 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1130 number generator (RNG). This was intended to include protection in the
1131 event of a fork() system call in order to ensure that the parent and child
1132 processes did not share the same RNG state. However this protection was not
1133 being used in the default case.
1135 A partial mitigation for this issue is that the output from a high
1136 precision timer is mixed into the RNG state so the likelihood of a parent
1137 and child process sharing state is significantly reduced.
1139 If an application already calls OPENSSL_init_crypto() explicitly using
1140 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1143 *Matthias St. Pierre*
1145 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1146 used even when parsing explicit parameters, when loading a serialized key
1147 or calling `EC_GROUP_new_from_ecpkparameters()`/
1148 `EC_GROUP_new_from_ecparameters()`.
1149 This prevents bypass of security hardening and performance gains,
1150 especially for curves with specialized EC_METHODs.
1151 By default, if a key encoded with explicit parameters is loaded and later
1152 serialized, the output is still encoded with explicit parameters, even if
1153 internally a "named" EC_GROUP is used for computation.
1157 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1158 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1159 NULL. After this change, only the cofactor parameter can be NULL. It also
1160 does some minimal sanity checks on the passed order.
1165 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1166 An attack is simple, if the first CMS_recipientInfo is valid but the
1167 second CMS_recipientInfo is chosen ciphertext. If the second
1168 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1169 encryption key will be replaced by garbage, and the message cannot be
1170 decoded, but if the RSA decryption fails, the correct encryption key is
1171 used and the recipient will not notice the attack.
1172 As a work around for this potential attack the length of the decrypted
1173 key must be equal to the cipher default key length, in case the
1174 certifiate is not given and all recipientInfo are tried out.
1175 The old behaviour can be re-enabled in the CMS code by setting the
1176 CMS_DEBUG_DECRYPT flag.
1181 * Early start up entropy quality from the DEVRANDOM seed source has been
1182 improved for older Linux systems. The RAND subsystem will wait for
1183 /dev/random to be producing output before seeding from /dev/urandom.
1184 The seeded state is stored for future library initialisations using
1185 a system global shared memory segment. The shared memory identifier
1186 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1187 the desired value. The default identifier is 114.
1191 * Correct the extended master secret constant on EBCDIC systems. Without this
1192 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1193 negotiate EMS will fail. Unfortunately this also means that TLS connections
1194 between EBCDIC systems with this fix, and EBCDIC systems without this
1195 fix will fail if they negotiate EMS.
1199 * Use Windows installation paths in the mingw builds
1201 Mingw isn't a POSIX environment per se, which means that Windows
1202 paths should be used for installation.
1207 * Changed DH_check to accept parameters with order q and 2q subgroups.
1208 With order 2q subgroups the bit 0 of the private key is not secret
1209 but DH_generate_key works around that by clearing bit 0 of the
1210 private key for those. This avoids leaking bit 0 of the private key.
1214 * Significantly reduce secure memory usage by the randomness pools.
1218 * Revert the DEVRANDOM_WAIT feature for Linux systems
1220 The DEVRANDOM_WAIT feature added a select() call to wait for the
1221 /dev/random device to become readable before reading from the
1222 /dev/urandom device.
1224 It turned out that this change had negative side effects on
1225 performance which were not acceptable. After some discussion it
1226 was decided to revert this feature and leave it up to the OS
1227 resp. the platform maintainer to ensure a proper initialization
1228 during early boot time.
1230 *Matthias St. Pierre*
1232 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1234 * Add build tests for C++. These are generated files that only do one
1235 thing, to include one public OpenSSL head file each. This tests that
1236 the public header files can be usefully included in a C++ application.
1238 This test isn't enabled by default. It can be enabled with the option
1239 'enable-buildtest-c++'.
1243 * Enable SHA3 pre-hashing for ECDSA and DSA.
1247 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1248 This changes the size when using the genpkey app when no size is given. It
1249 fixes an omission in earlier changes that changed all RSA, DSA and DH
1250 generation apps to use 2048 bits by default.
1254 * Reorganize the manual pages to consistently have RETURN VALUES,
1255 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1256 util/fix-doc-nits accordingly.
1258 *Paul Yang, Joshua Lock*
1260 * Add the missing accessor EVP_PKEY_get0_engine()
1264 * Have apps like 's_client' and 's_server' output the signature scheme
1265 along with other cipher suite parameters when debugging.
1269 * Make OPENSSL_config() error agnostic again.
1273 * Do the error handling in RSA decryption constant time.
1277 * Prevent over long nonces in ChaCha20-Poly1305.
1279 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1280 for every encryption operation. RFC 7539 specifies that the nonce value
1281 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1282 and front pads the nonce with 0 bytes if it is less than 12
1283 bytes. However it also incorrectly allows a nonce to be set of up to 16
1284 bytes. In this case only the last 12 bytes are significant and any
1285 additional leading bytes are ignored.
1287 It is a requirement of using this cipher that nonce values are
1288 unique. Messages encrypted using a reused nonce value are susceptible to
1289 serious confidentiality and integrity attacks. If an application changes
1290 the default nonce length to be longer than 12 bytes and then makes a
1291 change to the leading bytes of the nonce expecting the new value to be a
1292 new unique nonce then such an application could inadvertently encrypt
1293 messages with a reused nonce.
1295 Additionally the ignored bytes in a long nonce are not covered by the
1296 integrity guarantee of this cipher. Any application that relies on the
1297 integrity of these ignored leading bytes of a long nonce may be further
1298 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1299 is safe because no such use sets such a long nonce value. However user
1300 applications that use this cipher directly and set a non-default nonce
1301 length to be longer than 12 bytes may be vulnerable.
1303 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1309 * Add DEVRANDOM_WAIT feature for Linux systems
1311 On older Linux systems where the getrandom() system call is not available,
1312 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1313 Contrary to getrandom(), the /dev/urandom device will not block during
1314 early boot when the kernel CSPRNG has not been seeded yet.
1316 To mitigate this known weakness, use select() to wait for /dev/random to
1317 become readable before reading from /dev/urandom.
1319 * Ensure that SM2 only uses SM3 as digest algorithm
1323 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1325 * Change the info callback signals for the start and end of a post-handshake
1326 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1327 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1328 confused by this and assume that a TLSv1.2 renegotiation has started. This
1329 can break KeyUpdate handling. Instead we no longer signal the start and end
1330 of a post handshake message exchange (although the messages themselves are
1331 still signalled). This could break some applications that were expecting
1332 the old signals. However without this KeyUpdate is not usable for many
1337 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1339 * Timing vulnerability in DSA signature generation
1341 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1342 timing side channel attack. An attacker could use variations in the signing
1343 algorithm to recover the private key.
1345 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1350 * Timing vulnerability in ECDSA signature generation
1352 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1353 timing side channel attack. An attacker could use variations in the signing
1354 algorithm to recover the private key.
1356 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1361 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1362 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1363 of two gigabytes and the error handling improved.
1365 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1366 categorized as a normal bug, not a security issue, because the DRBG reseeds
1367 automatically and is fully functional even without additional randomness
1368 provided by the application.
1370 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1372 * Add a new ClientHello callback. Provides a callback interface that gives
1373 the application the ability to adjust the nascent SSL object at the
1374 earliest stage of ClientHello processing, immediately after extensions have
1375 been collected but before they have been processed. In particular, this
1376 callback can adjust the supported TLS versions in response to the contents
1381 * Add SM2 base algorithm support.
1385 * s390x assembly pack: add (improved) hardware-support for the following
1386 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1387 aes-cfb/cfb8, aes-ecb.
1391 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1392 parameter is no longer accepted, as it leads to a corrupt table. NULL
1393 pem_str is reserved for alias entries only.
1397 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1398 step for prime curves. The new implementation is based on formulae from
1399 differential addition-and-doubling in homogeneous projective coordinates
1400 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1401 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1402 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1403 to work in projective coordinates.
1405 *Billy Bob Brumley, Nicola Tuveri*
1407 * Change generating and checking of primes so that the error rate of not
1408 being prime depends on the intended use based on the size of the input.
1409 For larger primes this will result in more rounds of Miller-Rabin.
1410 The maximal error rate for primes with more than 1080 bits is lowered
1413 *Kurt Roeckx, Annie Yousar*
1415 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1419 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1420 moving between systems, and to avoid confusion when a Windows build is
1421 done with mingw vs with MSVC. For POSIX installs, there's still a
1422 symlink or copy named 'tsget' to avoid that confusion as well.
1426 * Revert blinding in ECDSA sign and instead make problematic addition
1427 length-invariant. Switch even to fixed-length Montgomery multiplication.
1431 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1432 step for binary curves. The new implementation is based on formulae from
1433 differential addition-and-doubling in mixed Lopez-Dahab projective
1434 coordinates, modified to independently blind the operands.
1436 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1438 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1439 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1440 EC_METHODs to implement their own specialized "ladder step", to take
1441 advantage of more favorable coordinate systems or more efficient
1442 differential addition-and-doubling algorithms.
1444 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1446 * Modified the random device based seed sources to keep the relevant
1447 file descriptors open rather than reopening them on each access.
1448 This allows such sources to operate in a chroot() jail without
1449 the associated device nodes being available. This behaviour can be
1450 controlled using RAND_keep_random_devices_open().
1454 * Numerous side-channel attack mitigations have been applied. This may have
1455 performance impacts for some algorithms for the benefit of improved
1456 security. Specific changes are noted in this change log by their respective
1461 * AIX shared library support overhaul. Switch to AIX "natural" way of
1462 handling shared libraries, which means collecting shared objects of
1463 different versions and bitnesses in one common archive. This allows to
1464 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1465 doesn't affect the way 3rd party applications are linked, only how
1466 multi-version installation is managed.
1470 * Make ec_group_do_inverse_ord() more robust and available to other
1471 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1472 mitigations are applied to the fallback BN_mod_inverse().
1473 When using this function rather than BN_mod_inverse() directly, new
1474 EC cryptosystem implementations are then safer-by-default.
1478 * Add coordinate blinding for EC_POINT and implement projective
1479 coordinate blinding for generic prime curves as a countermeasure to
1480 chosen point SCA attacks.
1482 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1484 * Add blinding to ECDSA and DSA signatures to protect against side channel
1485 attacks discovered by Keegan Ryan (NCC Group).
1489 * Enforce checking in the pkeyutl command line app to ensure that the input
1490 length does not exceed the maximum supported digest length when performing
1491 a sign, verify or verifyrecover operation.
1495 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1496 I/O in combination with something like select() or poll() will hang. This
1497 can be turned off again using SSL_CTX_clear_mode().
1498 Many applications do not properly handle non-application data records, and
1499 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1500 around the problems in those applications, but can also break some.
1501 It's recommended to read the manpages about SSL_read(), SSL_write(),
1502 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1503 SSL_CTX_set_read_ahead() again.
1507 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1508 now allow empty (zero character) pass phrases.
1512 * Apply blinding to binary field modular inversion and remove patent
1513 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1517 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1518 binary and prime elliptic curves.
1522 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1523 constant time fixed point multiplication.
1527 * Revise elliptic curve scalar multiplication with timing attack
1528 defenses: ec_wNAF_mul redirects to a constant time implementation
1529 when computing fixed point and variable point multiplication (which
1530 in OpenSSL are mostly used with secret scalars in keygen, sign,
1531 ECDH derive operations).
1532 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1535 * Updated CONTRIBUTING
1539 * Updated DRBG / RAND to request nonce and additional low entropy
1540 randomness from the system.
1542 *Matthias St. Pierre*
1544 * Updated 'openssl rehash' to use OpenSSL consistent default.
1548 * Moved the load of the ssl_conf module to libcrypto, which helps
1549 loading engines that libssl uses before libssl is initialised.
1553 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1557 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1559 *Ingo Schwarze, Rich Salz*
1561 * Added output of accepting IP address and port for 'openssl s_server'
1565 * Added a new API for TLSv1.3 ciphersuites:
1566 SSL_CTX_set_ciphersuites()
1567 SSL_set_ciphersuites()
1571 * Memory allocation failures consistently add an error to the error
1576 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1577 in libcrypto when run as setuid/setgid.
1581 * Load any config file by default when libssl is used.
1585 * Added new public header file <openssl/rand_drbg.h> and documentation
1586 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1588 *Matthias St. Pierre*
1590 * QNX support removed (cannot find contributors to get their approval
1591 for the license change).
1595 * TLSv1.3 replay protection for early data has been implemented. See the
1596 SSL_read_early_data() man page for further details.
1600 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1601 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1602 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1603 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1604 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1605 configuration has been separated out. See the ciphers man page or the
1606 SSL_CTX_set_ciphersuites() man page for more information.
1610 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1611 in responder mode now supports the new "-multi" option, which
1612 spawns the specified number of child processes to handle OCSP
1613 requests. The "-timeout" option now also limits the OCSP
1614 responder's patience to wait to receive the full client request
1615 on a newly accepted connection. Child processes are respawned
1616 as needed, and the CA index file is automatically reloaded
1617 when changed. This makes it possible to run the "ocsp" responder
1618 as a long-running service, making the OpenSSL CA somewhat more
1619 feature-complete. In this mode, most diagnostic messages logged
1620 after entering the event loop are logged via syslog(3) rather than
1625 * Added support for X448 and Ed448. Heavily based on original work by
1630 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1631 objects loaded. This adds the functions OSSL_STORE_expect() and
1632 OSSL_STORE_find() as well as needed tools to construct searches and
1633 get the search data out of them.
1637 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1638 version of OpenSSL should review their configuration settings to ensure
1639 that they are still appropriate for TLSv1.3. For further information see:
1640 <https://wiki.openssl.org/index.php/TLS1.3>
1644 * Grand redesign of the OpenSSL random generator
1646 The default RAND method now utilizes an AES-CTR DRBG according to
1647 NIST standard SP 800-90Ar1. The new random generator is essentially
1648 a port of the default random generator from the OpenSSL FIPS 2.0
1649 object module. It is a hybrid deterministic random bit generator
1650 using an AES-CTR bit stream and which seeds and reseeds itself
1651 automatically using trusted system entropy sources.
1653 Some of its new features are:
1654 - Support for multiple DRBG instances with seed chaining.
1655 - The default RAND method makes use of a DRBG.
1656 - There is a public and private DRBG instance.
1657 - The DRBG instances are fork-safe.
1658 - Keep all global DRBG instances on the secure heap if it is enabled.
1659 - The public and private DRBG instance are per thread for lock free
1662 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1664 * Changed Configure so it only says what it does and doesn't dump
1665 so much data. Instead, ./configdata.pm should be used as a script
1666 to display all sorts of configuration data.
1670 * Added processing of "make variables" to Configure.
1674 * Added SHA512/224 and SHA512/256 algorithm support.
1678 * The last traces of Netware support, first removed in 1.1.0, have
1683 * Get rid of Makefile.shared, and in the process, make the processing
1684 of certain files (rc.obj, or the .def/.map/.opt files produced from
1685 the ordinal files) more visible and hopefully easier to trace and
1686 debug (or make silent).
1690 * Make it possible to have environment variable assignments as
1691 arguments to config / Configure.
1695 * Add multi-prime RSA (RFC 8017) support.
1699 * Add SM3 implemented according to GB/T 32905-2016
1700 * Jack Lloyd <jack.lloyd@ribose.com>,
1701 Ronald Tse <ronald.tse@ribose.com>,
1702 Erick Borsboom <erick.borsboom@ribose.com> *
1704 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1705 as documented in RFC6066.
1706 Based on a patch from Tomasz Moń
1708 *Filipe Raimundo da Silva*
1710 * Add SM4 implemented according to GB/T 32907-2016.
1711 * Jack Lloyd <jack.lloyd@ribose.com>,
1712 Ronald Tse <ronald.tse@ribose.com>,
1713 Erick Borsboom <erick.borsboom@ribose.com> *
1715 * Reimplement -newreq-nodes and ERR_error_string_n; the
1716 original author does not agree with the license change.
1720 * Add ARIA AEAD TLS support.
1724 * Some macro definitions to support VS6 have been removed. Visual
1725 Studio 6 has not worked since 1.1.0
1729 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1730 without clearing the errors.
1734 * Add "atfork" functions. If building on a system that without
1735 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1736 requirements. The RAND facility now uses/requires this.
1744 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1745 not possible to disable entirely. However, it's still possible to
1746 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1749 To disable, configure with 'no-ui-console'. 'no-ui' is still
1750 possible to use as an alias. Check at compile time with the
1751 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1752 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1756 * Add a STORE module, which implements a uniform and URI based reader of
1757 stores that can contain keys, certificates, CRLs and numerous other
1758 objects. The main API is loosely based on a few stdio functions,
1759 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1760 OSSL_STORE_error and OSSL_STORE_close.
1761 The implementation uses backends called "loaders" to implement arbitrary
1762 URI schemes. There is one built in "loader" for the 'file' scheme.
1766 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1767 then adjusted to work on FreeBSD 8.4 as well.
1768 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1769 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1773 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1774 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1775 error code calls like this:
1777 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1779 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1780 that can be encoded in C. For the foreseeable future, this will only
1783 *Richard Levitte and Tim Hudson*
1785 * Removed BSD cryptodev engine.
1789 * Add a build target 'build_all_generated', to build all generated files
1790 and only that. This can be used to prepare everything that requires
1791 things like perl for a system that lacks perl and then move everything
1792 to that system and do the rest of the build there.
1796 * In the UI interface, make it possible to duplicate the user data. This
1797 can be used by engines that need to retain the data for a longer time
1798 than just the call where this user data is passed.
1802 * Ignore the '-named_curve auto' value for compatibility of applications
1805 *Tomas Mraz <tmraz@fedoraproject.org>*
1807 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1808 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1809 alerts across multiple records (some of which could be empty). In practice
1810 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1811 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1812 support this at all. Supporting it adds significant complexity to the
1813 record layer, and its removal is unlikely to cause interoperability
1818 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1819 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1820 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1825 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1826 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1828 *Richard Levitte, Andy Polyakov*
1830 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1835 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1836 platform rather than 'mingw'.
1840 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1841 success if they are asked to add an object which already exists
1842 in the store. This change cascades to other functions which load
1843 certificates and CRLs.
1847 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1848 facilitate stack unwinding even from assembly subroutines.
1852 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1853 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1857 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1858 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1859 which is the minimum version we support.
1863 * Certificate time validation (X509_cmp_time) enforces stricter
1864 compliance with RFC 5280. Fractional seconds and timezone offsets
1865 are no longer allowed.
1869 * Add support for ARIA
1873 * s_client will now send the Server Name Indication (SNI) extension by
1874 default unless the new "-noservername" option is used. The server name is
1875 based on the host provided to the "-connect" option unless overridden by
1876 using "-servername".
1880 * Add support for SipHash
1884 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1885 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1886 prevent issues where no progress is being made and the peer continually
1887 sends unrecognised record types, using up resources processing them.
1891 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1892 using the algorithm defined in
1893 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1897 * Heartbeat support has been removed; the ABI is changed for now.
1899 *Richard Levitte, Rich Salz*
1901 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1905 * The RSA "null" method, which was partially supported to avoid patent
1906 issues, has been replaced to always returns NULL.
1913 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1915 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1916 used even when parsing explicit parameters, when loading a serialized key
1917 or calling `EC_GROUP_new_from_ecpkparameters()`/
1918 `EC_GROUP_new_from_ecparameters()`.
1919 This prevents bypass of security hardening and performance gains,
1920 especially for curves with specialized EC_METHODs.
1921 By default, if a key encoded with explicit parameters is loaded and later
1922 serialized, the output is still encoded with explicit parameters, even if
1923 internally a "named" EC_GROUP is used for computation.
1927 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1928 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1929 NULL. After this change, only the cofactor parameter can be NULL. It also
1930 does some minimal sanity checks on the passed order.
1935 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1936 An attack is simple, if the first CMS_recipientInfo is valid but the
1937 second CMS_recipientInfo is chosen ciphertext. If the second
1938 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1939 encryption key will be replaced by garbage, and the message cannot be
1940 decoded, but if the RSA decryption fails, the correct encryption key is
1941 used and the recipient will not notice the attack.
1942 As a work around for this potential attack the length of the decrypted
1943 key must be equal to the cipher default key length, in case the
1944 certifiate is not given and all recipientInfo are tried out.
1945 The old behaviour can be re-enabled in the CMS code by setting the
1946 CMS_DEBUG_DECRYPT flag.
1951 * Use Windows installation paths in the mingw builds
1953 Mingw isn't a POSIX environment per se, which means that Windows
1954 paths should be used for installation.
1959 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
1961 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1962 This changes the size when using the genpkey app when no size is given. It
1963 fixes an omission in earlier changes that changed all RSA, DSA and DH
1964 generation apps to use 2048 bits by default.
1968 * Prevent over long nonces in ChaCha20-Poly1305.
1970 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1971 for every encryption operation. RFC 7539 specifies that the nonce value
1972 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1973 and front pads the nonce with 0 bytes if it is less than 12
1974 bytes. However it also incorrectly allows a nonce to be set of up to 16
1975 bytes. In this case only the last 12 bytes are significant and any
1976 additional leading bytes are ignored.
1978 It is a requirement of using this cipher that nonce values are
1979 unique. Messages encrypted using a reused nonce value are susceptible to
1980 serious confidentiality and integrity attacks. If an application changes
1981 the default nonce length to be longer than 12 bytes and then makes a
1982 change to the leading bytes of the nonce expecting the new value to be a
1983 new unique nonce then such an application could inadvertently encrypt
1984 messages with a reused nonce.
1986 Additionally the ignored bytes in a long nonce are not covered by the
1987 integrity guarantee of this cipher. Any application that relies on the
1988 integrity of these ignored leading bytes of a long nonce may be further
1989 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1990 is safe because no such use sets such a long nonce value. However user
1991 applications that use this cipher directly and set a non-default nonce
1992 length to be longer than 12 bytes may be vulnerable.
1994 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2000 * Added SCA hardening for modular field inversion in EC_GROUP through
2001 a new dedicated field_inv() pointer in EC_METHOD.
2002 This also addresses a leakage affecting conversions from projective
2003 to affine coordinates.
2005 *Billy Bob Brumley, Nicola Tuveri*
2007 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2008 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2012 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2016 * Remove the 'dist' target and add a tarball building script. The
2017 'dist' target has fallen out of use, and it shouldn't be
2018 necessary to configure just to create a source distribution.
2022 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2024 * Timing vulnerability in DSA signature generation
2026 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2027 timing side channel attack. An attacker could use variations in the signing
2028 algorithm to recover the private key.
2030 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2035 * Timing vulnerability in ECDSA signature generation
2037 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2038 timing side channel attack. An attacker could use variations in the signing
2039 algorithm to recover the private key.
2041 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2046 * Add coordinate blinding for EC_POINT and implement projective
2047 coordinate blinding for generic prime curves as a countermeasure to
2048 chosen point SCA attacks.
2050 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2052 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2054 * Client DoS due to large DH parameter
2056 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2057 malicious server can send a very large prime value to the client. This will
2058 cause the client to spend an unreasonably long period of time generating a
2059 key for this prime resulting in a hang until the client has finished. This
2060 could be exploited in a Denial Of Service attack.
2062 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2067 * Cache timing vulnerability in RSA Key Generation
2069 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2070 a cache timing side channel attack. An attacker with sufficient access to
2071 mount cache timing attacks during the RSA key generation process could
2072 recover the private key.
2074 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2075 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2080 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2081 parameter is no longer accepted, as it leads to a corrupt table. NULL
2082 pem_str is reserved for alias entries only.
2086 * Revert blinding in ECDSA sign and instead make problematic addition
2087 length-invariant. Switch even to fixed-length Montgomery multiplication.
2091 * Change generating and checking of primes so that the error rate of not
2092 being prime depends on the intended use based on the size of the input.
2093 For larger primes this will result in more rounds of Miller-Rabin.
2094 The maximal error rate for primes with more than 1080 bits is lowered
2097 *Kurt Roeckx, Annie Yousar*
2099 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2103 * Add blinding to ECDSA and DSA signatures to protect against side channel
2104 attacks discovered by Keegan Ryan (NCC Group).
2108 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2109 now allow empty (zero character) pass phrases.
2113 * Certificate time validation (X509_cmp_time) enforces stricter
2114 compliance with RFC 5280. Fractional seconds and timezone offsets
2115 are no longer allowed.
2119 * Fixed a text canonicalisation bug in CMS
2121 Where a CMS detached signature is used with text content the text goes
2122 through a canonicalisation process first prior to signing or verifying a
2123 signature. This process strips trailing space at the end of lines, converts
2124 line terminators to CRLF and removes additional trailing line terminators
2125 at the end of a file. A bug in the canonicalisation process meant that
2126 some characters, such as form-feed, were incorrectly treated as whitespace
2127 and removed. This is contrary to the specification (RFC5485). This fix
2128 could mean that detached text data signed with an earlier version of
2129 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2130 signed with a fixed OpenSSL may fail to verify with an earlier version of
2131 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2132 and use the "-binary" flag (for the "cms" command line application) or set
2133 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2137 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2139 * Constructed ASN.1 types with a recursive definition could exceed the stack
2141 Constructed ASN.1 types with a recursive definition (such as can be found
2142 in PKCS7) could eventually exceed the stack given malicious input with
2143 excessive recursion. This could result in a Denial Of Service attack. There
2144 are no such structures used within SSL/TLS that come from untrusted sources
2145 so this is considered safe.
2147 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2153 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2155 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2156 effectively reduced to only comparing the least significant bit of each
2157 byte. This allows an attacker to forge messages that would be considered as
2158 authenticated in an amount of tries lower than that guaranteed by the
2159 security claims of the scheme. The module can only be compiled by the
2160 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2162 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2168 * Add a build target 'build_all_generated', to build all generated files
2169 and only that. This can be used to prepare everything that requires
2170 things like perl for a system that lacks perl and then move everything
2171 to that system and do the rest of the build there.
2175 * Backport SSL_OP_NO_RENGOTIATION
2177 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2178 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2179 changes this is no longer possible in 1.1.0. Therefore the new
2180 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2181 1.1.0 to provide equivalent functionality.
2183 Note that if an application built against 1.1.0h headers (or above) is run
2184 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2185 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2189 * Removed the OS390-Unix config target. It relied on a script that doesn't
2194 * rsaz_1024_mul_avx2 overflow bug on x86_64
2196 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2197 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2198 Analysis suggests that attacks against RSA and DSA as a result of this
2199 defect would be very difficult to perform and are not believed likely.
2200 Attacks against DH1024 are considered just feasible, because most of the
2201 work necessary to deduce information about a private key may be performed
2202 offline. The amount of resources required for such an attack would be
2203 significant. However, for an attack on TLS to be meaningful, the server
2204 would have to share the DH1024 private key among multiple clients, which is
2205 no longer an option since CVE-2016-0701.
2207 This only affects processors that support the AVX2 but not ADX extensions
2208 like Intel Haswell (4th generation).
2210 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2211 was originally found via the OSS-Fuzz project.
2216 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2218 * bn_sqrx8x_internal carry bug on x86_64
2220 There is a carry propagating bug in the x86_64 Montgomery squaring
2221 procedure. No EC algorithms are affected. Analysis suggests that attacks
2222 against RSA and DSA as a result of this defect would be very difficult to
2223 perform and are not believed likely. Attacks against DH are considered just
2224 feasible (although very difficult) because most of the work necessary to
2225 deduce information about a private key may be performed offline. The amount
2226 of resources required for such an attack would be very significant and
2227 likely only accessible to a limited number of attackers. An attacker would
2228 additionally need online access to an unpatched system using the target
2229 private key in a scenario with persistent DH parameters and a private
2230 key that is shared between multiple clients.
2232 This only affects processors that support the BMI1, BMI2 and ADX extensions
2233 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2235 This issue was reported to OpenSSL by the OSS-Fuzz project.
2240 * Malformed X.509 IPAddressFamily could cause OOB read
2242 If an X.509 certificate has a malformed IPAddressFamily extension,
2243 OpenSSL could do a one-byte buffer overread. The most likely result
2244 would be an erroneous display of the certificate in text format.
2246 This issue was reported to OpenSSL by the OSS-Fuzz project.
2251 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2253 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2254 platform rather than 'mingw'.
2258 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2259 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2260 which is the minimum version we support.
2264 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2266 * Encrypt-Then-Mac renegotiation crash
2268 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2269 negotiated where it was not in the original handshake (or vice-versa) then
2270 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2271 and servers are affected.
2273 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2278 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2280 * Truncated packet could crash via OOB read
2282 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2283 cipher is being used, then a truncated packet can cause that host to
2284 perform an out-of-bounds read, usually resulting in a crash.
2286 This issue was reported to OpenSSL by Robert Święcki of Google.
2291 * Bad (EC)DHE parameters cause a client crash
2293 If a malicious server supplies bad parameters for a DHE or ECDHE key
2294 exchange then this can result in the client attempting to dereference a
2295 NULL pointer leading to a client crash. This could be exploited in a Denial
2298 This issue was reported to OpenSSL by Guido Vranken.
2303 * BN_mod_exp may produce incorrect results on x86_64
2305 There is a carry propagating bug in the x86_64 Montgomery squaring
2306 procedure. No EC algorithms are affected. Analysis suggests that attacks
2307 against RSA and DSA as a result of this defect would be very difficult to
2308 perform and are not believed likely. Attacks against DH are considered just
2309 feasible (although very difficult) because most of the work necessary to
2310 deduce information about a private key may be performed offline. The amount
2311 of resources required for such an attack would be very significant and
2312 likely only accessible to a limited number of attackers. An attacker would
2313 additionally need online access to an unpatched system using the target
2314 private key in a scenario with persistent DH parameters and a private
2315 key that is shared between multiple clients. For example this can occur by
2316 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2317 similar to CVE-2015-3193 but must be treated as a separate problem.
2319 This issue was reported to OpenSSL by the OSS-Fuzz project.
2324 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2326 * ChaCha20/Poly1305 heap-buffer-overflow
2328 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2329 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2330 crash. This issue is not considered to be exploitable beyond a DoS.
2332 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2337 * CMS Null dereference
2339 Applications parsing invalid CMS structures can crash with a NULL pointer
2340 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2341 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2342 structure callback if an attempt is made to free certain invalid encodings.
2343 Only CHOICE structures using a callback which do not handle NULL value are
2346 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2351 * Montgomery multiplication may produce incorrect results
2353 There is a carry propagating bug in the Broadwell-specific Montgomery
2354 multiplication procedure that handles input lengths divisible by, but
2355 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2356 and DH private keys are impossible. This is because the subroutine in
2357 question is not used in operations with the private key itself and an input
2358 of the attacker's direct choice. Otherwise the bug can manifest itself as
2359 transient authentication and key negotiation failures or reproducible
2360 erroneous outcome of public-key operations with specially crafted input.
2361 Among EC algorithms only Brainpool P-512 curves are affected and one
2362 presumably can attack ECDH key negotiation. Impact was not analyzed in
2363 detail, because pre-requisites for attack are considered unlikely. Namely
2364 multiple clients have to choose the curve in question and the server has to
2365 share the private key among them, neither of which is default behaviour.
2366 Even then only clients that chose the curve will be affected.
2368 This issue was publicly reported as transient failures and was not
2369 initially recognized as a security issue. Thanks to Richard Morgan for
2370 providing reproducible case.
2375 * Removed automatic addition of RPATH in shared libraries and executables,
2376 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2380 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2382 * Fix Use After Free for large message sizes
2384 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2385 message larger than approx 16k is received then the underlying buffer to
2386 store the incoming message is reallocated and moved. Unfortunately a
2387 dangling pointer to the old location is left which results in an attempt to
2388 write to the previously freed location. This is likely to result in a
2389 crash, however it could potentially lead to execution of arbitrary code.
2391 This issue only affects OpenSSL 1.1.0a.
2393 This issue was reported to OpenSSL by Robert Święcki.
2398 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2400 * OCSP Status Request extension unbounded memory growth
2402 A malicious client can send an excessively large OCSP Status Request
2403 extension. If that client continually requests renegotiation, sending a
2404 large OCSP Status Request extension each time, then there will be unbounded
2405 memory growth on the server. This will eventually lead to a Denial Of
2406 Service attack through memory exhaustion. Servers with a default
2407 configuration are vulnerable even if they do not support OCSP. Builds using
2408 the "no-ocsp" build time option are not affected.
2410 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2415 * SSL_peek() hang on empty record
2417 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2418 sends an empty record. This could be exploited by a malicious peer in a
2419 Denial Of Service attack.
2421 This issue was reported to OpenSSL by Alex Gaynor.
2426 * Excessive allocation of memory in tls_get_message_header() and
2427 dtls1_preprocess_fragment()
2429 A (D)TLS message includes 3 bytes for its length in the header for the
2430 message. This would allow for messages up to 16Mb in length. Messages of
2431 this length are excessive and OpenSSL includes a check to ensure that a
2432 peer is sending reasonably sized messages in order to avoid too much memory
2433 being consumed to service a connection. A flaw in the logic of version
2434 1.1.0 means that memory for the message is allocated too early, prior to
2435 the excessive message length check. Due to way memory is allocated in
2436 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2437 to service a connection. This could lead to a Denial of Service through
2438 memory exhaustion. However, the excessive message length check still takes
2439 place, and this would cause the connection to immediately fail. Assuming
2440 that the application calls SSL_free() on the failed connection in a timely
2441 manner then the 21Mb of allocated memory will then be immediately freed
2442 again. Therefore the excessive memory allocation will be transitory in
2443 nature. This then means that there is only a security impact if:
2445 1) The application does not call SSL_free() in a timely manner in the event
2446 that the connection fails
2448 2) The application is working in a constrained environment where there is
2449 very little free memory
2451 3) The attacker initiates multiple connection attempts such that there are
2452 multiple connections in a state where memory has been allocated for the
2453 connection; SSL_free() has not yet been called; and there is insufficient
2454 memory to service the multiple requests.
2456 Except in the instance of (1) above any Denial Of Service is likely to be
2457 transitory because as soon as the connection fails the memory is
2458 subsequently freed again in the SSL_free() call. However there is an
2459 increased risk during this period of application crashes due to the lack of
2460 memory - which would then mean a more serious Denial of Service.
2462 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2463 (CVE-2016-6307 and CVE-2016-6308)
2467 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2468 had to be removed. Primary reason is that vendor assembler can't
2469 assemble our modules with -KPIC flag. As result it, assembly
2470 support, was not even available as option. But its lack means
2471 lack of side-channel resistant code, which is incompatible with
2472 security by todays standards. Fortunately gcc is readily available
2473 prepackaged option, which we firmly point at...
2477 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2479 * Windows command-line tool supports UTF-8 opt-in option for arguments
2480 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2481 (to any value) allows Windows user to access PKCS#12 file generated
2482 with Windows CryptoAPI and protected with non-ASCII password, as well
2483 as files generated under UTF-8 locale on Linux also protected with
2488 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2489 have been disabled by default and removed from DEFAULT, just like RC4.
2490 See the RC4 item below to re-enable both.
2494 * The method for finding the storage location for the Windows RAND seed file
2495 has changed. First we check %RANDFILE%. If that is not set then we check
2496 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2497 all else fails we fall back to C:\.
2501 * The EVP_EncryptUpdate() function has had its return type changed from void
2502 to int. A return of 0 indicates and error while a return of 1 indicates
2507 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2508 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2509 off the constant time implementation for RSA, DSA and DH have been made
2510 no-ops and deprecated.
2514 * Windows RAND implementation was simplified to only get entropy by
2515 calling CryptGenRandom(). Various other RAND-related tickets
2518 *Joseph Wylie Yandle, Rich Salz*
2520 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2521 and `OPENSSL_LH_`, respectively. The old names are available
2522 with API compatibility. They new names are now completely documented.
2526 * Unify TYPE_up_ref(obj) methods signature.
2527 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2528 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2529 int (instead of void) like all others TYPE_up_ref() methods.
2530 So now these methods also check the return value of CRYPTO_atomic_add(),
2531 and the validity of object reference counter.
2533 *fdasilvayy@gmail.com*
2535 * With Windows Visual Studio builds, the .pdb files are installed
2536 alongside the installed libraries and executables. For a static
2537 library installation, ossl_static.pdb is the associate compiler
2538 generated .pdb file to be used when linking programs.
2542 * Remove openssl.spec. Packaging files belong with the packagers.
2546 * Automatic Darwin/OSX configuration has had a refresh, it will now
2547 recognise x86_64 architectures automatically. You can still decide
2548 to build for a different bitness with the environment variable
2549 KERNEL_BITS (can be 32 or 64), for example:
2551 KERNEL_BITS=32 ./config
2555 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2556 256 bit AES and HMAC with SHA256.
2560 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2564 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2568 * To enable users to have their own config files and build file templates,
2569 Configure looks in the directory indicated by the environment variable
2570 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2571 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2572 name and is used as is.
2576 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2577 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2578 X509_CERT_FILE_CTX was removed.
2582 * "shared" builds are now the default. To create only static libraries use
2583 the "no-shared" Configure option.
2587 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2588 All of these option have not worked for some while and are fundamental
2593 * Make various cleanup routines no-ops and mark them as deprecated. Most
2594 global cleanup functions are no longer required because they are handled
2595 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2596 Explicitly de-initing can cause problems (e.g. where a library that uses
2597 OpenSSL de-inits, but an application is still using it). The affected
2598 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2599 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2600 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2601 COMP_zlib_cleanup().
2605 * --strict-warnings no longer enables runtime debugging options
2606 such as REF_DEBUG. Instead, debug options are automatically
2607 enabled with '--debug' builds.
2609 *Andy Polyakov, Emilia Käsper*
2611 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2612 have been moved out of the public header files. New functions for managing
2613 these have been added.
2617 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2618 objects have been moved out of the public header files. New
2619 functions for managing these have been added.
2623 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2624 have been moved out of the public header files. New functions for managing
2625 these have been added.
2629 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2630 moved out of the public header files. New functions for managing these
2635 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2639 * Removed the mk1mf build scripts.
2643 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2644 it is always safe to #include a header now.
2648 * Removed the aged BC-32 config and all its supporting scripts
2652 * Removed support for Ultrix, Netware, and OS/2.
2656 * Add support for HKDF.
2658 *Alessandro Ghedini*
2660 * Add support for blake2b and blake2s
2664 * Added support for "pipelining". Ciphers that have the
2665 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2666 encryptions/decryptions simultaneously. There are currently no built-in
2667 ciphers with this property but the expectation is that engines will be able
2668 to offer it to significantly improve throughput. Support has been extended
2669 into libssl so that multiple records for a single connection can be
2670 processed in one go (for >=TLS 1.1).
2674 * Added the AFALG engine. This is an async capable engine which is able to
2675 offload work to the Linux kernel. In this initial version it only supports
2676 AES128-CBC. The kernel must be version 4.1.0 or greater.
2680 * OpenSSL now uses a new threading API. It is no longer necessary to
2681 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2682 are two supported threading models: pthreads and windows threads. It is
2683 also possible to configure OpenSSL at compile time for "no-threads". The
2684 old threading API should no longer be used. The functions have been
2685 replaced with "no-op" compatibility macros.
2687 *Alessandro Ghedini, Matt Caswell*
2689 * Modify behavior of ALPN to invoke callback after SNI/servername
2690 callback, such that updates to the SSL_CTX affect ALPN.
2694 * Add SSL_CIPHER queries for authentication and key-exchange.
2698 * Changes to the DEFAULT cipherlist:
2699 - Prefer (EC)DHE handshakes over plain RSA.
2700 - Prefer AEAD ciphers over legacy ciphers.
2701 - Prefer ECDSA over RSA when both certificates are available.
2702 - Prefer TLSv1.2 ciphers/PRF.
2703 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2708 * Change the ECC default curve list to be this, in order: x25519,
2709 secp256r1, secp521r1, secp384r1.
2713 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2714 disabled by default. They can be re-enabled using the
2715 enable-weak-ssl-ciphers option to Configure.
2719 * If the server has ALPN configured, but supports no protocols that the
2720 client advertises, send a fatal "no_application_protocol" alert.
2721 This behaviour is SHALL in RFC 7301, though it isn't universally
2722 implemented by other servers.
2726 * Add X25519 support.
2727 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2728 for public and private key encoding using the format documented in
2729 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2730 key generation and key derivation.
2732 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2737 * Deprecate SRP_VBASE_get_by_user.
2738 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2739 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2740 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2741 seed, even if the seed is configured.
2743 Users should use SRP_VBASE_get1_by_user instead. Note that in
2744 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2745 also that even though configuring the SRP seed attempts to hide
2746 invalid usernames by continuing the handshake with fake
2747 credentials, this behaviour is not constant time and no strong
2748 guarantees are made that the handshake is indistinguishable from
2749 that of a valid user.
2753 * Configuration change; it's now possible to build dynamic engines
2754 without having to build shared libraries and vice versa. This
2755 only applies to the engines in engines/, those in crypto/engine/
2756 will always be built into libcrypto (i.e. "static").
2758 Building dynamic engines is enabled by default; to disable, use
2759 the configuration option "disable-dynamic-engine".
2761 The only requirements for building dynamic engines are the
2762 presence of the DSO module and building with position independent
2763 code, so they will also automatically be disabled if configuring
2764 with "disable-dso" or "disable-pic".
2766 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2767 are also taken away from openssl/opensslconf.h, as they are
2772 * Configuration change; if there is a known flag to compile
2773 position independent code, it will always be applied on the
2774 libcrypto and libssl object files, and never on the application
2775 object files. This means other libraries that use routines from
2776 libcrypto / libssl can be made into shared libraries regardless
2777 of how OpenSSL was configured.
2779 If this isn't desirable, the configuration options "disable-pic"
2780 or "no-pic" can be used to disable the use of PIC. This will
2781 also disable building shared libraries and dynamic engines.
2785 * Removed JPAKE code. It was experimental and has no wide use.
2789 * The INSTALL_PREFIX Makefile variable has been renamed to
2790 DESTDIR. That makes for less confusion on what this variable
2791 is for. Also, the configuration option --install_prefix is
2796 * Heartbeat for TLS has been removed and is disabled by default
2797 for DTLS; configure with enable-heartbeats. Code that uses the
2798 old #define's might need to be updated.
2800 *Emilia Käsper, Rich Salz*
2802 * Rename REF_CHECK to REF_DEBUG.
2806 * New "unified" build system
2808 The "unified" build system is aimed to be a common system for all
2809 platforms we support. With it comes new support for VMS.
2811 This system builds supports building in a different directory tree
2812 than the source tree. It produces one Makefile (for unix family
2813 or lookalikes), or one descrip.mms (for VMS).
2815 The source of information to make the Makefile / descrip.mms is
2816 small files called 'build.info', holding the necessary
2817 information for each directory with source to compile, and a
2818 template in Configurations, like unix-Makefile.tmpl or
2821 With this change, the library names were also renamed on Windows
2822 and on VMS. They now have names that are closer to the standard
2823 on Unix, and include the major version number, and in certain
2824 cases, the architecture they are built for. See "Notes on shared
2825 libraries" in INSTALL.
2827 We rely heavily on the perl module Text::Template.
2831 * Added support for auto-initialisation and de-initialisation of the library.
2832 OpenSSL no longer requires explicit init or deinit routines to be called,
2833 except in certain circumstances. See the OPENSSL_init_crypto() and
2834 OPENSSL_init_ssl() man pages for further information.
2838 * The arguments to the DTLSv1_listen function have changed. Specifically the
2839 "peer" argument is now expected to be a BIO_ADDR object.
2841 * Rewrite of BIO networking library. The BIO library lacked consistent
2842 support of IPv6, and adding it required some more extensive
2843 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2844 which hold all types of addresses and chains of address information.
2845 It also introduces a new API, with functions like BIO_socket,
2846 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2847 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2848 have been adapted accordingly.
2852 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2857 * CRIME protection: disable compression by default, even if OpenSSL is
2858 compiled with zlib enabled. Applications can still enable compression
2859 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2860 using the SSL_CONF library to configure compression.
2864 * The signature of the session callback configured with
2865 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2866 was explicitly marked as `const unsigned char*` instead of
2871 * Always DPURIFY. Remove the use of uninitialized memory in the
2872 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2876 * Removed many obsolete configuration items, including
2877 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2878 MD2_CHAR, MD2_INT, MD2_LONG
2880 IDEA_SHORT, IDEA_LONG
2881 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2883 *Rich Salz, with advice from Andy Polyakov*
2885 * Many BN internals have been moved to an internal header file.
2887 *Rich Salz with help from Andy Polyakov*
2889 * Configuration and writing out the results from it has changed.
2890 Files such as Makefile include/openssl/opensslconf.h and are now
2891 produced through general templates, such as Makefile.in and
2892 crypto/opensslconf.h.in and some help from the perl module
2895 Also, the center of configuration information is no longer
2896 Makefile. Instead, Configure produces a perl module in
2897 configdata.pm which holds most of the config data (in the hash
2898 table %config), the target data that comes from the target
2899 configuration in one of the `Configurations/*.conf~ files (in
2904 * To clarify their intended purposes, the Configure options
2905 --prefix and --openssldir change their semantics, and become more
2906 straightforward and less interdependent.
2908 --prefix shall be used exclusively to give the location INSTALLTOP
2909 where programs, scripts, libraries, include files and manuals are
2910 going to be installed. The default is now /usr/local.
2912 --openssldir shall be used exclusively to give the default
2913 location OPENSSLDIR where certificates, private keys, CRLs are
2914 managed. This is also where the default openssl.cnf gets
2916 If the directory given with this option is a relative path, the
2917 values of both the --prefix value and the --openssldir value will
2918 be combined to become OPENSSLDIR.
2919 The default for --openssldir is INSTALLTOP/ssl.
2921 Anyone who uses --openssldir to specify where OpenSSL is to be
2922 installed MUST change to use --prefix instead.
2926 * The GOST engine was out of date and therefore it has been removed. An up
2927 to date GOST engine is now being maintained in an external repository.
2928 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2929 support for GOST ciphersuites (these are only activated if a GOST engine
2934 * EGD is no longer supported by default; use enable-egd when
2937 *Ben Kaduk and Rich Salz*
2939 * The distribution now has Makefile.in files, which are used to
2940 create Makefile's when Configure is run. *Configure must be run
2941 before trying to build now.*
2945 * The return value for SSL_CIPHER_description() for error conditions
2950 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2952 Obtaining and performing DNSSEC validation of TLSA records is
2953 the application's responsibility. The application provides
2954 the TLSA records of its choice to OpenSSL, and these are then
2955 used to authenticate the peer.
2957 The TLSA records need not even come from DNS. They can, for
2958 example, be used to implement local end-entity certificate or
2959 trust-anchor "pinning", where the "pin" data takes the form
2960 of TLSA records, which can augment or replace verification
2961 based on the usual WebPKI public certification authorities.
2965 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2966 continues to support deprecated interfaces in default builds.
2967 However, applications are strongly advised to compile their
2968 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2969 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2970 or the 1.1.0 releases.
2972 In environments in which all applications have been ported to
2973 not use any deprecated interfaces OpenSSL's Configure script
2974 should be used with the --api=1.1.0 option to entirely remove
2975 support for the deprecated features from the library and
2976 unconditionally disable them in the installed headers.
2977 Essentially the same effect can be achieved with the "no-deprecated"
2978 argument to Configure, except that this will always restrict
2979 the build to just the latest API, rather than a fixed API
2982 As applications are ported to future revisions of the API,
2983 they should update their compile-time OPENSSL_API_COMPAT define
2984 accordingly, but in most cases should be able to continue to
2985 compile with later releases.
2987 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2988 0x10000000L and 0x00908000L, respectively. However those
2989 versions did not support the OPENSSL_API_COMPAT feature, and
2990 so applications are not typically tested for explicit support
2991 of just the undeprecated features of either release.
2995 * Add support for setting the minimum and maximum supported protocol.
2996 It can bet set via the SSL_set_min_proto_version() and
2997 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2998 MaxProtocol. It's recommended to use the new APIs to disable
2999 protocols instead of disabling individual protocols using
3000 SSL_set_options() or SSL_CONF's Protocol. This change also
3001 removes support for disabling TLS 1.2 in the OpenSSL TLS
3002 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3006 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3010 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3011 and integrates ECDSA and ECDH functionality into EC. Implementations can
3012 now redirect key generation and no longer need to convert to or from
3015 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3016 include the ec.h header file instead.
3020 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3021 ciphers who are no longer supported and drops support the ephemeral RSA key
3022 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3026 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3027 opaque. For HMAC_CTX, the following constructors and destructors
3030 HMAC_CTX *HMAC_CTX_new(void);
3031 void HMAC_CTX_free(HMAC_CTX *ctx);
3033 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3034 destroy such methods has been added. See EVP_MD_meth_new(3) and
3035 EVP_CIPHER_meth_new(3) for documentation.
3038 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
3039 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
3040 EVP_MD_CTX_reset() should be called instead to reinitialise
3041 an already created structure.
3042 2) For consistency with the majority of our object creators and
3043 destructors, EVP_MD_CTX_(create|destroy) were renamed to
3044 EVP_MD_CTX_(new|free). The old names are retained as macros
3045 for deprecated builds.
3049 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3050 cryptographic operations to be performed asynchronously as long as an
3051 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3052 further details. Libssl has also had this capability integrated with the
3053 introduction of the new mode SSL_MODE_ASYNC and associated error
3054 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3055 pages. This work was developed in partnership with Intel Corp.
3059 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3060 always enabled now. If you want to disable the support you should
3061 exclude it using the list of supported ciphers. This also means that the
3062 "-no_ecdhe" option has been removed from s_server.
3066 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3067 SSL_{CTX_}set1_curves() which can set a list.
3071 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3072 curve you want to support using SSL_{CTX_}set1_curves().
3076 * State machine rewrite. The state machine code has been significantly
3077 refactored in order to remove much duplication of code and solve issues
3078 with the old code (see ssl/statem/README for further details). This change
3079 does have some associated API changes. Notably the SSL_state() function
3080 has been removed and replaced by SSL_get_state which now returns an
3081 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
3082 altogether. The previous handshake states defined in ssl.h and ssl3.h have
3087 * All instances of the string "ssleay" in the public API were replaced
3088 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3089 Some error codes related to internal RSA_eay API's were renamed.
3093 * The demo files in crypto/threads were moved to demo/threads.
3097 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3100 *Matt Caswell, Rich Salz*
3102 * New ASN.1 embed macro.
3104 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3105 structure is not allocated: it is part of the parent. That is instead of
3113 This reduces memory fragmentation and make it impossible to accidentally
3114 set a mandatory field to NULL.
3116 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3117 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3118 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3123 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3127 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3128 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3129 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3130 DES and RC4 ciphersuites.
3134 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3135 This changes the decoding behaviour for some invalid messages,
3136 though the change is mostly in the more lenient direction, and
3137 legacy behaviour is preserved as much as possible.
3141 * Fix no-stdio build.
3142 * David Woodhouse <David.Woodhouse@intel.com> and also
3143 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3145 * New testing framework
3146 The testing framework has been largely rewritten and is now using
3147 perl and the perl modules Test::Harness and an extended variant of
3148 Test::More called OpenSSL::Test to do its work. All test scripts in
3149 test/ have been rewritten into test recipes, and all direct calls to
3150 executables in test/Makefile have become individual recipes using the
3151 simplified testing OpenSSL::Test::Simple.
3153 For documentation on our testing modules, do:
3155 perldoc test/testlib/OpenSSL/Test/Simple.pm
3156 perldoc test/testlib/OpenSSL/Test.pm
3160 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3161 are used; the latter aborts on memory leaks (usually checked on exit).
3162 Some undocumented "set malloc, etc., hooks" functions were removed
3163 and others were changed. All are now documented.
3167 * In DSA_generate_parameters_ex, if the provided seed is too short,
3170 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3172 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3173 from RFC4279, RFC4785, RFC5487, RFC5489.
3175 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3176 original RSA_PSK patch.
3180 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3181 era flag was never set throughout the codebase (only read). Also removed
3182 SSL3_FLAGS_POP_BUFFER which was only used if
3183 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3187 * Changed the default name options in the "ca", "crl", "req" and "x509"
3188 to be "oneline" instead of "compat".
3192 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3193 not aware of clients that still exhibit this bug, and the workaround
3194 hasn't been working properly for a while.
3198 * The return type of BIO_number_read() and BIO_number_written() as well as
3199 the corresponding num_read and num_write members in the BIO structure has
3200 changed from unsigned long to uint64_t. On platforms where an unsigned
3201 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3206 * Given the pervasive nature of TLS extensions it is inadvisable to run
3207 OpenSSL without support for them. It also means that maintaining
3208 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3209 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3213 * Removed support for the two export grade static DH ciphersuites
3214 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3215 were newly added (along with a number of other static DH ciphersuites) to
3216 1.0.2. However the two export ones have *never* worked since they were
3217 introduced. It seems strange in any case to be adding new export
3218 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3222 * Version negotiation has been rewritten. In particular SSLv23_method(),
3223 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3224 and turned into macros which simply call the new preferred function names
3225 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3226 should use the new names instead. Also as part of this change the ssl23.h
3227 header file has been removed.
3231 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3232 code and the associated standard is no longer considered fit-for-purpose.
3236 * RT2547 was closed. When generating a private key, try to make the
3237 output file readable only by the owner. This behavior change might
3238 be noticeable when interacting with other software.
3240 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3245 * Added HTTP GET support to the ocsp command.
3249 * Changed default digest for the dgst and enc commands from MD5 to
3254 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3258 * Added support for TLS extended master secret from
3259 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3260 initial patch which was a great help during development.
3264 * All libssl internal structures have been removed from the public header
3265 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3266 now redundant). Users should not attempt to access internal structures
3267 directly. Instead they should use the provided API functions.
3271 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3272 Access to deprecated functions can be re-enabled by running config with
3273 "enable-deprecated". In addition applications wishing to use deprecated
3274 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3275 will, by default, disable some transitive includes that previously existed
3276 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3280 * Added support for OCB mode. OpenSSL has been granted a patent license
3281 compatible with the OpenSSL license for use of OCB. Details are available
3282 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3283 for OCB can be removed by calling config with no-ocb.
3287 * SSLv2 support has been removed. It still supports receiving a SSLv2
3288 compatible client hello.
3292 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3293 done while fixing the error code for the key-too-small case.
3295 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3297 * CA.sh has been removed; use CA.pl instead.
3301 * Removed old DES API.
3305 * Remove various unsupported platforms:
3311 Sinix/ReliantUNIX RM400
3316 16-bit platforms such as WIN16
3320 * Clean up OPENSSL_NO_xxx #define's
3321 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3322 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3323 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3324 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3325 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3326 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3327 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3328 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3329 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3330 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3334 * Cleaned up dead code
3335 Remove all but one '#ifdef undef' which is to be looked at.
3339 * Clean up calling of xxx_free routines.
3340 Just like free(), fix most of the xxx_free routines to accept
3341 NULL. Remove the non-null checks from callers. Save much code.
3345 * Add secure heap for storage of private keys (when possible).
3346 Add BIO_s_secmem(), CBIGNUM, etc.
3347 Contributed by Akamai Technologies under our Corporate CLA.
3351 * Experimental support for a new, fast, unbiased prime candidate generator,
3352 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3354 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3356 * New output format NSS in the sess_id command line tool. This allows
3357 exporting the session id and the master key in NSS keylog format.
3359 *Martin Kaiser <martin@kaiser.cx>*
3361 * Harmonize version and its documentation. -f flag is used to display
3364 *mancha <mancha1@zoho.com>*
3366 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3367 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3369 *mancha <mancha1@zoho.com>*
3371 * Fix some double frees. These are not thought to be exploitable.
3373 *mancha <mancha1@zoho.com>*
3375 * A missing bounds check in the handling of the TLS heartbeat extension
3376 can be used to reveal up to 64k of memory to a connected client or
3379 Thanks for Neel Mehta of Google Security for discovering this bug and to
3380 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3381 preparing the fix [CVE-2014-0160][]
3383 *Adam Langley, Bodo Moeller*
3385 * Fix for the attack described in the paper "Recovering OpenSSL
3386 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3387 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3388 <http://eprint.iacr.org/2014/140>
3390 Thanks to Yuval Yarom and Naomi Benger for discovering this
3391 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3393 *Yuval Yarom and Naomi Benger*
3395 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3396 this fixes a limitation in previous versions of OpenSSL.
3400 * Experimental encrypt-then-mac support.
3402 Experimental support for encrypt then mac from
3403 draft-gutmann-tls-encrypt-then-mac-02.txt
3405 To enable it set the appropriate extension number (0x42 for the test
3406 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3408 For non-compliant peers (i.e. just about everything) this should have no
3411 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3415 * Add EVP support for key wrapping algorithms, to avoid problems with
3416 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3417 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3418 algorithms and include tests cases.
3422 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3427 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3428 MGF1 digest and OAEP label.
3432 * Make openssl verify return errors.
3434 *Chris Palmer <palmer@google.com> and Ben Laurie*
3436 * New function ASN1_TIME_diff to calculate the difference between two
3437 ASN1_TIME structures or one structure and the current time.
3441 * Update fips_test_suite to support multiple command line options. New
3442 test to induce all self test errors in sequence and check expected
3447 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3448 sign or verify all in one operation.
3452 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3453 test programs and fips_test_suite. Includes functionality to parse
3454 the minimal script output of fipsalgest.pl directly.
3458 * Add authorisation parameter to FIPS_module_mode_set().
3462 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3466 * Use separate DRBG fields for internal and external flags. New function
3467 FIPS_drbg_health_check() to perform on demand health checking. Add
3468 generation tests to fips_test_suite with reduced health check interval to
3469 demonstrate periodic health checking. Add "nodh" option to
3470 fips_test_suite to skip very slow DH test.
3474 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3479 * More extensive health check for DRBG checking many more failure modes.
3480 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3481 combination: call this in fips_test_suite.
3485 * Add support for canonical generation of DSA parameter 'g'. See
3488 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3489 POST to handle HMAC cases.
3493 * Add functions FIPS_module_version() and FIPS_module_version_text()
3494 to return numerical and string versions of the FIPS module number.
3498 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3499 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3500 outside the validated module in the FIPS capable OpenSSL.
3504 * Minor change to DRBG entropy callback semantics. In some cases
3505 there is no multiple of the block length between min_len and
3506 max_len. Allow the callback to return more than max_len bytes
3507 of entropy but discard any extra: it is the callback's responsibility
3508 to ensure that the extra data discarded does not impact the
3509 requested amount of entropy.
3513 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3514 information in FIPS186-3, SP800-57 and SP800-131A.
3518 * CCM support via EVP. Interface is very similar to GCM case except we
3519 must supply all data in one chunk (i.e. no update, final) and the
3520 message length must be supplied if AAD is used. Add algorithm test
3525 * Initial version of POST overhaul. Add POST callback to allow the status
3526 of POST to be monitored and/or failures induced. Modify fips_test_suite
3527 to use callback. Always run all selftests even if one fails.
3531 * XTS support including algorithm test driver in the fips_gcmtest program.
3532 Note: this does increase the maximum key length from 32 to 64 bytes but
3533 there should be no binary compatibility issues as existing applications
3534 will never use XTS mode.
3538 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3539 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3540 performs algorithm blocking for unapproved PRNG types. Also do not
3541 set PRNG type in FIPS_mode_set(): leave this to the application.
3542 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3543 the standard OpenSSL PRNG: set additional data to a date time vector.
3547 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3548 This shouldn't present any incompatibility problems because applications
3549 shouldn't be using these directly and any that are will need to rethink
3550 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3554 * Extensive self tests and health checking required by SP800-90 DRBG.
3555 Remove strength parameter from FIPS_drbg_instantiate and always
3556 instantiate at maximum supported strength.
3560 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3564 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3568 * New function DH_compute_key_padded() to compute a DH key and pad with
3569 leading zeroes if needed: this complies with SP800-56A et al.
3573 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3574 anything, incomplete, subject to change and largely untested at present.
3578 * Modify fipscanisteronly build option to only build the necessary object
3579 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3583 * Add experimental option FIPSSYMS to give all symbols in
3584 fipscanister.o and FIPS or fips prefix. This will avoid
3585 conflicts with future versions of OpenSSL. Add perl script
3586 util/fipsas.pl to preprocess assembly language source files
3587 and rename any affected symbols.
3591 * Add selftest checks and algorithm block of non-fips algorithms in
3592 FIPS mode. Remove DES2 from selftests.
3596 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3597 return internal method without any ENGINE dependencies. Add new
3598 tiny fips sign and verify functions.
3602 * New build option no-ec2m to disable characteristic 2 code.
3606 * New build option "fipscanisteronly". This only builds fipscanister.o
3607 and (currently) associated fips utilities. Uses the file Makefile.fips
3608 instead of Makefile.org as the prototype.
3612 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3613 Update fips_gcmtest to use IV generator.
3617 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3618 setting output buffer to NULL. The `*Final` function must be
3619 called although it will not retrieve any additional data. The tag
3620 can be set or retrieved with a ctrl. The IV length is by default 12
3621 bytes (96 bits) but can be set to an alternative value. If the IV
3622 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3627 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3628 underlying do_cipher function handles all cipher semantics itself
3629 including padding and finalisation. This is useful if (for example)
3630 an ENGINE cipher handles block padding itself. The behaviour of
3631 do_cipher is subtly changed if this flag is set: the return value
3632 is the number of characters written to the output buffer (zero is
3633 no longer an error code) or a negative error code. Also if the
3634 input buffer is NULL and length 0 finalisation should be performed.
3638 * If a candidate issuer certificate is already part of the constructed
3639 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3643 * Improve forward-security support: add functions
3645 void SSL_CTX_set_not_resumable_session_callback(
3646 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3647 void SSL_set_not_resumable_session_callback(
3648 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3650 for use by SSL/TLS servers; the callback function will be called whenever a
3651 new session is created, and gets to decide whether the session may be
3652 cached to make it resumable (return 0) or not (return 1). (As by the
3653 SSL/TLS protocol specifications, the session_id sent by the server will be
3654 empty to indicate that the session is not resumable; also, the server will
3655 not generate RFC 4507 (RFC 5077) session tickets.)
3657 A simple reasonable callback implementation is to return is_forward_secure.
3658 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3659 by the SSL/TLS server library, indicating whether it can provide forward
3662 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3664 * New -verify_name option in command line utilities to set verification
3669 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3670 Add CMAC pkey methods.
3674 * Experimental renegotiation in s_server -www mode. If the client
3675 browses /reneg connection is renegotiated. If /renegcert it is
3676 renegotiated requesting a certificate.
3680 * Add an "external" session cache for debugging purposes to s_server. This
3681 should help trace issues which normally are only apparent in deployed
3682 multi-process servers.
3686 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3687 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3688 BIO_set_cipher() and some obscure PEM functions were changed so they
3689 can now return an error. The RAND changes required a change to the
3690 RAND_METHOD structure.
3694 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3695 a gcc attribute to warn if the result of a function is ignored. This
3696 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3697 whose return value is often ignored.
3701 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3702 These allow SCTs (signed certificate timestamps) to be requested and
3703 validated when establishing a connection.
3705 *Rob Percival <robpercival@google.com>*
3710 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3712 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3713 used even when parsing explicit parameters, when loading a serialized key
3714 or calling `EC_GROUP_new_from_ecpkparameters()`/
3715 `EC_GROUP_new_from_ecparameters()`.
3716 This prevents bypass of security hardening and performance gains,
3717 especially for curves with specialized EC_METHODs.
3718 By default, if a key encoded with explicit parameters is loaded and later
3719 serialized, the output is still encoded with explicit parameters, even if
3720 internally a "named" EC_GROUP is used for computation.
3724 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3725 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3726 NULL. After this change, only the cofactor parameter can be NULL. It also
3727 does some minimal sanity checks on the passed order.
3732 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3733 An attack is simple, if the first CMS_recipientInfo is valid but the
3734 second CMS_recipientInfo is chosen ciphertext. If the second
3735 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3736 encryption key will be replaced by garbage, and the message cannot be
3737 decoded, but if the RSA decryption fails, the correct encryption key is
3738 used and the recipient will not notice the attack.
3739 As a work around for this potential attack the length of the decrypted
3740 key must be equal to the cipher default key length, in case the
3741 certifiate is not given and all recipientInfo are tried out.
3742 The old behaviour can be re-enabled in the CMS code by setting the
3743 CMS_DEBUG_DECRYPT flag.
3748 * Document issue with installation paths in diverse Windows builds
3750 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3751 binaries and run-time config file.
3756 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3758 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3759 This changes the size when using the genpkey app when no size is given. It
3760 fixes an omission in earlier changes that changed all RSA, DSA and DH
3761 generation apps to use 2048 bits by default.
3765 * Add FIPS support for Android Arm 64-bit
3767 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3768 Module in Version 2.0.10. For some reason, the corresponding target
3769 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3770 built with FIPS support on Android Arm 64-bit. This omission has been
3773 *Matthias St. Pierre*
3775 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3777 * 0-byte record padding oracle
3779 If an application encounters a fatal protocol error and then calls
3780 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3781 then OpenSSL can respond differently to the calling application if a 0 byte
3782 record is received with invalid padding compared to if a 0 byte record is
3783 received with an invalid MAC. If the application then behaves differently
3784 based on that in a way that is detectable to the remote peer, then this
3785 amounts to a padding oracle that could be used to decrypt data.
3787 In order for this to be exploitable "non-stitched" ciphersuites must be in
3788 use. Stitched ciphersuites are optimised implementations of certain
3789 commonly used ciphersuites. Also the application must call SSL_shutdown()
3790 twice even if a protocol error has occurred (applications should not do
3791 this but some do anyway).
3793 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3794 Aviram, with additional investigation by Steven Collison and Andrew
3795 Hourselt. It was reported to OpenSSL on 10th December 2018.
3800 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3804 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3806 * Microarchitecture timing vulnerability in ECC scalar multiplication
3808 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3809 shown to be vulnerable to a microarchitecture timing side channel attack.
3810 An attacker with sufficient access to mount local timing attacks during
3811 ECDSA signature generation could recover the private key.
3813 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3814 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3820 * Timing vulnerability in DSA signature generation
3822 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3823 timing side channel attack. An attacker could use variations in the signing
3824 algorithm to recover the private key.
3826 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3831 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3832 Module, accidentally introduced while backporting security fixes from the
3833 development branch and hindering the use of ECC in FIPS mode.
3837 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3839 * Client DoS due to large DH parameter
3841 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3842 malicious server can send a very large prime value to the client. This will
3843 cause the client to spend an unreasonably long period of time generating a
3844 key for this prime resulting in a hang until the client has finished. This
3845 could be exploited in a Denial Of Service attack.
3847 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3852 * Cache timing vulnerability in RSA Key Generation
3854 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3855 a cache timing side channel attack. An attacker with sufficient access to
3856 mount cache timing attacks during the RSA key generation process could
3857 recover the private key.
3859 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3860 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3865 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3866 parameter is no longer accepted, as it leads to a corrupt table. NULL
3867 pem_str is reserved for alias entries only.
3871 * Revert blinding in ECDSA sign and instead make problematic addition
3872 length-invariant. Switch even to fixed-length Montgomery multiplication.
3876 * Change generating and checking of primes so that the error rate of not
3877 being prime depends on the intended use based on the size of the input.
3878 For larger primes this will result in more rounds of Miller-Rabin.
3879 The maximal error rate for primes with more than 1080 bits is lowered
3882 *Kurt Roeckx, Annie Yousar*
3884 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3888 * Add blinding to ECDSA and DSA signatures to protect against side channel
3889 attacks discovered by Keegan Ryan (NCC Group).
3893 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3894 now allow empty (zero character) pass phrases.
3898 * Certificate time validation (X509_cmp_time) enforces stricter
3899 compliance with RFC 5280. Fractional seconds and timezone offsets
3900 are no longer allowed.
3904 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3906 * Constructed ASN.1 types with a recursive definition could exceed the stack
3908 Constructed ASN.1 types with a recursive definition (such as can be found
3909 in PKCS7) could eventually exceed the stack given malicious input with
3910 excessive recursion. This could result in a Denial Of Service attack. There
3911 are no such structures used within SSL/TLS that come from untrusted sources
3912 so this is considered safe.
3914 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3920 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3922 * Read/write after SSL object in error state
3924 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3925 mechanism. The intent was that if a fatal error occurred during a handshake
3926 then OpenSSL would move into the error state and would immediately fail if
3927 you attempted to continue the handshake. This works as designed for the
3928 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3929 SSL_connect()), however due to a bug it does not work correctly if
3930 SSL_read() or SSL_write() is called directly. In that scenario, if the
3931 handshake fails then a fatal error will be returned in the initial function
3932 call. If SSL_read()/SSL_write() is subsequently called by the application
3933 for the same SSL object then it will succeed and the data is passed without
3934 being decrypted/encrypted directly from the SSL/TLS record layer.
3936 In order to exploit this issue an application bug would have to be present
3937 that resulted in a call to SSL_read()/SSL_write() being issued after having
3938 already received a fatal error.
3940 This issue was reported to OpenSSL by David Benjamin (Google).
3945 * rsaz_1024_mul_avx2 overflow bug on x86_64
3947 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3948 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3949 Analysis suggests that attacks against RSA and DSA as a result of this
3950 defect would be very difficult to perform and are not believed likely.
3951 Attacks against DH1024 are considered just feasible, because most of the
3952 work necessary to deduce information about a private key may be performed
3953 offline. The amount of resources required for such an attack would be
3954 significant. However, for an attack on TLS to be meaningful, the server
3955 would have to share the DH1024 private key among multiple clients, which is
3956 no longer an option since CVE-2016-0701.
3958 This only affects processors that support the AVX2 but not ADX extensions
3959 like Intel Haswell (4th generation).
3961 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3962 was originally found via the OSS-Fuzz project.
3967 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
3969 * bn_sqrx8x_internal carry bug on x86_64
3971 There is a carry propagating bug in the x86_64 Montgomery squaring
3972 procedure. No EC algorithms are affected. Analysis suggests that attacks
3973 against RSA and DSA as a result of this defect would be very difficult to
3974 perform and are not believed likely. Attacks against DH are considered just
3975 feasible (although very difficult) because most of the work necessary to
3976 deduce information about a private key may be performed offline. The amount
3977 of resources required for such an attack would be very significant and
3978 likely only accessible to a limited number of attackers. An attacker would
3979 additionally need online access to an unpatched system using the target
3980 private key in a scenario with persistent DH parameters and a private
3981 key that is shared between multiple clients.
3983 This only affects processors that support the BMI1, BMI2 and ADX extensions
3984 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3986 This issue was reported to OpenSSL by the OSS-Fuzz project.
3991 * Malformed X.509 IPAddressFamily could cause OOB read
3993 If an X.509 certificate has a malformed IPAddressFamily extension,
3994 OpenSSL could do a one-byte buffer overread. The most likely result
3995 would be an erroneous display of the certificate in text format.
3997 This issue was reported to OpenSSL by the OSS-Fuzz project.
4002 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4004 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4005 platform rather than 'mingw'.
4009 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4011 * Truncated packet could crash via OOB read
4013 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4014 cipher is being used, then a truncated packet can cause that host to
4015 perform an out-of-bounds read, usually resulting in a crash.
4017 This issue was reported to OpenSSL by Robert Święcki of Google.
4022 * BN_mod_exp may produce incorrect results on x86_64
4024 There is a carry propagating bug in the x86_64 Montgomery squaring
4025 procedure. No EC algorithms are affected. Analysis suggests that attacks
4026 against RSA and DSA as a result of this defect would be very difficult to
4027 perform and are not believed likely. Attacks against DH are considered just
4028 feasible (although very difficult) because most of the work necessary to
4029 deduce information about a private key may be performed offline. The amount
4030 of resources required for such an attack would be very significant and
4031 likely only accessible to a limited number of attackers. An attacker would
4032 additionally need online access to an unpatched system using the target
4033 private key in a scenario with persistent DH parameters and a private
4034 key that is shared between multiple clients. For example this can occur by
4035 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4036 similar to CVE-2015-3193 but must be treated as a separate problem.
4038 This issue was reported to OpenSSL by the OSS-Fuzz project.
4043 * Montgomery multiplication may produce incorrect results
4045 There is a carry propagating bug in the Broadwell-specific Montgomery
4046 multiplication procedure that handles input lengths divisible by, but
4047 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4048 and DH private keys are impossible. This is because the subroutine in
4049 question is not used in operations with the private key itself and an input
4050 of the attacker's direct choice. Otherwise the bug can manifest itself as
4051 transient authentication and key negotiation failures or reproducible
4052 erroneous outcome of public-key operations with specially crafted input.
4053 Among EC algorithms only Brainpool P-512 curves are affected and one
4054 presumably can attack ECDH key negotiation. Impact was not analyzed in
4055 detail, because pre-requisites for attack are considered unlikely. Namely
4056 multiple clients have to choose the curve in question and the server has to
4057 share the private key among them, neither of which is default behaviour.
4058 Even then only clients that chose the curve will be affected.
4060 This issue was publicly reported as transient failures and was not
4061 initially recognized as a security issue. Thanks to Richard Morgan for
4062 providing reproducible case.
4067 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4068 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4069 prevent issues where no progress is being made and the peer continually
4070 sends unrecognised record types, using up resources processing them.
4074 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4076 * Missing CRL sanity check
4078 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4079 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4080 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4082 This issue only affects the OpenSSL 1.0.2i
4087 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4089 * OCSP Status Request extension unbounded memory growth
4091 A malicious client can send an excessively large OCSP Status Request
4092 extension. If that client continually requests renegotiation, sending a
4093 large OCSP Status Request extension each time, then there will be unbounded
4094 memory growth on the server. This will eventually lead to a Denial Of
4095 Service attack through memory exhaustion. Servers with a default
4096 configuration are vulnerable even if they do not support OCSP. Builds using
4097 the "no-ocsp" build time option are not affected.
4099 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4104 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4107 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4113 * OOB write in MDC2_Update()
4115 An overflow can occur in MDC2_Update() either if called directly or
4116 through the EVP_DigestUpdate() function using MDC2. If an attacker
4117 is able to supply very large amounts of input data after a previous
4118 call to EVP_EncryptUpdate() with a partial block then a length check
4119 can overflow resulting in a heap corruption.
4121 The amount of data needed is comparable to SIZE_MAX which is impractical
4124 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4129 * Malformed SHA512 ticket DoS
4131 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4132 DoS attack where a malformed ticket will result in an OOB read which will
4135 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4136 a custom server callback and ticket lookup mechanism.
4138 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4143 * OOB write in BN_bn2dec()
4145 The function BN_bn2dec() does not check the return value of BN_div_word().
4146 This can cause an OOB write if an application uses this function with an
4147 overly large BIGNUM. This could be a problem if an overly large certificate
4148 or CRL is printed out from an untrusted source. TLS is not affected because
4149 record limits will reject an oversized certificate before it is parsed.
4151 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4156 * OOB read in TS_OBJ_print_bio()
4158 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4159 the total length the OID text representation would use and not the amount
4160 of data written. This will result in OOB reads when large OIDs are
4163 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4168 * Pointer arithmetic undefined behaviour
4170 Avoid some undefined pointer arithmetic
4172 A common idiom in the codebase is to check limits in the following manner:
4175 Where "p" points to some malloc'd data of SIZE bytes and
4178 "len" here could be from some externally supplied data (e.g. from a TLS
4181 The rules of C pointer arithmetic are such that "p + len" is only well
4182 defined where len <= SIZE. Therefore the above idiom is actually
4183 undefined behaviour.
4185 For example this could cause problems if some malloc implementation
4186 provides an address for "p" such that "p + len" actually overflows for
4187 values of len that are too big and therefore p + len < limit.
4189 This issue was reported to OpenSSL by Guido Vranken
4194 * Constant time flag not preserved in DSA signing
4196 Operations in the DSA signing algorithm should run in constant time in
4197 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4198 implementation means that a non-constant time codepath is followed for
4199 certain operations. This has been demonstrated through a cache-timing
4200 attack to be sufficient for an attacker to recover the private DSA key.
4202 This issue was reported by César Pereida (Aalto University), Billy Brumley
4203 (Tampere University of Technology), and Yuval Yarom (The University of
4204 Adelaide and NICTA).
4209 * DTLS buffered message DoS
4211 In a DTLS connection where handshake messages are delivered out-of-order
4212 those messages that OpenSSL is not yet ready to process will be buffered
4213 for later use. Under certain circumstances, a flaw in the logic means that
4214 those messages do not get removed from the buffer even though the handshake
4215 has been completed. An attacker could force up to approx. 15 messages to
4216 remain in the buffer when they are no longer required. These messages will
4217 be cleared when the DTLS connection is closed. The default maximum size for
4218 a message is 100k. Therefore the attacker could force an additional 1500k
4219 to be consumed per connection. By opening many simulataneous connections an
4220 attacker could cause a DoS attack through memory exhaustion.
4222 This issue was reported to OpenSSL by Quan Luo.
4227 * DTLS replay protection DoS
4229 A flaw in the DTLS replay attack protection mechanism means that records
4230 that arrive for future epochs update the replay protection "window" before
4231 the MAC for the record has been validated. This could be exploited by an
4232 attacker by sending a record for the next epoch (which does not have to
4233 decrypt or have a valid MAC), with a very large sequence number. This means
4234 that all subsequent legitimate packets are dropped causing a denial of
4235 service for a specific DTLS connection.
4237 This issue was reported to OpenSSL by the OCAP audit team.
4242 * Certificate message OOB reads
4244 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4245 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4246 theoretical DoS risk but this has not been observed in practice on common
4249 The messages affected are client certificate, client certificate request
4250 and server certificate. As a result the attack can only be performed
4251 against a client or a server which enables client authentication.
4253 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4258 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4260 * Prevent padding oracle in AES-NI CBC MAC check
4262 A MITM attacker can use a padding oracle attack to decrypt traffic
4263 when the connection uses an AES CBC cipher and the server support
4266 This issue was introduced as part of the fix for Lucky 13 padding
4267 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4268 constant time by making sure that always the same bytes are read and
4269 compared against either the MAC or padding bytes. But it no longer
4270 checked that there was enough data to have both the MAC and padding
4273 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4278 * Fix EVP_EncodeUpdate overflow
4280 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4281 Base64 encoding of binary data. If an attacker is able to supply very large
4282 amounts of input data then a length check can overflow resulting in a heap
4285 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4286 the `PEM_write_bio*` family of functions. These are mainly used within the
4287 OpenSSL command line applications, so any application which processes data
4288 from an untrusted source and outputs it as a PEM file should be considered
4289 vulnerable to this issue. User applications that call these APIs directly
4290 with large amounts of untrusted data may also be vulnerable.
4292 This issue was reported by Guido Vranken.
4297 * Fix EVP_EncryptUpdate overflow
4299 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4300 is able to supply very large amounts of input data after a previous call to
4301 EVP_EncryptUpdate() with a partial block then a length check can overflow
4302 resulting in a heap corruption. Following an analysis of all OpenSSL
4303 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4304 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4305 the first called function after an EVP_EncryptInit(), and therefore that
4306 specific call must be safe. The second form is where the length passed to
4307 EVP_EncryptUpdate() can be seen from the code to be some small value and
4308 therefore there is no possibility of an overflow. Since all instances are
4309 one of these two forms, it is believed that there can be no overflows in
4310 internal code due to this problem. It should be noted that
4311 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4312 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4313 of these calls have also been analysed too and it is believed there are no
4314 instances in internal usage where an overflow could occur.
4316 This issue was reported by Guido Vranken.
4321 * Prevent ASN.1 BIO excessive memory allocation
4323 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4324 a short invalid encoding can cause allocation of large amounts of memory
4325 potentially consuming excessive resources or exhausting memory.
4327 Any application parsing untrusted data through d2i BIO functions is
4328 affected. The memory based functions such as d2i_X509() are *not* affected.
4329 Since the memory based functions are used by the TLS library, TLS
4330 applications are not affected.
4332 This issue was reported by Brian Carpenter.
4339 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4340 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4341 in arbitrary stack data being returned in the buffer.
4343 This issue was reported by Guido Vranken.
4348 * Modify behavior of ALPN to invoke callback after SNI/servername
4349 callback, such that updates to the SSL_CTX affect ALPN.
4353 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4358 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4359 methods are enabled and ssl2 is disabled the methods return NULL.
4363 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4365 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4366 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4367 provide any "EXPORT" or "LOW" strength ciphers.
4371 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4372 is by default disabled at build-time. Builds that are not configured with
4373 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4374 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4375 will need to explicitly call either of:
4377 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4379 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4381 as appropriate. Even if either of those is used, or the application
4382 explicitly uses the version-specific SSLv2_method() or its client and
4383 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4384 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4385 ciphers, and SSLv2 56-bit DES are no longer available.
4390 * Fix a double-free in DSA code
4392 A double free bug was discovered when OpenSSL parses malformed DSA private
4393 keys and could lead to a DoS attack or memory corruption for applications
4394 that receive DSA private keys from untrusted sources. This scenario is
4397 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4403 * Disable SRP fake user seed to address a server memory leak.
4405 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4407 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4408 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4409 was changed to ignore the "fake user" SRP seed, even if the seed
4412 Users should use SRP_VBASE_get1_by_user instead. Note that in
4413 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4414 also that even though configuring the SRP seed attempts to hide
4415 invalid usernames by continuing the handshake with fake
4416 credentials, this behaviour is not constant time and no strong
4417 guarantees are made that the handshake is indistinguishable from
4418 that of a valid user.
4423 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4425 In the BN_hex2bn function the number of hex digits is calculated using an
4426 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4427 large values of |i| this can result in |bn_expand| not allocating any
4428 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4429 field as NULL leading to a subsequent NULL ptr deref. For very large values
4430 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4431 In this case memory is allocated to the internal BIGNUM data field, but it
4432 is insufficiently sized leading to heap corruption. A similar issue exists
4433 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4434 is ever called by user applications with very large untrusted hex/dec data.
4435 This is anticipated to be a rare occurrence.
4437 All OpenSSL internal usage of these functions use data that is not expected
4438 to be untrusted, e.g. config file data or application command line
4439 arguments. If user developed applications generate config file data based
4440 on untrusted data then it is possible that this could also lead to security
4441 consequences. This is also anticipated to be rare.
4443 This issue was reported to OpenSSL by Guido Vranken.
4448 * Fix memory issues in `BIO_*printf` functions
4450 The internal |fmtstr| function used in processing a "%s" format string in
4451 the `BIO_*printf` functions could overflow while calculating the length of a
4452 string and cause an OOB read when printing very long strings.
4454 Additionally the internal |doapr_outch| function can attempt to write to an
4455 OOB memory location (at an offset from the NULL pointer) in the event of a
4456 memory allocation failure. In 1.0.2 and below this could be caused where
4457 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4458 could be in processing a very long "%s" format string. Memory leaks can
4461 The first issue may mask the second issue dependent on compiler behaviour.
4462 These problems could enable attacks where large amounts of untrusted data
4463 is passed to the `BIO_*printf` functions. If applications use these functions
4464 in this way then they could be vulnerable. OpenSSL itself uses these
4465 functions when printing out human-readable dumps of ASN.1 data. Therefore
4466 applications that print this data could be vulnerable if the data is from
4467 untrusted sources. OpenSSL command line applications could also be
4468 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4469 as command line arguments.
4471 Libssl is not considered directly vulnerable. Additionally certificates etc
4472 received via remote connections via libssl are also unlikely to be able to
4473 trigger these issues because of message size limits enforced within libssl.
4475 This issue was reported to OpenSSL Guido Vranken.
4480 * Side channel attack on modular exponentiation
4482 A side-channel attack was found which makes use of cache-bank conflicts on
4483 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4484 of RSA keys. The ability to exploit this issue is limited as it relies on
4485 an attacker who has control of code in a thread running on the same
4486 hyper-threaded core as the victim thread which is performing decryptions.
4488 This issue was reported to OpenSSL by Yuval Yarom, The University of
4489 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4490 Nadia Heninger, University of Pennsylvania with more information at
4491 <http://cachebleed.info>.
4496 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4497 if no keysize is specified with default_bits. This fixes an
4498 omission in an earlier change that changed all RSA/DSA key generation
4499 apps to use 2048 bits by default.
4503 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4505 * DH small subgroups
4507 Historically OpenSSL only ever generated DH parameters based on "safe"
4508 primes. More recently (in version 1.0.2) support was provided for
4509 generating X9.42 style parameter files such as those required for RFC 5114
4510 support. The primes used in such files may not be "safe". Where an
4511 application is using DH configured with parameters based on primes that are
4512 not "safe" then an attacker could use this fact to find a peer's private
4513 DH exponent. This attack requires that the attacker complete multiple
4514 handshakes in which the peer uses the same private DH exponent. For example
4515 this could be used to discover a TLS server's private DH exponent if it's
4516 reusing the private DH exponent or it's using a static DH ciphersuite.
4518 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4519 TLS. It is not on by default. If the option is not set then the server
4520 reuses the same private DH exponent for the life of the server process and
4521 would be vulnerable to this attack. It is believed that many popular
4522 applications do set this option and would therefore not be at risk.
4524 The fix for this issue adds an additional check where a "q" parameter is
4525 available (as is the case in X9.42 based parameters). This detects the
4526 only known attack, and is the only possible defense for static DH
4527 ciphersuites. This could have some performance impact.
4529 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4530 default and cannot be disabled. This could have some performance impact.
4532 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4537 * SSLv2 doesn't block disabled ciphers
4539 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4540 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4541 been disabled, provided that the SSLv2 protocol was not also disabled via
4544 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4545 and Sebastian Schinzel.
4550 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4552 * BN_mod_exp may produce incorrect results on x86_64
4554 There is a carry propagating bug in the x86_64 Montgomery squaring
4555 procedure. No EC algorithms are affected. Analysis suggests that attacks
4556 against RSA and DSA as a result of this defect would be very difficult to
4557 perform and are not believed likely. Attacks against DH are considered just
4558 feasible (although very difficult) because most of the work necessary to
4559 deduce information about a private key may be performed offline. The amount
4560 of resources required for such an attack would be very significant and
4561 likely only accessible to a limited number of attackers. An attacker would
4562 additionally need online access to an unpatched system using the target
4563 private key in a scenario with persistent DH parameters and a private
4564 key that is shared between multiple clients. For example this can occur by
4565 default in OpenSSL DHE based SSL/TLS ciphersuites.
4567 This issue was reported to OpenSSL by Hanno Böck.
4572 * Certificate verify crash with missing PSS parameter
4574 The signature verification routines will crash with a NULL pointer
4575 dereference if presented with an ASN.1 signature using the RSA PSS
4576 algorithm and absent mask generation function parameter. Since these
4577 routines are used to verify certificate signature algorithms this can be
4578 used to crash any certificate verification operation and exploited in a
4579 DoS attack. Any application which performs certificate verification is
4580 vulnerable including OpenSSL clients and servers which enable client
4583 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4588 * X509_ATTRIBUTE memory leak
4590 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4591 memory. This structure is used by the PKCS#7 and CMS routines so any
4592 application which reads PKCS#7 or CMS data from untrusted sources is
4593 affected. SSL/TLS is not affected.
4595 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4601 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4602 This changes the decoding behaviour for some invalid messages,
4603 though the change is mostly in the more lenient direction, and
4604 legacy behaviour is preserved as much as possible.
4608 * In DSA_generate_parameters_ex, if the provided seed is too short,
4611 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4613 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4615 * Alternate chains certificate forgery
4617 During certificate verification, OpenSSL will attempt to find an
4618 alternative certificate chain if the first attempt to build such a chain
4619 fails. An error in the implementation of this logic can mean that an
4620 attacker could cause certain checks on untrusted certificates to be
4621 bypassed, such as the CA flag, enabling them to use a valid leaf
4622 certificate to act as a CA and "issue" an invalid certificate.
4624 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4629 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4631 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4632 incompatibility in the handling of HMAC. The previous ABI has now been
4637 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4639 * Malformed ECParameters causes infinite loop
4641 When processing an ECParameters structure OpenSSL enters an infinite loop
4642 if the curve specified is over a specially malformed binary polynomial
4645 This can be used to perform denial of service against any
4646 system which processes public keys, certificate requests or
4647 certificates. This includes TLS clients and TLS servers with
4648 client authentication enabled.
4650 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4655 * Exploitable out-of-bounds read in X509_cmp_time
4657 X509_cmp_time does not properly check the length of the ASN1_TIME
4658 string and can read a few bytes out of bounds. In addition,
4659 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4662 An attacker can use this to craft malformed certificates and CRLs of
4663 various sizes and potentially cause a segmentation fault, resulting in
4664 a DoS on applications that verify certificates or CRLs. TLS clients
4665 that verify CRLs are affected. TLS clients and servers with client
4666 authentication enabled may be affected if they use custom verification
4669 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4670 independently by Hanno Böck.
4675 * PKCS7 crash with missing EnvelopedContent
4677 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4678 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4679 with missing content and trigger a NULL pointer dereference on parsing.
4681 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4682 structures from untrusted sources are affected. OpenSSL clients and
4683 servers are not affected.
4685 This issue was reported to OpenSSL by Michal Zalewski (Google).
4690 * CMS verify infinite loop with unknown hash function
4692 When verifying a signedData message the CMS code can enter an infinite loop
4693 if presented with an unknown hash function OID. This can be used to perform
4694 denial of service against any system which verifies signedData messages using
4696 This issue was reported to OpenSSL by Johannes Bauer.
4701 * Race condition handling NewSessionTicket
4703 If a NewSessionTicket is received by a multi-threaded client when attempting to
4704 reuse a previous ticket then a race condition can occur potentially leading to
4705 a double free of the ticket data.
4710 * Only support 256-bit or stronger elliptic curves with the
4711 'ecdh_auto' setting (server) or by default (client). Of supported
4712 curves, prefer P-256 (both).
4716 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4718 * ClientHello sigalgs DoS fix
4720 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4721 invalid signature algorithms extension a NULL pointer dereference will
4722 occur. This can be exploited in a DoS attack against the server.
4724 This issue was was reported to OpenSSL by David Ramos of Stanford
4728 *Stephen Henson and Matt Caswell*
4730 * Multiblock corrupted pointer fix
4732 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4733 feature only applies on 64 bit x86 architecture platforms that support AES
4734 NI instructions. A defect in the implementation of "multiblock" can cause
4735 OpenSSL's internal write buffer to become incorrectly set to NULL when
4736 using non-blocking IO. Typically, when the user application is using a
4737 socket BIO for writing, this will only result in a failed connection.
4738 However if some other BIO is used then it is likely that a segmentation
4739 fault will be triggered, thus enabling a potential DoS attack.
4741 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4746 * Segmentation fault in DTLSv1_listen fix
4748 The DTLSv1_listen function is intended to be stateless and processes the
4749 initial ClientHello from many peers. It is common for user code to loop
4750 over the call to DTLSv1_listen until a valid ClientHello is received with
4751 an associated cookie. A defect in the implementation of DTLSv1_listen means
4752 that state is preserved in the SSL object from one invocation to the next
4753 that can lead to a segmentation fault. Errors processing the initial
4754 ClientHello can trigger this scenario. An example of such an error could be
4755 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4758 This issue was reported to OpenSSL by Per Allansson.
4763 * Segmentation fault in ASN1_TYPE_cmp fix
4765 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4766 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4767 certificate signature algorithm consistency this can be used to crash any
4768 certificate verification operation and exploited in a DoS attack. Any
4769 application which performs certificate verification is vulnerable including
4770 OpenSSL clients and servers which enable client authentication.
4775 * Segmentation fault for invalid PSS parameters fix
4777 The signature verification routines will crash with a NULL pointer
4778 dereference if presented with an ASN.1 signature using the RSA PSS
4779 algorithm and invalid parameters. Since these routines are used to verify
4780 certificate signature algorithms this can be used to crash any
4781 certificate verification operation and exploited in a DoS attack. Any
4782 application which performs certificate verification is vulnerable including
4783 OpenSSL clients and servers which enable client authentication.
4785 This issue was was reported to OpenSSL by Brian Carpenter.
4790 * ASN.1 structure reuse memory corruption fix
4792 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4793 memory corruption via an invalid write. Such reuse is and has been
4794 strongly discouraged and is believed to be rare.
4796 Applications that parse structures containing CHOICE or ANY DEFINED BY
4797 components may be affected. Certificate parsing (d2i_X509 and related
4798 functions) are however not affected. OpenSSL clients and servers are
4804 * PKCS7 NULL pointer dereferences fix
4806 The PKCS#7 parsing code does not handle missing outer ContentInfo
4807 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4808 missing content and trigger a NULL pointer dereference on parsing.
4810 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4811 otherwise parse PKCS#7 structures from untrusted sources are
4812 affected. OpenSSL clients and servers are not affected.
4814 This issue was reported to OpenSSL by Michal Zalewski (Google).
4819 * DoS via reachable assert in SSLv2 servers fix
4821 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4822 servers that both support SSLv2 and enable export cipher suites by sending
4823 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4825 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4826 (OpenSSL development team).
4831 * Empty CKE with client auth and DHE fix
4833 If client auth is used then a server can seg fault in the event of a DHE
4834 ciphersuite being selected and a zero length ClientKeyExchange message
4835 being sent by the client. This could be exploited in a DoS attack.
4840 * Handshake with unseeded PRNG fix
4842 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4843 with an unseeded PRNG. The conditions are:
4844 - The client is on a platform where the PRNG has not been seeded
4845 automatically, and the user has not seeded manually
4846 - A protocol specific client method version has been used (i.e. not
4847 SSL_client_methodv23)
4848 - A ciphersuite is used that does not require additional random data from
4849 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4851 If the handshake succeeds then the client random that has been used will
4852 have been generated from a PRNG with insufficient entropy and therefore the
4853 output may be predictable.
4855 For example using the following command with an unseeded openssl will
4856 succeed on an unpatched platform:
4858 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4863 * Use After Free following d2i_ECPrivatekey error fix
4865 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4866 could cause a use after free condition. This, in turn, could cause a double
4867 free in several private key parsing functions (such as d2i_PrivateKey
4868 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4869 for applications that receive EC private keys from untrusted
4870 sources. This scenario is considered rare.
4872 This issue was discovered by the BoringSSL project and fixed in their
4878 * X509_to_X509_REQ NULL pointer deref fix
4880 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4881 the certificate key is invalid. This function is rarely used in practice.
4883 This issue was discovered by Brian Carpenter.
4888 * Removed the export ciphers from the DEFAULT ciphers
4892 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4894 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4895 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4896 So far those who have to target multiple platforms would compromise
4897 and argue that binary targeting say ARMv5 would still execute on
4898 ARMv8. "Universal" build resolves this compromise by providing
4899 near-optimal performance even on newer platforms.
4903 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4904 (other platforms pending).
4906 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4908 * Add support for the SignedCertificateTimestampList certificate and
4909 OCSP response extensions from RFC6962.
4913 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4914 for corner cases. (Certain input points at infinity could lead to
4915 bogus results, with non-infinity inputs mapped to infinity too.)
4919 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4920 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4921 common cases are optimized and there still is room for further
4922 improvements. Vector Permutation AES for Altivec is also added.
4926 * Add support for little-endian ppc64 Linux target.
4928 *Marcelo Cerri (IBM)*
4930 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4931 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4932 are optimized and there still is room for further improvements.
4933 Both 32- and 64-bit modes are supported.
4935 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4937 * Improved ARMv7 NEON support.
4941 * Support for SPARC Architecture 2011 crypto extensions, first
4942 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4943 SHA256/512, MD5, GHASH and modular exponentiation.
4945 *Andy Polyakov, David Miller*
4947 * Accelerated modular exponentiation for Intel processors, a.k.a.
4950 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4952 * Support for new and upcoming Intel processors, including AVX2,
4953 BMI and SHA ISA extensions. This includes additional "stitched"
4954 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4957 This work was sponsored by Intel Corp.
4961 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4962 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4963 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4967 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4968 this fixes a limitation in previous versions of OpenSSL.
4972 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4973 MGF1 digest and OAEP label.
4977 * Add EVP support for key wrapping algorithms, to avoid problems with
4978 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4979 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4980 algorithms and include tests cases.
4984 * Add functions to allocate and set the fields of an ECDSA_METHOD
4987 *Douglas E. Engert, Steve Henson*
4989 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4990 difference in days and seconds between two tm or ASN1_TIME structures.
4994 * Add -rev test option to s_server to just reverse order of characters
4995 received by client and send back to server. Also prints an abbreviated
4996 summary of the connection parameters.
5000 * New option -brief for s_client and s_server to print out a brief summary
5001 of connection parameters.
5005 * Add callbacks for arbitrary TLS extensions.
5007 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5009 * New option -crl_download in several openssl utilities to download CRLs
5010 from CRLDP extension in certificates.
5014 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5018 * New function X509_CRL_diff to generate a delta CRL from the difference
5019 of two full CRLs. Add support to "crl" utility.
5023 * New functions to set lookup_crls function and to retrieve
5024 X509_STORE from X509_STORE_CTX.
5028 * Print out deprecated issuer and subject unique ID fields in
5033 * Extend OCSP I/O functions so they can be used for simple general purpose
5034 HTTP as well as OCSP. New wrapper function which can be used to download
5035 CRLs using the OCSP API.
5039 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5043 * `SSL_CONF*` functions. These provide a common framework for application
5044 configuration using configuration files or command lines.
5048 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5049 message callback and prints the results. Needs compile time option
5050 "enable-ssl-trace". New options to s_client and s_server to enable
5055 * New ctrl and macro to retrieve supported points extensions.
5056 Print out extension in s_server and s_client.
5060 * New functions to retrieve certificate signature and signature
5065 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5070 * New Suite B modes for TLS code. These use and enforce the requirements
5071 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5072 only use Suite B curves. The Suite B modes can be set by using the
5073 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5077 * New chain verification flags for Suite B levels of security. Check
5078 algorithms are acceptable when flags are set in X509_verify_cert.
5082 * Make tls1_check_chain return a set of flags indicating checks passed
5083 by a certificate chain. Add additional tests to handle client
5084 certificates: checks for matching certificate type and issuer name
5089 * If an attempt is made to use a signature algorithm not in the peer
5090 preference list abort the handshake. If client has no suitable
5091 signature algorithms in response to a certificate request do not
5092 use the certificate.
5096 * If server EC tmp key is not in client preference list abort handshake.
5100 * Add support for certificate stores in CERT structure. This makes it
5101 possible to have different stores per SSL structure or one store in
5102 the parent SSL_CTX. Include distinct stores for certificate chain
5103 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5104 to build and store a certificate chain in CERT structure: returning
5105 an error if the chain cannot be built: this will allow applications
5106 to test if a chain is correctly configured.
5108 Note: if the CERT based stores are not set then the parent SSL_CTX
5109 store is used to retain compatibility with existing behaviour.
5113 * New function ssl_set_client_disabled to set a ciphersuite disabled
5114 mask based on the current session, check mask when sending client
5115 hello and checking the requested ciphersuite.
5119 * New ctrls to retrieve and set certificate types in a certificate
5120 request message. Print out received values in s_client. If certificate
5121 types is not set with custom values set sensible values based on
5122 supported signature algorithms.
5126 * Support for distinct client and server supported signature algorithms.
5130 * Add certificate callback. If set this is called whenever a certificate
5131 is required by client or server. An application can decide which
5132 certificate chain to present based on arbitrary criteria: for example
5133 supported signature algorithms. Add very simple example to s_server.
5134 This fixes many of the problems and restrictions of the existing client
5135 certificate callback: for example you can now clear an existing
5136 certificate and specify the whole chain.
5140 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5141 the certificate can be used for (if anything). Set valid_flags field
5142 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5143 to have similar checks in it.
5145 Add new "cert_flags" field to CERT structure and include a "strict mode".
5146 This enforces some TLS certificate requirements (such as only permitting
5147 certificate signature algorithms contained in the supported algorithms
5148 extension) which some implementations ignore: this option should be used
5149 with caution as it could cause interoperability issues.
5153 * Update and tidy signature algorithm extension processing. Work out
5154 shared signature algorithms based on preferences and peer algorithms
5155 and print them out in s_client and s_server. Abort handshake if no
5156 shared signature algorithms.
5160 * Add new functions to allow customised supported signature algorithms
5161 for SSL and SSL_CTX structures. Add options to s_client and s_server
5166 * New function SSL_certs_clear() to delete all references to certificates
5167 from an SSL structure. Before this once a certificate had been added
5168 it couldn't be removed.
5172 * Integrate hostname, email address and IP address checking with certificate
5173 verification. New verify options supporting checking in openssl utility.
5177 * Fixes and wildcard matching support to hostname and email checking
5178 functions. Add manual page.
5180 *Florian Weimer (Red Hat Product Security Team)*
5182 * New functions to check a hostname email or IP address against a
5183 certificate. Add options x509 utility to print results of checks against
5188 * Fix OCSP checking.
5190 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5192 * Initial experimental support for explicitly trusted non-root CAs.
5193 OpenSSL still tries to build a complete chain to a root but if an
5194 intermediate CA has a trust setting included that is used. The first
5195 setting is used: whether to trust (e.g., -addtrust option to the x509
5200 * Add -trusted_first option which attempts to find certificates in the
5201 trusted store even if an untrusted chain is also supplied.
5205 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5206 platform support for Linux and Android.
5210 * Support for linux-x32, ILP32 environment in x86_64 framework.
5214 * Experimental multi-implementation support for FIPS capable OpenSSL.
5215 When in FIPS mode the approved implementations are used as normal,
5216 when not in FIPS mode the internal unapproved versions are used instead.
5217 This means that the FIPS capable OpenSSL isn't forced to use the
5218 (often lower performance) FIPS implementations outside FIPS mode.
5222 * Transparently support X9.42 DH parameters when calling
5223 PEM_read_bio_DHparameters. This means existing applications can handle
5224 the new parameter format automatically.
5228 * Initial experimental support for X9.42 DH parameter format: mainly
5229 to support use of 'q' parameter for RFC5114 parameters.
5233 * Add DH parameters from RFC5114 including test data to dhtest.
5237 * Support for automatic EC temporary key parameter selection. If enabled
5238 the most preferred EC parameters are automatically used instead of
5239 hardcoded fixed parameters. Now a server just has to call:
5240 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5241 support ECDH and use the most appropriate parameters.
5245 * Enhance and tidy EC curve and point format TLS extension code. Use
5246 static structures instead of allocation if default values are used.
5247 New ctrls to set curves we wish to support and to retrieve shared curves.
5248 Print out shared curves in s_server. New options to s_server and s_client
5249 to set list of supported curves.
5253 * New ctrls to retrieve supported signature algorithms and
5254 supported curve values as an array of NIDs. Extend openssl utility
5255 to print out received values.
5259 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5260 between NIDs and the more common NIST names such as "P-256". Enhance
5261 ecparam utility and ECC method to recognise the NIST names for curves.
5265 * Enhance SSL/TLS certificate chain handling to support different
5266 chains for each certificate instead of one chain in the parent SSL_CTX.
5270 * Support for fixed DH ciphersuite client authentication: where both
5271 server and client use DH certificates with common parameters.
5275 * Support for fixed DH ciphersuites: those requiring DH server
5280 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5282 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5283 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5284 X509_CINF_get_signature were reverted post internal team review.
5289 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5291 * OCSP Status Request extension unbounded memory growth
5293 A malicious client can send an excessively large OCSP Status Request
5294 extension. If that client continually requests renegotiation, sending a
5295 large OCSP Status Request extension each time, then there will be unbounded
5296 memory growth on the server. This will eventually lead to a Denial Of
5297 Service attack through memory exhaustion. Servers with a default
5298 configuration are vulnerable even if they do not support OCSP. Builds using
5299 the "no-ocsp" build time option are not affected.
5301 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5306 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5309 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5315 * OOB write in MDC2_Update()
5317 An overflow can occur in MDC2_Update() either if called directly or
5318 through the EVP_DigestUpdate() function using MDC2. If an attacker
5319 is able to supply very large amounts of input data after a previous
5320 call to EVP_EncryptUpdate() with a partial block then a length check
5321 can overflow resulting in a heap corruption.
5323 The amount of data needed is comparable to SIZE_MAX which is impractical
5326 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5331 * Malformed SHA512 ticket DoS
5333 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5334 DoS attack where a malformed ticket will result in an OOB read which will
5337 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5338 a custom server callback and ticket lookup mechanism.
5340 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5345 * OOB write in BN_bn2dec()
5347 The function BN_bn2dec() does not check the return value of BN_div_word().
5348 This can cause an OOB write if an application uses this function with an
5349 overly large BIGNUM. This could be a problem if an overly large certificate
5350 or CRL is printed out from an untrusted source. TLS is not affected because
5351 record limits will reject an oversized certificate before it is parsed.
5353 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5358 * OOB read in TS_OBJ_print_bio()
5360 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5361 the total length the OID text representation would use and not the amount
5362 of data written. This will result in OOB reads when large OIDs are
5365 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5370 * Pointer arithmetic undefined behaviour
5372 Avoid some undefined pointer arithmetic
5374 A common idiom in the codebase is to check limits in the following manner:
5377 Where "p" points to some malloc'd data of SIZE bytes and
5380 "len" here could be from some externally supplied data (e.g. from a TLS
5383 The rules of C pointer arithmetic are such that "p + len" is only well
5384 defined where len <= SIZE. Therefore the above idiom is actually
5385 undefined behaviour.
5387 For example this could cause problems if some malloc implementation
5388 provides an address for "p" such that "p + len" actually overflows for
5389 values of len that are too big and therefore p + len < limit.
5391 This issue was reported to OpenSSL by Guido Vranken
5396 * Constant time flag not preserved in DSA signing
5398 Operations in the DSA signing algorithm should run in constant time in
5399 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5400 implementation means that a non-constant time codepath is followed for
5401 certain operations. This has been demonstrated through a cache-timing
5402 attack to be sufficient for an attacker to recover the private DSA key.
5404 This issue was reported by César Pereida (Aalto University), Billy Brumley
5405 (Tampere University of Technology), and Yuval Yarom (The University of
5406 Adelaide and NICTA).
5411 * DTLS buffered message DoS
5413 In a DTLS connection where handshake messages are delivered out-of-order
5414 those messages that OpenSSL is not yet ready to process will be buffered
5415 for later use. Under certain circumstances, a flaw in the logic means that
5416 those messages do not get removed from the buffer even though the handshake
5417 has been completed. An attacker could force up to approx. 15 messages to
5418 remain in the buffer when they are no longer required. These messages will
5419 be cleared when the DTLS connection is closed. The default maximum size for
5420 a message is 100k. Therefore the attacker could force an additional 1500k
5421 to be consumed per connection. By opening many simulataneous connections an
5422 attacker could cause a DoS attack through memory exhaustion.
5424 This issue was reported to OpenSSL by Quan Luo.
5429 * DTLS replay protection DoS
5431 A flaw in the DTLS replay attack protection mechanism means that records
5432 that arrive for future epochs update the replay protection "window" before
5433 the MAC for the record has been validated. This could be exploited by an
5434 attacker by sending a record for the next epoch (which does not have to
5435 decrypt or have a valid MAC), with a very large sequence number. This means
5436 that all subsequent legitimate packets are dropped causing a denial of
5437 service for a specific DTLS connection.
5439 This issue was reported to OpenSSL by the OCAP audit team.
5444 * Certificate message OOB reads
5446 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5447 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5448 theoretical DoS risk but this has not been observed in practice on common
5451 The messages affected are client certificate, client certificate request
5452 and server certificate. As a result the attack can only be performed
5453 against a client or a server which enables client authentication.
5455 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5460 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5462 * Prevent padding oracle in AES-NI CBC MAC check
5464 A MITM attacker can use a padding oracle attack to decrypt traffic
5465 when the connection uses an AES CBC cipher and the server support
5468 This issue was introduced as part of the fix for Lucky 13 padding
5469 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5470 constant time by making sure that always the same bytes are read and
5471 compared against either the MAC or padding bytes. But it no longer
5472 checked that there was enough data to have both the MAC and padding
5475 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5480 * Fix EVP_EncodeUpdate overflow
5482 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5483 Base64 encoding of binary data. If an attacker is able to supply very large
5484 amounts of input data then a length check can overflow resulting in a heap
5487 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5488 the `PEM_write_bio*` family of functions. These are mainly used within the
5489 OpenSSL command line applications, so any application which processes data
5490 from an untrusted source and outputs it as a PEM file should be considered
5491 vulnerable to this issue. User applications that call these APIs directly
5492 with large amounts of untrusted data may also be vulnerable.
5494 This issue was reported by Guido Vranken.
5499 * Fix EVP_EncryptUpdate overflow
5501 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5502 is able to supply very large amounts of input data after a previous call to
5503 EVP_EncryptUpdate() with a partial block then a length check can overflow
5504 resulting in a heap corruption. Following an analysis of all OpenSSL
5505 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5506 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5507 the first called function after an EVP_EncryptInit(), and therefore that
5508 specific call must be safe. The second form is where the length passed to
5509 EVP_EncryptUpdate() can be seen from the code to be some small value and
5510 therefore there is no possibility of an overflow. Since all instances are
5511 one of these two forms, it is believed that there can be no overflows in
5512 internal code due to this problem. It should be noted that
5513 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5514 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5515 of these calls have also been analysed too and it is believed there are no
5516 instances in internal usage where an overflow could occur.
5518 This issue was reported by Guido Vranken.
5523 * Prevent ASN.1 BIO excessive memory allocation
5525 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5526 a short invalid encoding can casuse allocation of large amounts of memory
5527 potentially consuming excessive resources or exhausting memory.
5529 Any application parsing untrusted data through d2i BIO functions is
5530 affected. The memory based functions such as d2i_X509() are *not* affected.
5531 Since the memory based functions are used by the TLS library, TLS
5532 applications are not affected.
5534 This issue was reported by Brian Carpenter.
5541 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5542 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5543 in arbitrary stack data being returned in the buffer.
5545 This issue was reported by Guido Vranken.
5550 * Modify behavior of ALPN to invoke callback after SNI/servername
5551 callback, such that updates to the SSL_CTX affect ALPN.
5555 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5560 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5561 methods are enabled and ssl2 is disabled the methods return NULL.
5565 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5567 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5568 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5569 provide any "EXPORT" or "LOW" strength ciphers.
5573 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5574 is by default disabled at build-time. Builds that are not configured with
5575 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5576 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5577 will need to explicitly call either of:
5579 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5581 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5583 as appropriate. Even if either of those is used, or the application
5584 explicitly uses the version-specific SSLv2_method() or its client and
5585 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5586 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5587 ciphers, and SSLv2 56-bit DES are no longer available.
5592 * Fix a double-free in DSA code
5594 A double free bug was discovered when OpenSSL parses malformed DSA private
5595 keys and could lead to a DoS attack or memory corruption for applications
5596 that receive DSA private keys from untrusted sources. This scenario is
5599 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5605 * Disable SRP fake user seed to address a server memory leak.
5607 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5609 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5610 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5611 was changed to ignore the "fake user" SRP seed, even if the seed
5614 Users should use SRP_VBASE_get1_by_user instead. Note that in
5615 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5616 also that even though configuring the SRP seed attempts to hide
5617 invalid usernames by continuing the handshake with fake
5618 credentials, this behaviour is not constant time and no strong
5619 guarantees are made that the handshake is indistinguishable from
5620 that of a valid user.
5625 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5627 In the BN_hex2bn function the number of hex digits is calculated using an
5628 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5629 large values of |i| this can result in |bn_expand| not allocating any
5630 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5631 field as NULL leading to a subsequent NULL ptr deref. For very large values
5632 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5633 In this case memory is allocated to the internal BIGNUM data field, but it
5634 is insufficiently sized leading to heap corruption. A similar issue exists
5635 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5636 is ever called by user applications with very large untrusted hex/dec data.
5637 This is anticipated to be a rare occurrence.
5639 All OpenSSL internal usage of these functions use data that is not expected
5640 to be untrusted, e.g. config file data or application command line
5641 arguments. If user developed applications generate config file data based
5642 on untrusted data then it is possible that this could also lead to security
5643 consequences. This is also anticipated to be rare.
5645 This issue was reported to OpenSSL by Guido Vranken.
5650 * Fix memory issues in `BIO_*printf` functions
5652 The internal |fmtstr| function used in processing a "%s" format string in
5653 the `BIO_*printf` functions could overflow while calculating the length of a
5654 string and cause an OOB read when printing very long strings.
5656 Additionally the internal |doapr_outch| function can attempt to write to an
5657 OOB memory location (at an offset from the NULL pointer) in the event of a
5658 memory allocation failure. In 1.0.2 and below this could be caused where
5659 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5660 could be in processing a very long "%s" format string. Memory leaks can
5663 The first issue may mask the second issue dependent on compiler behaviour.
5664 These problems could enable attacks where large amounts of untrusted data
5665 is passed to the `BIO_*printf` functions. If applications use these functions
5666 in this way then they could be vulnerable. OpenSSL itself uses these
5667 functions when printing out human-readable dumps of ASN.1 data. Therefore
5668 applications that print this data could be vulnerable if the data is from
5669 untrusted sources. OpenSSL command line applications could also be
5670 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5671 as command line arguments.
5673 Libssl is not considered directly vulnerable. Additionally certificates etc
5674 received via remote connections via libssl are also unlikely to be able to
5675 trigger these issues because of message size limits enforced within libssl.
5677 This issue was reported to OpenSSL Guido Vranken.
5682 * Side channel attack on modular exponentiation
5684 A side-channel attack was found which makes use of cache-bank conflicts on
5685 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5686 of RSA keys. The ability to exploit this issue is limited as it relies on
5687 an attacker who has control of code in a thread running on the same
5688 hyper-threaded core as the victim thread which is performing decryptions.
5690 This issue was reported to OpenSSL by Yuval Yarom, The University of
5691 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5692 Nadia Heninger, University of Pennsylvania with more information at
5693 <http://cachebleed.info>.
5698 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5699 if no keysize is specified with default_bits. This fixes an
5700 omission in an earlier change that changed all RSA/DSA key generation
5701 apps to use 2048 bits by default.
5705 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5707 * Protection for DH small subgroup attacks
5709 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5710 switched on by default and cannot be disabled. This could have some
5715 * SSLv2 doesn't block disabled ciphers
5717 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5718 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5719 been disabled, provided that the SSLv2 protocol was not also disabled via
5722 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5723 and Sebastian Schinzel.
5728 * Reject DH handshakes with parameters shorter than 1024 bits.
5732 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5734 * Certificate verify crash with missing PSS parameter
5736 The signature verification routines will crash with a NULL pointer
5737 dereference if presented with an ASN.1 signature using the RSA PSS
5738 algorithm and absent mask generation function parameter. Since these
5739 routines are used to verify certificate signature algorithms this can be
5740 used to crash any certificate verification operation and exploited in a
5741 DoS attack. Any application which performs certificate verification is
5742 vulnerable including OpenSSL clients and servers which enable client
5745 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5750 * X509_ATTRIBUTE memory leak
5752 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5753 memory. This structure is used by the PKCS#7 and CMS routines so any
5754 application which reads PKCS#7 or CMS data from untrusted sources is
5755 affected. SSL/TLS is not affected.
5757 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5763 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5764 This changes the decoding behaviour for some invalid messages,
5765 though the change is mostly in the more lenient direction, and
5766 legacy behaviour is preserved as much as possible.
5770 * In DSA_generate_parameters_ex, if the provided seed is too short,
5771 use a random seed, as already documented.
5773 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5775 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5777 * Alternate chains certificate forgery
5779 During certificate verfification, OpenSSL will attempt to find an
5780 alternative certificate chain if the first attempt to build such a chain
5781 fails. An error in the implementation of this logic can mean that an
5782 attacker could cause certain checks on untrusted certificates to be
5783 bypassed, such as the CA flag, enabling them to use a valid leaf
5784 certificate to act as a CA and "issue" an invalid certificate.
5786 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5792 * Race condition handling PSK identify hint
5794 If PSK identity hints are received by a multi-threaded client then
5795 the values are wrongly updated in the parent SSL_CTX structure. This can
5796 result in a race condition potentially leading to a double free of the
5802 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5804 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5805 incompatibility in the handling of HMAC. The previous ABI has now been
5808 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5810 * Malformed ECParameters causes infinite loop
5812 When processing an ECParameters structure OpenSSL enters an infinite loop
5813 if the curve specified is over a specially malformed binary polynomial
5816 This can be used to perform denial of service against any
5817 system which processes public keys, certificate requests or
5818 certificates. This includes TLS clients and TLS servers with
5819 client authentication enabled.
5821 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5826 * Exploitable out-of-bounds read in X509_cmp_time
5828 X509_cmp_time does not properly check the length of the ASN1_TIME
5829 string and can read a few bytes out of bounds. In addition,
5830 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5833 An attacker can use this to craft malformed certificates and CRLs of
5834 various sizes and potentially cause a segmentation fault, resulting in
5835 a DoS on applications that verify certificates or CRLs. TLS clients
5836 that verify CRLs are affected. TLS clients and servers with client
5837 authentication enabled may be affected if they use custom verification
5840 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5841 independently by Hanno Böck.
5846 * PKCS7 crash with missing EnvelopedContent
5848 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5849 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5850 with missing content and trigger a NULL pointer dereference on parsing.
5852 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5853 structures from untrusted sources are affected. OpenSSL clients and
5854 servers are not affected.
5856 This issue was reported to OpenSSL by Michal Zalewski (Google).
5861 * CMS verify infinite loop with unknown hash function
5863 When verifying a signedData message the CMS code can enter an infinite loop
5864 if presented with an unknown hash function OID. This can be used to perform
5865 denial of service against any system which verifies signedData messages using
5867 This issue was reported to OpenSSL by Johannes Bauer.
5872 * Race condition handling NewSessionTicket
5874 If a NewSessionTicket is received by a multi-threaded client when attempting to
5875 reuse a previous ticket then a race condition can occur potentially leading to
5876 a double free of the ticket data.
5881 * Reject DH handshakes with parameters shorter than 768 bits.
5883 *Kurt Roeckx and Emilia Kasper*
5885 * dhparam: generate 2048-bit parameters by default.
5887 *Kurt Roeckx and Emilia Kasper*
5889 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5891 * Segmentation fault in ASN1_TYPE_cmp fix
5893 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5894 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5895 certificate signature algorithm consistency this can be used to crash any
5896 certificate verification operation and exploited in a DoS attack. Any
5897 application which performs certificate verification is vulnerable including
5898 OpenSSL clients and servers which enable client authentication.
5903 * ASN.1 structure reuse memory corruption fix
5905 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5906 memory corruption via an invalid write. Such reuse is and has been
5907 strongly discouraged and is believed to be rare.
5909 Applications that parse structures containing CHOICE or ANY DEFINED BY
5910 components may be affected. Certificate parsing (d2i_X509 and related
5911 functions) are however not affected. OpenSSL clients and servers are
5917 * PKCS7 NULL pointer dereferences fix
5919 The PKCS#7 parsing code does not handle missing outer ContentInfo
5920 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5921 missing content and trigger a NULL pointer dereference on parsing.
5923 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5924 otherwise parse PKCS#7 structures from untrusted sources are
5925 affected. OpenSSL clients and servers are not affected.
5927 This issue was reported to OpenSSL by Michal Zalewski (Google).
5932 * DoS via reachable assert in SSLv2 servers fix
5934 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5935 servers that both support SSLv2 and enable export cipher suites by sending
5936 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5938 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5939 (OpenSSL development team).
5944 * Use After Free following d2i_ECPrivatekey error fix
5946 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5947 could cause a use after free condition. This, in turn, could cause a double
5948 free in several private key parsing functions (such as d2i_PrivateKey
5949 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5950 for applications that receive EC private keys from untrusted
5951 sources. This scenario is considered rare.
5953 This issue was discovered by the BoringSSL project and fixed in their
5959 * X509_to_X509_REQ NULL pointer deref fix
5961 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5962 the certificate key is invalid. This function is rarely used in practice.
5964 This issue was discovered by Brian Carpenter.
5969 * Removed the export ciphers from the DEFAULT ciphers
5973 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
5975 * Build fixes for the Windows and OpenVMS platforms
5977 *Matt Caswell and Richard Levitte*
5979 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
5981 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5982 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5983 dereference. This could lead to a Denial Of Service attack. Thanks to
5984 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5989 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5990 dtls1_buffer_record function under certain conditions. In particular this
5991 could occur if an attacker sent repeated DTLS records with the same
5992 sequence number but for the next epoch. The memory leak could be exploited
5993 by an attacker in a Denial of Service attack through memory exhaustion.
5994 Thanks to Chris Mueller for reporting this issue.
5999 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6000 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6001 method would be set to NULL which could later result in a NULL pointer
6002 dereference. Thanks to Frank Schmirler for reporting this issue.
6007 * Abort handshake if server key exchange message is omitted for ephemeral
6010 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6011 reporting this issue.
6016 * Remove non-export ephemeral RSA code on client and server. This code
6017 violated the TLS standard by allowing the use of temporary RSA keys in
6018 non-export ciphersuites and could be used by a server to effectively
6019 downgrade the RSA key length used to a value smaller than the server
6020 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6021 INRIA or reporting this issue.
6026 * Fixed issue where DH client certificates are accepted without verification.
6027 An OpenSSL server will accept a DH certificate for client authentication
6028 without the certificate verify message. This effectively allows a client to
6029 authenticate without the use of a private key. This only affects servers
6030 which trust a client certificate authority which issues certificates
6031 containing DH keys: these are extremely rare and hardly ever encountered.
6032 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6038 * Ensure that the session ID context of an SSL is updated when its
6039 SSL_CTX is updated via SSL_set_SSL_CTX.
6041 The session ID context is typically set from the parent SSL_CTX,
6042 and can vary with the CTX.
6046 * Fix various certificate fingerprint issues.
6048 By using non-DER or invalid encodings outside the signed portion of a
6049 certificate the fingerprint can be changed without breaking the signature.
6050 Although no details of the signed portion of the certificate can be changed
6051 this can cause problems with some applications: e.g. those using the
6052 certificate fingerprint for blacklists.
6054 1. Reject signatures with non zero unused bits.
6056 If the BIT STRING containing the signature has non zero unused bits reject
6057 the signature. All current signature algorithms require zero unused bits.
6059 2. Check certificate algorithm consistency.
6061 Check the AlgorithmIdentifier inside TBS matches the one in the
6062 certificate signature. NB: this will result in signature failure
6063 errors for some broken certificates.
6065 Thanks to Konrad Kraszewski from Google for reporting this issue.
6067 3. Check DSA/ECDSA signatures use DER.
6069 Re-encode DSA/ECDSA signatures and compare with the original received
6070 signature. Return an error if there is a mismatch.
6072 This will reject various cases including garbage after signature
6073 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6074 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6075 (negative or with leading zeroes).
6077 Further analysis was conducted and fixes were developed by Stephen Henson
6078 of the OpenSSL core team.
6084 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6085 results on some platforms, including x86_64. This bug occurs at random
6086 with a very low probability, and is not known to be exploitable in any
6087 way, though its exact impact is difficult to determine. Thanks to Pieter
6088 Wuille (Blockstream) who reported this issue and also suggested an initial
6089 fix. Further analysis was conducted by the OpenSSL development team and
6090 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6091 the OpenSSL core team.
6096 * Do not resume sessions on the server if the negotiated protocol
6097 version does not match the session's version. Resuming with a different
6098 version, while not strictly forbidden by the RFC, is of questionable
6099 sanity and breaks all known clients.
6101 *David Benjamin, Emilia Käsper*
6103 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6104 early CCS messages during renegotiation. (Note that because
6105 renegotiation is encrypted, this early CCS was not exploitable.)
6109 * Tighten client-side session ticket handling during renegotiation:
6110 ensure that the client only accepts a session ticket if the server sends
6111 the extension anew in the ServerHello. Previously, a TLS client would
6112 reuse the old extension state and thus accept a session ticket if one was
6113 announced in the initial ServerHello.
6115 Similarly, ensure that the client requires a session ticket if one
6116 was advertised in the ServerHello. Previously, a TLS client would
6117 ignore a missing NewSessionTicket message.
6121 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6125 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6126 sends a carefully crafted handshake message, to cause OpenSSL to fail
6127 to free up to 64k of memory causing a memory leak. This could be
6128 exploited in a Denial Of Service attack. This issue affects OpenSSL
6129 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6130 whether SRTP is used or configured. Implementations of OpenSSL that
6131 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6133 The fix was developed by the OpenSSL team.
6138 * Session Ticket Memory Leak.
6140 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6141 integrity of that ticket is first verified. In the event of a session
6142 ticket integrity check failing, OpenSSL will fail to free memory
6143 causing a memory leak. By sending a large number of invalid session
6144 tickets an attacker could exploit this issue in a Denial Of Service
6150 * Build option no-ssl3 is incomplete.
6152 When OpenSSL is configured with "no-ssl3" as a build option, servers
6153 could accept and complete a SSL 3.0 handshake, and clients could be
6154 configured to send them.
6157 *Akamai and the OpenSSL team*
6159 * Add support for TLS_FALLBACK_SCSV.
6160 Client applications doing fallback retries should call
6161 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6164 *Adam Langley, Bodo Moeller*
6166 * Add additional DigestInfo checks.
6168 Re-encode DigestInto in DER and check against the original when
6169 verifying RSA signature: this will reject any improperly encoded
6170 DigestInfo structures.
6172 Note: this is a precautionary measure and no attacks are currently known.
6176 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6178 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6179 SRP code can be overrun an internal buffer. Add sanity check that
6180 g, A, B < N to SRP code.
6182 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6183 Group for discovering this issue.
6188 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6189 TLS 1.0 instead of higher protocol versions when the ClientHello message
6190 is badly fragmented. This allows a man-in-the-middle attacker to force a
6191 downgrade to TLS 1.0 even if both the server and the client support a
6192 higher protocol version, by modifying the client's TLS records.
6194 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6195 researching this issue.
6200 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6201 to a denial of service attack. A malicious server can crash the client
6202 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6203 ciphersuite and sending carefully crafted handshake messages.
6205 Thanks to Felix Gröbert (Google) for discovering and researching this
6211 * By sending carefully crafted DTLS packets an attacker could cause openssl
6212 to leak memory. This can be exploited through a Denial of Service attack.
6213 Thanks to Adam Langley for discovering and researching this issue.
6218 * An attacker can force openssl to consume large amounts of memory whilst
6219 processing DTLS handshake messages. This can be exploited through a
6220 Denial of Service attack.
6221 Thanks to Adam Langley for discovering and researching this issue.
6226 * An attacker can force an error condition which causes openssl to crash
6227 whilst processing DTLS packets due to memory being freed twice. This
6228 can be exploited through a Denial of Service attack.
6229 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6235 * If a multithreaded client connects to a malicious server using a resumed
6236 session and the server sends an ec point format extension it could write
6237 up to 255 bytes to freed memory.
6239 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6245 * A malicious server can crash an OpenSSL client with a null pointer
6246 dereference (read) by specifying an SRP ciphersuite even though it was not
6247 properly negotiated with the client. This can be exploited through a
6248 Denial of Service attack.
6250 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6251 discovering and researching this issue.
6256 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6257 X509_name_oneline, X509_name_print_ex et al. to leak some information
6258 from the stack. Applications may be affected if they echo pretty printing
6259 output to the attacker.
6261 Thanks to Ivan Fratric (Google) for discovering this issue.
6264 *Emilia Käsper, and Steve Henson*
6266 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6267 for corner cases. (Certain input points at infinity could lead to
6268 bogus results, with non-infinity inputs mapped to infinity too.)
6272 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6274 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6275 handshake can force the use of weak keying material in OpenSSL
6276 SSL/TLS clients and servers.
6278 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6279 researching this issue. [CVE-2014-0224][]
6281 *KIKUCHI Masashi, Steve Henson*
6283 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6284 OpenSSL DTLS client the code can be made to recurse eventually crashing
6287 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6290 *Imre Rad, Steve Henson*
6292 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6293 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6294 client or server. This is potentially exploitable to run arbitrary
6295 code on a vulnerable client or server.
6297 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6299 *Jüri Aedla, Steve Henson*
6301 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6302 are subject to a denial of service attack.
6304 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6305 this issue. [CVE-2014-3470][]
6307 *Felix Gröbert, Ivan Fratric, Steve Henson*
6309 * Harmonize version and its documentation. -f flag is used to display
6312 *mancha <mancha1@zoho.com>*
6314 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6315 in i2d_ECPrivateKey.
6317 *mancha <mancha1@zoho.com>*
6319 * Fix some double frees. These are not thought to be exploitable.
6321 *mancha <mancha1@zoho.com>*
6323 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6325 * A missing bounds check in the handling of the TLS heartbeat extension
6326 can be used to reveal up to 64k of memory to a connected client or
6329 Thanks for Neel Mehta of Google Security for discovering this bug and to
6330 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6331 preparing the fix [CVE-2014-0160][]
6333 *Adam Langley, Bodo Moeller*
6335 * Fix for the attack described in the paper "Recovering OpenSSL
6336 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6337 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6338 <http://eprint.iacr.org/2014/140>
6340 Thanks to Yuval Yarom and Naomi Benger for discovering this
6341 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6343 *Yuval Yarom and Naomi Benger*
6345 * TLS pad extension: draft-agl-tls-padding-03
6347 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6348 TLS client Hello record length value would otherwise be > 255 and
6349 less that 512 pad with a dummy extension containing zeroes so it
6350 is at least 512 bytes long.
6352 *Adam Langley, Steve Henson*
6354 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6356 * Fix for TLS record tampering bug. A carefully crafted invalid
6357 handshake could crash OpenSSL with a NULL pointer exception.
6358 Thanks to Anton Johansson for reporting this issues.
6361 * Keep original DTLS digest and encryption contexts in retransmission
6362 structures so we can use the previous session parameters if they need
6363 to be resent. [CVE-2013-6450][]
6367 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6368 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6369 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6370 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6371 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6372 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6374 *Rob Stradling, Adam Langley*
6376 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6378 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6379 supporting platforms or when small records were transferred.
6381 *Andy Polyakov, Steve Henson*
6383 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6385 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6387 This addresses the flaw in CBC record processing discovered by
6388 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6389 at: <http://www.isg.rhul.ac.uk/tls/>
6391 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6392 Security Group at Royal Holloway, University of London
6393 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6394 Emilia Käsper for the initial patch.
6397 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6399 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6400 ciphersuites which can be exploited in a denial of service attack.
6401 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6402 and detecting this bug and to Wolfgang Ettlinger
6403 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6408 * Return an error when checking OCSP signatures when key is NULL.
6409 This fixes a DoS attack. [CVE-2013-0166][]
6413 * Make openssl verify return errors.
6415 *Chris Palmer <palmer@google.com> and Ben Laurie*
6417 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6418 the right response is stapled. Also change SSL_get_certificate()
6419 so it returns the certificate actually sent.
6420 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6422 *Rob Stradling <rob.stradling@comodo.com>*
6424 * Fix possible deadlock when decoding public keys.
6428 * Don't use TLS 1.0 record version number in initial client hello
6433 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6435 * Sanity check record length before skipping explicit IV in TLS
6436 1.2, 1.1 and DTLS to fix DoS attack.
6438 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6439 fuzzing as a service testing platform.
6444 * Initialise tkeylen properly when encrypting CMS messages.
6445 Thanks to Solar Designer of Openwall for reporting this issue.
6449 * In FIPS mode don't try to use composite ciphers as they are not
6454 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6456 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6457 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6458 mean any application compiled against OpenSSL 1.0.0 headers setting
6459 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6460 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6461 0x10000000L Any application which was previously compiled against
6462 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6463 will need to be recompiled as a result. Letting be results in
6464 inability to disable specifically TLS 1.1 and in client context,
6465 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6469 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6470 disable just protocol X, but all protocols above X *if* there are
6471 protocols *below* X still enabled. In more practical terms it means
6472 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6473 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6474 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6479 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6481 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6482 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6483 in CRYPTO_realloc_clean.
6485 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6486 issue and to Adam Langley <agl@chromium.org> for fixing it.
6489 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6491 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6495 * Workarounds for some broken servers that "hang" if a client hello
6496 record length exceeds 255 bytes.
6498 1. Do not use record version number > TLS 1.0 in initial client
6499 hello: some (but not all) hanging servers will now work.
6500 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6501 the number of ciphers sent in the client hello. This should be
6502 set to an even number, such as 50, for example by passing:
6503 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6504 Most broken servers should now work.
6505 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6506 TLS 1.2 client support entirely.
6510 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6514 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6516 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6517 STRING form instead of a DigestInfo.
6521 * The format used for MDC2 RSA signatures is inconsistent between EVP
6522 and the RSA_sign/RSA_verify functions. This was made more apparent when
6523 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6524 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6525 the correct format in RSA_verify so both forms transparently work.
6529 * Some servers which support TLS 1.0 can choke if we initially indicate
6530 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6531 encrypted premaster secret. As a workaround use the maximum permitted
6532 client version in client hello, this should keep such servers happy
6533 and still work with previous versions of OpenSSL.
6537 * Add support for TLS/DTLS heartbeats.
6539 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6541 * Add support for SCTP.
6543 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6545 * Improved PRNG seeding for VOS.
6547 *Paul Green <Paul.Green@stratus.com>*
6549 * Extensive assembler packs updates, most notably:
6551 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6552 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6553 - x86_64: bit-sliced AES implementation;
6554 - ARM: NEON support, contemporary platforms optimizations;
6555 - s390x: z196 support;
6556 - `*`: GHASH and GF(2^m) multiplication implementations;
6560 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6561 (removal of unnecessary code)
6563 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6565 * Add TLS key material exporter from RFC 5705.
6569 * Add DTLS-SRTP negotiation from RFC 5764.
6573 * Add Next Protocol Negotiation,
6574 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6575 disabled with a no-npn flag to config or Configure. Code donated
6578 *Adam Langley <agl@google.com> and Ben Laurie*
6580 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6581 NIST-P256, NIST-P521, with constant-time single point multiplication on
6582 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6583 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6584 Code made available under Apache License version 2.0.
6586 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6587 line to include this in your build of OpenSSL, and run "make depend" (or
6588 "make update"). This enables the following EC_METHODs:
6590 EC_GFp_nistp224_method()
6591 EC_GFp_nistp256_method()
6592 EC_GFp_nistp521_method()
6594 EC_GROUP_new_by_curve_name() will automatically use these (while
6595 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6598 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6600 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6601 all platforms. Move ssize_t definition from e_os.h to the public
6602 header file e_os2.h as it now appears in public header file cms.h
6606 * New -sigopt option to the ca, req and x509 utilities. Additional
6607 signature parameters can be passed using this option and in
6612 * Add RSA PSS signing function. This will generate and set the
6613 appropriate AlgorithmIdentifiers for PSS based on those in the
6614 corresponding EVP_MD_CTX structure. No application support yet.
6618 * Support for companion algorithm specific ASN1 signing routines.
6619 New function ASN1_item_sign_ctx() signs a pre-initialised
6620 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6621 the appropriate parameters.
6625 * Add new algorithm specific ASN1 verification initialisation function
6626 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6627 handling will be the same no matter what EVP_PKEY_METHOD is used.
6628 Add a PSS handler to support verification of PSS signatures: checked
6629 against a number of sample certificates.
6633 * Add signature printing for PSS. Add PSS OIDs.
6635 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6637 * Add algorithm specific signature printing. An individual ASN1 method
6638 can now print out signatures instead of the standard hex dump.
6640 More complex signatures (e.g. PSS) can print out more meaningful
6641 information. Include DSA version that prints out the signature
6646 * Password based recipient info support for CMS library: implementing
6651 * Split password based encryption into PBES2 and PBKDF2 functions. This
6652 neatly separates the code into cipher and PBE sections and is required
6653 for some algorithms that split PBES2 into separate pieces (such as
6654 password based CMS).
6658 * Session-handling fixes:
6659 - Fix handling of connections that are resuming with a session ID,
6660 but also support Session Tickets.
6661 - Fix a bug that suppressed issuing of a new ticket if the client
6662 presented a ticket with an expired session.
6663 - Try to set the ticket lifetime hint to something reasonable.
6664 - Make tickets shorter by excluding irrelevant information.
6665 - On the client side, don't ignore renewed tickets.
6667 *Adam Langley, Bodo Moeller (Google)*
6669 * Fix PSK session representation.
6673 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6675 This work was sponsored by Intel.
6679 * Add GCM support to TLS library. Some custom code is needed to split
6680 the IV between the fixed (from PRF) and explicit (from TLS record)
6681 portions. This adds all GCM ciphersuites supported by RFC5288 and
6682 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6683 add a special AESGCM string for GCM only.
6687 * Expand range of ctrls for AES GCM. Permit setting invocation
6688 field on decrypt and retrieval of invocation field only on encrypt.
6692 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6693 As required by RFC5289 these ciphersuites cannot be used if for
6694 versions of TLS earlier than 1.2.
6698 * For FIPS capable OpenSSL interpret a NULL default public key method
6699 as unset and return the appropriate default but do *not* set the default.
6700 This means we can return the appropriate method in applications that
6701 switch between FIPS and non-FIPS modes.
6705 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6706 ENGINE is used then we cannot handle that in the FIPS module so we
6707 keep original code iff non-FIPS operations are allowed.
6711 * Add -attime option to openssl utilities.
6713 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6715 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6719 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6720 FIPS EC methods unconditionally for now.
6724 * New build option no-ec2m to disable characteristic 2 code.
6728 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6729 all cases can be covered as some introduce binary incompatibilities.
6733 * Redirect RSA operations to FIPS module including keygen,
6734 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6738 * Add similar low level API blocking to ciphers.
6742 * Low level digest APIs are not approved in FIPS mode: any attempt
6743 to use these will cause a fatal error. Applications that *really* want
6744 to use them can use the `private_*` version instead.
6748 * Redirect cipher operations to FIPS module for FIPS builds.
6752 * Redirect digest operations to FIPS module for FIPS builds.
6756 * Update build system to add "fips" flag which will link in fipscanister.o
6757 for static and shared library builds embedding a signature if needed.
6761 * Output TLS supported curves in preference order instead of numerical
6762 order. This is currently hardcoded for the highest order curves first.
6763 This should be configurable so applications can judge speed vs strength.
6767 * Add TLS v1.2 server support for client authentication.
6771 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6776 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6777 FIPS modules versions.
6781 * Add TLS v1.2 client side support for client authentication. Keep cache
6782 of handshake records longer as we don't know the hash algorithm to use
6783 until after the certificate request message is received.
6787 * Initial TLS v1.2 client support. Add a default signature algorithms
6788 extension including all the algorithms we support. Parse new signature
6789 format in client key exchange. Relax some ECC signing restrictions for
6790 TLS v1.2 as indicated in RFC5246.
6794 * Add server support for TLS v1.2 signature algorithms extension. Switch
6795 to new signature format when needed using client digest preference.
6796 All server ciphersuites should now work correctly in TLS v1.2. No client
6797 support yet and no support for client certificates.
6801 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6802 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6803 ciphersuites. At present only RSA key exchange ciphersuites work with
6804 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6805 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6806 and version checking.
6810 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6811 with this defined it will not be affected by any changes to ssl internal
6812 structures. Add several utility functions to allow openssl application
6813 to work with OPENSSL_NO_SSL_INTERN defined.
6817 * A long standing patch to add support for SRP from EdelWeb (Peter
6818 Sylvester and Christophe Renou) was integrated.
6819 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6820 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6823 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6827 * Permit abbreviated handshakes when renegotiating using the function
6828 SSL_renegotiate_abbreviated().
6830 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6832 * Add call to ENGINE_register_all_complete() to
6833 ENGINE_load_builtin_engines(), so some implementations get used
6834 automatically instead of needing explicit application support.
6838 * Add support for TLS key exporter as described in RFC5705.
6840 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6842 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6843 a few changes are required:
6845 Add SSL_OP_NO_TLSv1_1 flag.
6846 Add TLSv1_1 methods.
6847 Update version checking logic to handle version 1.1.
6848 Add explicit IV handling (ported from DTLS code).
6849 Add command line options to s_client/s_server.
6856 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6858 * X509_ATTRIBUTE memory leak
6860 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6861 memory. This structure is used by the PKCS#7 and CMS routines so any
6862 application which reads PKCS#7 or CMS data from untrusted sources is
6863 affected. SSL/TLS is not affected.
6865 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6871 * Race condition handling PSK identify hint
6873 If PSK identity hints are received by a multi-threaded client then
6874 the values are wrongly updated in the parent SSL_CTX structure. This can
6875 result in a race condition potentially leading to a double free of the
6881 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6883 * Malformed ECParameters causes infinite loop
6885 When processing an ECParameters structure OpenSSL enters an infinite loop
6886 if the curve specified is over a specially malformed binary polynomial
6889 This can be used to perform denial of service against any
6890 system which processes public keys, certificate requests or
6891 certificates. This includes TLS clients and TLS servers with
6892 client authentication enabled.
6894 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6899 * Exploitable out-of-bounds read in X509_cmp_time
6901 X509_cmp_time does not properly check the length of the ASN1_TIME
6902 string and can read a few bytes out of bounds. In addition,
6903 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6906 An attacker can use this to craft malformed certificates and CRLs of
6907 various sizes and potentially cause a segmentation fault, resulting in
6908 a DoS on applications that verify certificates or CRLs. TLS clients
6909 that verify CRLs are affected. TLS clients and servers with client
6910 authentication enabled may be affected if they use custom verification
6913 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6914 independently by Hanno Böck.
6919 * PKCS7 crash with missing EnvelopedContent
6921 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6922 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6923 with missing content and trigger a NULL pointer dereference on parsing.
6925 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6926 structures from untrusted sources are affected. OpenSSL clients and
6927 servers are not affected.
6929 This issue was reported to OpenSSL by Michal Zalewski (Google).
6934 * CMS verify infinite loop with unknown hash function
6936 When verifying a signedData message the CMS code can enter an infinite loop
6937 if presented with an unknown hash function OID. This can be used to perform
6938 denial of service against any system which verifies signedData messages using
6940 This issue was reported to OpenSSL by Johannes Bauer.
6945 * Race condition handling NewSessionTicket
6947 If a NewSessionTicket is received by a multi-threaded client when attempting to
6948 reuse a previous ticket then a race condition can occur potentially leading to
6949 a double free of the ticket data.
6954 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
6956 * Segmentation fault in ASN1_TYPE_cmp fix
6958 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6959 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6960 certificate signature algorithm consistency this can be used to crash any
6961 certificate verification operation and exploited in a DoS attack. Any
6962 application which performs certificate verification is vulnerable including
6963 OpenSSL clients and servers which enable client authentication.
6968 * ASN.1 structure reuse memory corruption fix
6970 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6971 memory corruption via an invalid write. Such reuse is and has been
6972 strongly discouraged and is believed to be rare.
6974 Applications that parse structures containing CHOICE or ANY DEFINED BY
6975 components may be affected. Certificate parsing (d2i_X509 and related
6976 functions) are however not affected. OpenSSL clients and servers are
6982 * PKCS7 NULL pointer dereferences fix
6984 The PKCS#7 parsing code does not handle missing outer ContentInfo
6985 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6986 missing content and trigger a NULL pointer dereference on parsing.
6988 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6989 otherwise parse PKCS#7 structures from untrusted sources are
6990 affected. OpenSSL clients and servers are not affected.
6992 This issue was reported to OpenSSL by Michal Zalewski (Google).
6997 * DoS via reachable assert in SSLv2 servers fix
6999 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7000 servers that both support SSLv2 and enable export cipher suites by sending
7001 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7003 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7004 (OpenSSL development team).
7009 * Use After Free following d2i_ECPrivatekey error fix
7011 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7012 could cause a use after free condition. This, in turn, could cause a double
7013 free in several private key parsing functions (such as d2i_PrivateKey
7014 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7015 for applications that receive EC private keys from untrusted
7016 sources. This scenario is considered rare.
7018 This issue was discovered by the BoringSSL project and fixed in their
7024 * X509_to_X509_REQ NULL pointer deref fix
7026 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7027 the certificate key is invalid. This function is rarely used in practice.
7029 This issue was discovered by Brian Carpenter.
7034 * Removed the export ciphers from the DEFAULT ciphers
7038 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7040 * Build fixes for the Windows and OpenVMS platforms
7042 *Matt Caswell and Richard Levitte*
7044 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7046 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7047 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7048 dereference. This could lead to a Denial Of Service attack. Thanks to
7049 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7054 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7055 dtls1_buffer_record function under certain conditions. In particular this
7056 could occur if an attacker sent repeated DTLS records with the same
7057 sequence number but for the next epoch. The memory leak could be exploited
7058 by an attacker in a Denial of Service attack through memory exhaustion.
7059 Thanks to Chris Mueller for reporting this issue.
7064 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7065 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7066 method would be set to NULL which could later result in a NULL pointer
7067 dereference. Thanks to Frank Schmirler for reporting this issue.
7072 * Abort handshake if server key exchange message is omitted for ephemeral
7075 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7076 reporting this issue.
7081 * Remove non-export ephemeral RSA code on client and server. This code
7082 violated the TLS standard by allowing the use of temporary RSA keys in
7083 non-export ciphersuites and could be used by a server to effectively
7084 downgrade the RSA key length used to a value smaller than the server
7085 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7086 INRIA or reporting this issue.
7091 * Fixed issue where DH client certificates are accepted without verification.
7092 An OpenSSL server will accept a DH certificate for client authentication
7093 without the certificate verify message. This effectively allows a client to
7094 authenticate without the use of a private key. This only affects servers
7095 which trust a client certificate authority which issues certificates
7096 containing DH keys: these are extremely rare and hardly ever encountered.
7097 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7103 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7104 results on some platforms, including x86_64. This bug occurs at random
7105 with a very low probability, and is not known to be exploitable in any
7106 way, though its exact impact is difficult to determine. Thanks to Pieter
7107 Wuille (Blockstream) who reported this issue and also suggested an initial
7108 fix. Further analysis was conducted by the OpenSSL development team and
7109 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7110 the OpenSSL core team.
7115 * Fix various certificate fingerprint issues.
7117 By using non-DER or invalid encodings outside the signed portion of a
7118 certificate the fingerprint can be changed without breaking the signature.
7119 Although no details of the signed portion of the certificate can be changed
7120 this can cause problems with some applications: e.g. those using the
7121 certificate fingerprint for blacklists.
7123 1. Reject signatures with non zero unused bits.
7125 If the BIT STRING containing the signature has non zero unused bits reject
7126 the signature. All current signature algorithms require zero unused bits.
7128 2. Check certificate algorithm consistency.
7130 Check the AlgorithmIdentifier inside TBS matches the one in the
7131 certificate signature. NB: this will result in signature failure
7132 errors for some broken certificates.
7134 Thanks to Konrad Kraszewski from Google for reporting this issue.
7136 3. Check DSA/ECDSA signatures use DER.
7138 Reencode DSA/ECDSA signatures and compare with the original received
7139 signature. Return an error if there is a mismatch.
7141 This will reject various cases including garbage after signature
7142 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7143 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7144 (negative or with leading zeroes).
7146 Further analysis was conducted and fixes were developed by Stephen Henson
7147 of the OpenSSL core team.
7153 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7155 * Session Ticket Memory Leak.
7157 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7158 integrity of that ticket is first verified. In the event of a session
7159 ticket integrity check failing, OpenSSL will fail to free memory
7160 causing a memory leak. By sending a large number of invalid session
7161 tickets an attacker could exploit this issue in a Denial Of Service
7167 * Build option no-ssl3 is incomplete.
7169 When OpenSSL is configured with "no-ssl3" as a build option, servers
7170 could accept and complete a SSL 3.0 handshake, and clients could be
7171 configured to send them.
7174 *Akamai and the OpenSSL team*
7176 * Add support for TLS_FALLBACK_SCSV.
7177 Client applications doing fallback retries should call
7178 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7181 *Adam Langley, Bodo Moeller*
7183 * Add additional DigestInfo checks.
7185 Reencode DigestInto in DER and check against the original when
7186 verifying RSA signature: this will reject any improperly encoded
7187 DigestInfo structures.
7189 Note: this is a precautionary measure and no attacks are currently known.
7193 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7195 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7196 to a denial of service attack. A malicious server can crash the client
7197 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7198 ciphersuite and sending carefully crafted handshake messages.
7200 Thanks to Felix Gröbert (Google) for discovering and researching this
7206 * By sending carefully crafted DTLS packets an attacker could cause openssl
7207 to leak memory. This can be exploited through a Denial of Service attack.
7208 Thanks to Adam Langley for discovering and researching this issue.
7213 * An attacker can force openssl to consume large amounts of memory whilst
7214 processing DTLS handshake messages. This can be exploited through a
7215 Denial of Service attack.
7216 Thanks to Adam Langley for discovering and researching this issue.
7221 * An attacker can force an error condition which causes openssl to crash
7222 whilst processing DTLS packets due to memory being freed twice. This
7223 can be exploited through a Denial of Service attack.
7224 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7230 * If a multithreaded client connects to a malicious server using a resumed
7231 session and the server sends an ec point format extension it could write
7232 up to 255 bytes to freed memory.
7234 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7240 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7241 X509_name_oneline, X509_name_print_ex et al. to leak some information
7242 from the stack. Applications may be affected if they echo pretty printing
7243 output to the attacker.
7245 Thanks to Ivan Fratric (Google) for discovering this issue.
7248 *Emilia Käsper, and Steve Henson*
7250 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7251 for corner cases. (Certain input points at infinity could lead to
7252 bogus results, with non-infinity inputs mapped to infinity too.)
7256 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7258 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7259 handshake can force the use of weak keying material in OpenSSL
7260 SSL/TLS clients and servers.
7262 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7263 researching this issue. [CVE-2014-0224][]
7265 *KIKUCHI Masashi, Steve Henson*
7267 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7268 OpenSSL DTLS client the code can be made to recurse eventually crashing
7271 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7274 *Imre Rad, Steve Henson*
7276 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7277 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7278 client or server. This is potentially exploitable to run arbitrary
7279 code on a vulnerable client or server.
7281 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7283 *Jüri Aedla, Steve Henson*
7285 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7286 are subject to a denial of service attack.
7288 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7289 this issue. [CVE-2014-3470][]
7291 *Felix Gröbert, Ivan Fratric, Steve Henson*
7293 * Harmonize version and its documentation. -f flag is used to display
7296 *mancha <mancha1@zoho.com>*
7298 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7299 in i2d_ECPrivateKey.
7301 *mancha <mancha1@zoho.com>*
7303 * Fix some double frees. These are not thought to be exploitable.
7305 *mancha <mancha1@zoho.com>*
7307 * Fix for the attack described in the paper "Recovering OpenSSL
7308 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7309 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7310 <http://eprint.iacr.org/2014/140>
7312 Thanks to Yuval Yarom and Naomi Benger for discovering this
7313 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7315 *Yuval Yarom and Naomi Benger*
7317 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7319 * Keep original DTLS digest and encryption contexts in retransmission
7320 structures so we can use the previous session parameters if they need
7321 to be resent. [CVE-2013-6450][]
7325 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7326 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7327 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7328 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7329 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7330 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7332 *Rob Stradling, Adam Langley*
7334 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7336 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7338 This addresses the flaw in CBC record processing discovered by
7339 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7340 at: <http://www.isg.rhul.ac.uk/tls/>
7342 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7343 Security Group at Royal Holloway, University of London
7344 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7345 Emilia Käsper for the initial patch.
7348 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7350 * Return an error when checking OCSP signatures when key is NULL.
7351 This fixes a DoS attack. [CVE-2013-0166][]
7355 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7356 the right response is stapled. Also change SSL_get_certificate()
7357 so it returns the certificate actually sent.
7358 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7359 (This is a backport)
7361 *Rob Stradling <rob.stradling@comodo.com>*
7363 * Fix possible deadlock when decoding public keys.
7367 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7369 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7372 * Sanity check record length before skipping explicit IV in DTLS
7375 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7376 fuzzing as a service testing platform.
7381 * Initialise tkeylen properly when encrypting CMS messages.
7382 Thanks to Solar Designer of Openwall for reporting this issue.
7386 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7388 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7389 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7390 in CRYPTO_realloc_clean.
7392 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7393 issue and to Adam Langley <agl@chromium.org> for fixing it.
7396 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7398 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7400 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7401 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7402 content decryption and always return the same error. Note: this attack
7403 needs on average 2^20 messages so it only affects automated senders. The
7404 old behaviour can be re-enabled in the CMS code by setting the
7405 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7406 an MMA defence is not necessary.
7407 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7408 this issue. [CVE-2012-0884][]
7412 * Fix CVE-2011-4619: make sure we really are receiving a
7413 client hello before rejecting multiple SGC restarts. Thanks to
7414 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7418 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7420 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7421 Thanks to Antonio Martin, Enterprise Secure Access Research and
7422 Development, Cisco Systems, Inc. for discovering this bug and
7423 preparing a fix. [CVE-2012-0050][]
7427 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7429 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7430 of the Vaudenay padding oracle attack on CBC mode encryption
7431 which enables an efficient plaintext recovery attack against
7432 the OpenSSL implementation of DTLS. Their attack exploits timing
7433 differences arising during decryption processing. A research
7434 paper describing this attack can be found at:
7435 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7436 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7437 Security Group at Royal Holloway, University of London
7438 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7439 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7440 for preparing the fix. [CVE-2011-4108][]
7442 *Robin Seggelmann, Michael Tuexen*
7444 * Clear bytes used for block padding of SSL 3.0 records.
7447 *Adam Langley (Google)*
7449 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7450 Kadianakis <desnacked@gmail.com> for discovering this issue and
7451 Adam Langley for preparing the fix. [CVE-2011-4619][]
7453 *Adam Langley (Google)*
7455 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7457 *Andrey Kulikov <amdeich@gmail.com>*
7459 * Prevent malformed RFC3779 data triggering an assertion failure.
7460 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7461 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7463 *Rob Austein <sra@hactrn.net>*
7465 * Improved PRNG seeding for VOS.
7467 *Paul Green <Paul.Green@stratus.com>*
7469 * Fix ssl_ciph.c set-up race.
7471 *Adam Langley (Google)*
7473 * Fix spurious failures in ecdsatest.c.
7475 *Emilia Käsper (Google)*
7477 * Fix the BIO_f_buffer() implementation (which was mixing different
7478 interpretations of the `..._len` fields).
7480 *Adam Langley (Google)*
7482 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7483 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7484 threads won't reuse the same blinding coefficients.
7486 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7487 lock to call BN_BLINDING_invert_ex, and avoids one use of
7488 BN_BLINDING_update for each BN_BLINDING structure (previously,
7489 the last update always remained unused).
7491 *Emilia Käsper (Google)*
7493 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7495 *Bob Buckholz (Google)*
7497 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7499 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7500 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7502 *Kaspar Brand <ossl@velox.ch>*
7504 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7505 for multi-threaded use of ECDH. [CVE-2011-3210][]
7507 *Adam Langley (Google)*
7509 * Fix x509_name_ex_d2i memory leak on bad inputs.
7513 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7514 signature public key algorithm by using OID xref utilities instead.
7515 Before this you could only use some ECC ciphersuites with SHA1 only.
7519 * Add protection against ECDSA timing attacks as mentioned in the paper
7520 by Billy Bob Brumley and Nicola Tuveri, see:
7521 <http://eprint.iacr.org/2011/232.pdf>
7523 *Billy Bob Brumley and Nicola Tuveri*
7525 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7527 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7529 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7531 * Fix bug in string printing code: if *any* escaping is enabled we must
7532 escape the escape character (backslash) or the resulting string is
7537 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7539 * Disable code workaround for ancient and obsolete Netscape browsers
7540 and servers: an attacker can use it in a ciphersuite downgrade attack.
7541 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7545 * Fixed J-PAKE implementation error, originally discovered by
7546 Sebastien Martini, further info and confirmation from Stefan
7547 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7551 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7553 * Fix extension code to avoid race conditions which can result in a buffer
7554 overrun vulnerability: resumed sessions must not be modified as they can
7555 be shared by multiple threads. CVE-2010-3864
7559 * Fix WIN32 build system to correctly link an ENGINE directory into
7564 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7566 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7569 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7571 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7573 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7574 context. The operation can be customised via the ctrl mechanism in
7575 case ENGINEs want to include additional functionality.
7579 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7583 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7584 output hashes compatible with older versions of OpenSSL.
7586 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7588 * Fix compression algorithm handling: if resuming a session use the
7589 compression algorithm of the resumed session instead of determining
7590 it from client hello again. Don't allow server to change algorithm.
7594 * Add load_crls() function to apps tidying load_certs() too. Add option
7595 to verify utility to allow additional CRLs to be included.
7599 * Update OCSP request code to permit adding custom headers to the request:
7600 some responders need this.
7604 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7607 *Julia Lawall <julia@diku.dk>*
7609 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7610 needlessly dereferenced structures, used obsolete functions and
7611 didn't handle all updated verify codes correctly.
7615 * Disable MD2 in the default configuration.
7619 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7620 indicate the initial BIO being pushed or popped. This makes it possible
7621 to determine whether the BIO is the one explicitly called or as a result
7622 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7623 it handles reference counts correctly and doesn't zero out the I/O bio
7624 when it is not being explicitly popped. WARNING: applications which
7625 included workarounds for the old buggy behaviour will need to be modified
7626 or they could free up already freed BIOs.
7630 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7631 renaming to all platforms (within the 0.9.8 branch, this was
7632 done conditionally on Netware platforms to avoid a name clash).
7634 *Guenter <lists@gknw.net>*
7636 * Add ECDHE and PSK support to DTLS.
7638 *Michael Tuexen <tuexen@fh-muenster.de>*
7640 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7645 * Add "missing" function EVP_MD_flags() (without this the only way to
7646 retrieve a digest flags is by accessing the structure directly. Update
7647 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7648 or cipher is registered as in the "from" argument. Print out all
7649 registered digests in the dgst usage message instead of manually
7650 attempting to work them out.
7654 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7655 this allows the use of compression and extensions. Change default cipher
7656 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7657 by default unless an application cipher string requests it.
7661 * Alter match criteria in PKCS12_parse(). It used to try to use local
7662 key ids to find matching certificates and keys but some PKCS#12 files
7663 don't follow the (somewhat unwritten) rules and this strategy fails.
7664 Now just gather all certificates together and the first private key
7665 then look for the first certificate that matches the key.
7669 * Support use of registered digest and cipher names for dgst and cipher
7670 commands instead of having to add each one as a special case. So now
7677 openssl dgst -sha256 foo
7679 and this works for ENGINE based algorithms too.
7683 * Update Gost ENGINE to support parameter files.
7685 *Victor B. Wagner <vitus@cryptocom.ru>*
7687 * Support GeneralizedTime in ca utility.
7689 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7691 * Enhance the hash format used for certificate directory links. The new
7692 form uses the canonical encoding (meaning equivalent names will work
7693 even if they aren't identical) and uses SHA1 instead of MD5. This form
7694 is incompatible with the older format and as a result c_rehash should
7695 be used to rebuild symbolic links.
7699 * Make PKCS#8 the default write format for private keys, replacing the
7700 traditional format. This form is standardised, more secure and doesn't
7701 include an implicit MD5 dependency.
7705 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7706 committed to OpenSSL should pass this lot as a minimum.
7710 * Add session ticket override functionality for use by EAP-FAST.
7712 *Jouni Malinen <j@w1.fi>*
7714 * Modify HMAC functions to return a value. Since these can be implemented
7715 in an ENGINE errors can occur.
7719 * Type-checked OBJ_bsearch_ex.
7723 * Type-checked OBJ_bsearch. Also some constification necessitated
7724 by type-checking. Still to come: TXT_DB, bsearch(?),
7725 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7730 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7731 seconds to a tm structure directly, instead of going through OS
7732 specific date routines. This avoids any issues with OS routines such
7733 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7734 and X509_time_adj_ex() to cover the extended range. The existing
7735 X509_time_adj() is still usable and will no longer have any date issues.
7739 * Delta CRL support. New use deltas option which will attempt to locate
7740 and search any appropriate delta CRLs available.
7742 This work was sponsored by Google.
7746 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7747 code and add additional score elements. Validate alternate CRL paths
7748 as part of the CRL checking and indicate a new error "CRL path validation
7749 error" in this case. Applications wanting additional details can use
7750 the verify callback and check the new "parent" field. If this is not
7751 NULL CRL path validation is taking place. Existing applications won't
7752 see this because it requires extended CRL support which is off by
7755 This work was sponsored by Google.
7759 * Support for freshest CRL extension.
7761 This work was sponsored by Google.
7765 * Initial indirect CRL support. Currently only supported in the CRLs
7766 passed directly and not via lookup. Process certificate issuer
7767 CRL entry extension and lookup CRL entries by bother issuer name
7768 and serial number. Check and process CRL issuer entry in IDP extension.
7770 This work was sponsored by Google.
7774 * Add support for distinct certificate and CRL paths. The CRL issuer
7775 certificate is validated separately in this case. Only enabled if
7776 an extended CRL support flag is set: this flag will enable additional
7777 CRL functionality in future.
7779 This work was sponsored by Google.
7783 * Add support for policy mappings extension.
7785 This work was sponsored by Google.
7789 * Fixes to pathlength constraint, self issued certificate handling,
7790 policy processing to align with RFC3280 and PKITS tests.
7792 This work was sponsored by Google.
7796 * Support for name constraints certificate extension. DN, email, DNS
7797 and URI types are currently supported.
7799 This work was sponsored by Google.
7803 * To cater for systems that provide a pointer-based thread ID rather
7804 than numeric, deprecate the current numeric thread ID mechanism and
7805 replace it with a structure and associated callback type. This
7806 mechanism allows a numeric "hash" to be extracted from a thread ID in
7807 either case, and on platforms where pointers are larger than 'long',
7808 mixing is done to help ensure the numeric 'hash' is usable even if it
7809 can't be guaranteed unique. The default mechanism is to use "&errno"
7810 as a pointer-based thread ID to distinguish between threads.
7812 Applications that want to provide their own thread IDs should now use
7813 CRYPTO_THREADID_set_callback() to register a callback that will call
7814 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7816 Note that ERR_remove_state() is now deprecated, because it is tied
7817 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7818 to free the current thread's error state should be replaced by
7819 ERR_remove_thread_state(NULL).
7821 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7822 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7823 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7824 application was previously providing a numeric thread callback that
7825 was inappropriate for distinguishing threads, then uniqueness might
7826 have been obtained with &errno that happened immediately in the
7827 intermediate development versions of OpenSSL; this is no longer the
7828 case, the numeric thread callback will now override the automatic use
7831 *Geoff Thorpe, with help from Bodo Moeller*
7833 * Initial support for different CRL issuing certificates. This covers a
7834 simple case where the self issued certificates in the chain exist and
7835 the real CRL issuer is higher in the existing chain.
7837 This work was sponsored by Google.
7841 * Removed effectively defunct crypto/store from the build.
7845 * Revamp of STACK to provide stronger type-checking. Still to come:
7846 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7847 ASN1_STRING, CONF_VALUE.
7851 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7852 RAM on SSL connections. This option can save about 34k per idle SSL.
7856 * Revamp of LHASH to provide stronger type-checking. Still to come:
7857 STACK, TXT_DB, bsearch, qsort.
7861 * Initial support for Cryptographic Message Syntax (aka CMS) based
7862 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7863 support for data, signedData, compressedData, digestedData and
7864 encryptedData, envelopedData types included. Scripts to check against
7865 RFC4134 examples draft and interop and consistency checks of many
7866 content types and variants.
7870 * Add options to enc utility to support use of zlib compression BIO.
7874 * Extend mk1mf to support importing of options and assembly language
7875 files from Configure script, currently only included in VC-WIN32.
7876 The assembly language rules can now optionally generate the source
7877 files from the associated perl scripts.
7881 * Implement remaining functionality needed to support GOST ciphersuites.
7882 Interop testing has been performed using CryptoPro implementations.
7884 *Victor B. Wagner <vitus@cryptocom.ru>*
7886 * s390x assembler pack.
7890 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7895 * Implement Opaque PRF Input TLS extension as specified in
7896 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7897 official specification yet and no extension type assignment by
7898 IANA exists, this extension (for now) will have to be explicitly
7899 enabled when building OpenSSL by providing the extension number
7900 to use. For example, specify an option
7902 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7904 to the "config" or "Configure" script to enable the extension,
7905 assuming extension number 0x9527 (which is a completely arbitrary
7906 and unofficial assignment based on the MD5 hash of the Internet
7907 Draft). Note that by doing so, you potentially lose
7908 interoperability with other TLS implementations since these might
7909 be using the same extension number for other purposes.
7911 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7912 opaque PRF input value to use in the handshake. This will create
7913 an internal copy of the length-'len' string at 'src', and will
7914 return non-zero for success.
7916 To get more control and flexibility, provide a callback function
7919 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7920 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7924 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7927 Callback function 'cb' will be called in handshakes, and is
7928 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7929 Argument 'arg' is for application purposes (the value as given to
7930 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7931 be provided to the callback function). The callback function
7932 has to return non-zero to report success: usually 1 to use opaque
7933 PRF input just if possible, or 2 to enforce use of the opaque PRF
7934 input. In the latter case, the library will abort the handshake
7935 if opaque PRF input is not successfully negotiated.
7937 Arguments 'peerinput' and 'len' given to the callback function
7938 will always be NULL and 0 in the case of a client. A server will
7939 see the client's opaque PRF input through these variables if
7940 available (NULL and 0 otherwise). Note that if the server
7941 provides an opaque PRF input, the length must be the same as the
7942 length of the client's opaque PRF input.
7944 Note that the callback function will only be called when creating
7945 a new session (session resumption can resume whatever was
7946 previously negotiated), and will not be called in SSL 2.0
7947 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7948 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7949 for applications that need to enforce opaque PRF input.
7953 * Update ssl code to support digests other than SHA1+MD5 for handshake
7956 *Victor B. Wagner <vitus@cryptocom.ru>*
7958 * Add RFC4507 support to OpenSSL. This includes the corrections in
7959 RFC4507bis. The encrypted ticket format is an encrypted encoded
7960 SSL_SESSION structure, that way new session features are automatically
7963 If a client application caches session in an SSL_SESSION structure
7964 support is transparent because tickets are now stored in the encoded
7967 The SSL_CTX structure automatically generates keys for ticket
7968 protection in servers so again support should be possible
7969 with no application modification.
7971 If a client or server wishes to disable RFC4507 support then the option
7972 SSL_OP_NO_TICKET can be set.
7974 Add a TLS extension debugging callback to allow the contents of any client
7975 or server extensions to be examined.
7977 This work was sponsored by Google.
7981 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7982 OpenSSL should now compile cleanly on gcc 4.2
7984 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7986 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7987 support including streaming MAC support: this is required for GOST
7988 ciphersuite support.
7990 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7992 * Add option -stream to use PKCS#7 streaming in smime utility. New
7993 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7994 to output in BER and PEM format.
7998 * Experimental support for use of HMAC via EVP_PKEY interface. This
7999 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8000 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8001 ENGINE support for HMAC keys which are unextractable. New -mac and
8002 -macopt options to dgst utility.
8006 * New option -sigopt to dgst utility. Update dgst to use
8007 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8008 alternative signing parameters such as X9.31 or PSS in the dgst
8013 * Change ssl_cipher_apply_rule(), the internal function that does
8014 the work each time a ciphersuite string requests enabling
8015 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8016 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8017 the order of disabled ciphersuites such that those ciphersuites
8018 that most recently went from enabled to disabled not only stay
8019 in order with respect to each other, but also have higher priority
8020 than other disabled ciphersuites the next time ciphersuites are
8023 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8024 the same ciphersuites as with "HIGH" alone, but in a specific
8025 order where the PSK ciphersuites come first (since they are the
8026 most recently disabled ciphersuites when "HIGH" is parsed).
8028 Also, change ssl_create_cipher_list() (using this new
8029 functionality) such that between otherwise identical
8030 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8035 * Change ssl_create_cipher_list() so that it automatically
8036 arranges the ciphersuites in reasonable order before starting
8037 to process the rule string. Thus, the definition for "DEFAULT"
8038 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8039 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8040 This makes it much easier to arrive at a reasonable default order
8041 in applications for which anonymous ciphers are OK (meaning
8042 that you can't actually use DEFAULT).
8044 *Bodo Moeller; suggested by Victor Duchovni*
8046 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8047 processing) into multiple integers instead of setting
8048 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8049 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8050 (These masks as well as the individual bit definitions are hidden
8051 away into the non-exported interface ssl/ssl_locl.h, so this
8052 change to the definition of the SSL_CIPHER structure shouldn't
8053 affect applications.) This give us more bits for each of these
8054 categories, so there is no longer a need to coagulate AES128 and
8055 AES256 into a single algorithm bit, and to coagulate Camellia128
8056 and Camellia256 into a single algorithm bit, which has led to all
8059 Thus, among other things, the kludge introduced in 0.9.7m and
8060 0.9.8e for masking out AES256 independently of AES128 or masking
8061 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8063 With the change, we also introduce new ciphersuite aliases that
8064 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8069 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8070 Use the leftmost N bytes of the signature input if the input is
8071 larger than the prime q (with N being the size in bytes of q).
8075 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8076 it yet and it is largely untested.
8080 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8084 * Initial incomplete changes to avoid need for function casts in OpenSSL
8085 some compilers (gcc 4.2 and later) reject their use. Safestack is
8086 reimplemented. Update ASN1 to avoid use of legacy functions.
8090 * Win32/64 targets are linked with Winsock2.
8094 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8095 to external functions. This can be used to increase CRL handling
8096 efficiency especially when CRLs are very large by (for example) storing
8097 the CRL revoked certificates in a database.
8101 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8102 new CRLs added to a directory can be used. New command line option
8103 -verify_return_error to s_client and s_server. This causes real errors
8104 to be returned by the verify callback instead of carrying on no matter
8105 what. This reflects the way a "real world" verify callback would behave.
8109 * GOST engine, supporting several GOST algorithms and public key formats.
8110 Kindly donated by Cryptocom.
8114 * Partial support for Issuing Distribution Point CRL extension. CRLs
8115 partitioned by DP are handled but no indirect CRL or reason partitioning
8116 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8117 selected via a scoring technique which handles IDP and AKID in CRLs.
8121 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8122 will ultimately be used for all verify operations: this will remove the
8123 X509_STORE dependency on certificate verification and allow alternative
8124 lookup methods. X509_STORE based implementations of these two callbacks.
8128 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8129 Modify get_crl() to find a valid (unexpired) CRL if possible.
8133 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8134 this would be called X509_CRL_cmp() but that name is already used by
8135 a function that just compares CRL issuer names. Cache several CRL
8136 extensions in X509_CRL structure and cache CRLDP in X509.
8140 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8141 this maps equivalent X509_NAME structures into a consistent structure.
8142 Name comparison can then be performed rapidly using memcmp().
8146 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8151 * Allow digests to supply their own micalg string for S/MIME type using
8152 the ctrl EVP_MD_CTRL_MICALG.
8156 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8157 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8158 ctrl. It can then customise the structure before and/or after signing
8163 * New function OBJ_add_sigid() to allow application defined signature OIDs
8164 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8165 to free up any added signature OIDs.
8169 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8170 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8171 digest and cipher tables. New options added to openssl utility:
8172 list-message-digest-algorithms and list-cipher-algorithms.
8176 * Change the array representation of binary polynomials: the list
8177 of degrees of non-zero coefficients is now terminated with -1.
8178 Previously it was terminated with 0, which was also part of the
8179 value; thus, the array representation was not applicable to
8180 polynomials where t^0 has coefficient zero. This change makes
8181 the array representation useful in a more general context.
8185 * Various modifications and fixes to SSL/TLS cipher string
8186 handling. For ECC, the code now distinguishes between fixed ECDH
8187 with RSA certificates on the one hand and with ECDSA certificates
8188 on the other hand, since these are separate ciphersuites. The
8189 unused code for Fortezza ciphersuites has been removed.
8191 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8192 (not "ECDHE"). For consistency with the code for DH
8193 certificates, use of ECDH certificates is now considered ECDH
8194 authentication, not RSA or ECDSA authentication (the latter is
8195 merely the CA's signing algorithm and not actively used in the
8198 The temporary ciphersuite alias "ECCdraft" is no longer
8199 available, and ECC ciphersuites are no longer excluded from "ALL"
8200 and "DEFAULT". The following aliases now exist for RFC 4492
8201 ciphersuites, most of these by analogy with the DH case:
8203 kECDHr - ECDH cert, signed with RSA
8204 kECDHe - ECDH cert, signed with ECDSA
8205 kECDH - ECDH cert (signed with either RSA or ECDSA)
8206 kEECDH - ephemeral ECDH
8207 ECDH - ECDH cert or ephemeral ECDH
8213 AECDH - anonymous ECDH
8214 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8218 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8219 Use correct micalg parameters depending on digest(s) in signed message.
8223 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8224 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8228 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8229 an engine to register a method. Add ENGINE lookups for methods and
8230 functional reference processing.
8234 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8235 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8240 * New -resign option to smime utility. This adds one or more signers
8241 to an existing PKCS#7 signedData structure. Also -md option to use an
8242 alternative message digest algorithm for signing.
8246 * Tidy up PKCS#7 routines and add new functions to make it easier to
8247 create PKCS7 structures containing multiple signers. Update smime
8248 application to support multiple signers.
8252 * New -macalg option to pkcs12 utility to allow setting of an alternative
8257 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8258 Reorganize PBE internals to lookup from a static table using NIDs,
8259 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8260 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8261 PRF which will be automatically used with PBES2.
8265 * Replace the algorithm specific calls to generate keys in "req" with the
8270 * Update PKCS#7 enveloped data routines to use new API. This is now
8271 supported by any public key method supporting the encrypt operation. A
8272 ctrl is added to allow the public key algorithm to examine or modify
8273 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8278 * Add a ctrl to asn1 method to allow a public key algorithm to express
8279 a default digest type to use. In most cases this will be SHA1 but some
8280 algorithms (such as GOST) need to specify an alternative digest. The
8281 return value indicates how strong the preference is 1 means optional and
8282 2 is mandatory (that is it is the only supported type). Modify
8283 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8284 use the default md. Update openssl utilities to use the default digest
8285 type for signing if it is not explicitly indicated.
8289 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8290 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8291 signing method from the key type. This effectively removes the link
8292 between digests and public key types.
8296 * Add an OID cross reference table and utility functions. Its purpose is to
8297 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8298 rsaEncryption. This will allow some of the algorithm specific hackery
8299 needed to use the correct OID to be removed.
8303 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8304 structures for PKCS7_sign(). They are now set up by the relevant public
8309 * Add provisional EC pkey method with support for ECDSA and ECDH.
8313 * Add support for key derivation (agreement) in the API, DH method and
8318 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8319 public and private key formats. As a side effect these add additional
8320 command line functionality not previously available: DSA signatures can be
8321 generated and verified using pkeyutl and DH key support and generation in
8328 *Oliver Tappe <zooey@hirschkaefer.de>*
8330 * New make target "install_html_docs" installs HTML renditions of the
8333 *Oliver Tappe <zooey@hirschkaefer.de>*
8335 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8336 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8337 support key and parameter generation and add initial key generation
8338 functionality for RSA.
8342 * Add functions for main EVP_PKEY_method operations. The undocumented
8343 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8344 `EVP_PKEY_{encrypt,decrypt}_old`.
8348 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8349 key API, doesn't do much yet.
8353 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8354 public key algorithms. New option to openssl utility:
8355 "list-public-key-algorithms" to print out info.
8359 * Implement the Supported Elliptic Curves Extension for
8360 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8364 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8365 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8369 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8370 utilities such as rsa, dsa, dsaparam etc except they process any key
8375 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8376 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8377 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8382 * Initial support for pluggable public key ASN1.
8383 De-spaghettify the public key ASN1 handling. Move public and private
8384 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8385 algorithm specific handling to a single module within the relevant
8386 algorithm directory. Add functions to allow (near) opaque processing
8387 of public and private key structures.
8391 * Implement the Supported Point Formats Extension for
8392 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8396 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8397 for the psk identity [hint] and the psk callback functions to the
8398 SSL_SESSION, SSL and SSL_CTX structure.
8401 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8405 SSL_CTX_use_psk_identity_hint
8406 SSL_get_psk_identity_hint
8407 SSL_get_psk_identity
8408 SSL_use_psk_identity_hint
8410 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8412 * Add RFC 3161 compliant time stamp request creation, response generation
8413 and response verification functionality.
8415 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8417 * Add initial support for TLS extensions, specifically for the server_name
8418 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8419 have new members for a host name. The SSL data structure has an
8420 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8421 stored in that context to allow for session resumption, even after the
8422 SSL has been switched to a new SSL_CTX in reaction to a client's
8423 server_name extension.
8425 New functions (subject to change):
8427 SSL_get_servername()
8428 SSL_get_servername_type()
8431 New CTRL codes and macros (subject to change):
8433 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8434 - SSL_CTX_set_tlsext_servername_callback()
8435 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8436 - SSL_CTX_set_tlsext_servername_arg()
8437 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8439 openssl s_client has a new '-servername ...' option.
8441 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8442 '-key2 ...', '-servername_fatal' (subject to change). This allows
8443 testing the HostName extension for a specific single host name ('-cert'
8444 and '-key' remain fallbacks for handshakes without HostName
8445 negotiation). If the unrecognized_name alert has to be sent, this by
8446 default is a warning; it becomes fatal with the '-servername_fatal'
8449 *Peter Sylvester, Remy Allais, Christophe Renou*
8451 * Whirlpool hash implementation is added.
8455 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8456 bn(64,32). Because of instruction set limitations it doesn't have
8457 any negative impact on performance. This was done mostly in order
8458 to make it possible to share assembler modules, such as bn_mul_mont
8459 implementations, between 32- and 64-bit builds without hassle.
8463 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8464 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8469 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8470 dedicated Montgomery multiplication procedure, is introduced.
8471 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8472 "64-bit" performance on certain 32-bit targets.
8476 * New option SSL_OP_NO_COMP to disable use of compression selectively
8477 in SSL structures. New SSL ctrl to set maximum send fragment size.
8478 Save memory by setting the I/O buffer sizes dynamically instead of
8479 using the maximum available value.
8483 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8484 in addition to the text details.
8488 * Very, very preliminary EXPERIMENTAL support for printing of general
8489 ASN1 structures. This currently produces rather ugly output and doesn't
8490 handle several customised structures at all.
8494 * Integrated support for PVK file format and some related formats such
8495 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8496 these in the 'rsa' and 'dsa' utilities.
8500 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8504 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8505 place for the (very old) "NETSCAPE" format certificates which are now
8506 handled using new ASN1 code equivalents.
8510 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8511 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8512 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8516 * Modify CRL distribution points extension code to print out previously
8517 unsupported fields. Enhance extension setting code to allow setting of
8522 * Add print and set support for Issuing Distribution Point CRL extension.
8526 * Change 'Configure' script to enable Camellia by default.
8533 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8535 * When rejecting SSL/TLS records due to an incorrect version number, never
8536 update s->server with a new major version number. As of
8537 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8538 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8539 the previous behavior could result in a read attempt at NULL when
8540 receiving specific incorrect SSL/TLS records once record payload
8541 protection is active. [CVE-2010-0740][]
8543 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8545 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8546 could be crashed if the relevant tables were not present (e.g. chrooted).
8548 *Tomas Hoger <thoger@redhat.com>*
8550 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8552 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8554 *Martin Olsson, Neel Mehta*
8556 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8557 accommodate for stack sorting, always a write lock!).
8561 * On some versions of WIN32 Heap32Next is very slow. This can cause
8562 excessive delays in the RAND_poll(): over a minute. As a workaround
8563 include a time check in the inner Heap32Next loop too.
8567 * The code that handled flushing of data in SSL/TLS originally used the
8568 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8569 the problem outlined in PR#1949. The fix suggested there however can
8570 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8571 of Apache). So instead simplify the code to flush unconditionally.
8572 This should be fine since flushing with no data to flush is a no op.
8576 * Handle TLS versions 2.0 and later properly and correctly use the
8577 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8578 off ancient servers have a habit of sticking around for a while...
8582 * Modify compression code so it frees up structures without using the
8583 ex_data callbacks. This works around a problem where some applications
8584 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8585 restarting) then use compression (e.g. SSL with compression) later.
8586 This results in significant per-connection memory leaks and
8587 has caused some security issues including CVE-2008-1678 and
8592 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8593 change when encrypting or decrypting.
8597 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8598 connect and renegotiate with servers which do not support RI.
8599 Until RI is more widely deployed this option is enabled by default.
8603 * Add "missing" ssl ctrls to clear options and mode.
8607 * If client attempts to renegotiate and doesn't support RI respond with
8608 a no_renegotiation alert as required by RFC5746. Some renegotiating
8609 TLS clients will continue a connection gracefully when they receive
8610 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8611 waiting for a server hello which it will never receive. Now we treat a
8612 received no_renegotiation alert as a fatal error. This is because
8613 applications requesting a renegotiation might well expect it to succeed
8614 and would have no code in place to handle the server denying it so the
8615 only safe thing to do is to terminate the connection.
8619 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8620 peer supports secure renegotiation and 0 otherwise. Print out peer
8621 renegotiation support in s_client/s_server.
8625 * Replace the highly broken and deprecated SPKAC certification method with
8626 the updated NID creation version. This should correctly handle UTF8.
8630 * Implement RFC5746. Re-enable renegotiation but require the extension
8631 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8632 turns out to be a bad idea. It has been replaced by
8633 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8634 SSL_CTX_set_options(). This is really not recommended unless you
8635 know what you are doing.
8637 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8639 * Fixes to stateless session resumption handling. Use initial_ctx when
8640 issuing and attempting to decrypt tickets in case it has changed during
8641 servername handling. Use a non-zero length session ID when attempting
8642 stateless session resumption: this makes it possible to determine if
8643 a resumption has occurred immediately after receiving server hello
8644 (several places in OpenSSL subtly assume this) instead of later in
8649 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8650 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8651 fixes for a few places where the return code is not checked
8654 *Julia Lawall <julia@diku.dk>*
8656 * Add --strict-warnings option to Configure script to include devteam
8657 warnings in other configurations.
8661 * Add support for --libdir option and LIBDIR variable in makefiles. This
8662 makes it possible to install openssl libraries in locations which
8663 have names other than "lib", for example "/usr/lib64" which some
8666 *Steve Henson, based on patch from Jeremy Utley*
8668 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8669 X690 8.9.12 and can produce some misleading textual output of OIDs.
8671 *Steve Henson, reported by Dan Kaminsky*
8673 * Delete MD2 from algorithm tables. This follows the recommendation in
8674 several standards that it is not used in new applications due to
8675 several cryptographic weaknesses. For binary compatibility reasons
8676 the MD2 API is still compiled in by default.
8680 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8685 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8686 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8689 *Guenter <lists@gknw.net>*
8691 * Fix the server certificate chain building code to use X509_verify_cert(),
8692 it used to have an ad-hoc builder which was unable to cope with anything
8693 other than a simple chain.
8695 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8697 * Don't check self signed certificate signatures in X509_verify_cert()
8698 by default (a flag can override this): it just wastes time without
8699 adding any security. As a useful side effect self signed root CAs
8700 with non-FIPS digests are now usable in FIPS mode.
8704 * In dtls1_process_out_of_seq_message() the check if the current message
8705 is already buffered was missing. For every new message was memory
8706 allocated, allowing an attacker to perform an denial of service attack
8707 with sending out of seq handshake messages until there is no memory
8708 left. Additionally every future message was buffered, even if the
8709 sequence number made no sense and would be part of another handshake.
8710 So only messages with sequence numbers less than 10 in advance will be
8711 buffered. [CVE-2009-1378][]
8713 *Robin Seggelmann, discovered by Daniel Mentz*
8715 * Records are buffered if they arrive with a future epoch to be
8716 processed after finishing the corresponding handshake. There is
8717 currently no limitation to this buffer allowing an attacker to perform
8718 a DOS attack with sending records with future epochs until there is no
8719 memory left. This patch adds the pqueue_size() function to determine
8720 the size of a buffer and limits the record buffer to 100 entries.
8723 *Robin Seggelmann, discovered by Daniel Mentz*
8725 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8726 parent structure is freed. [CVE-2009-1379][]
8730 * Handle non-blocking I/O properly in SSL_shutdown() call.
8732 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8734 * Add `2.5.4.*` OIDs
8736 *Ilya O. <vrghost@gmail.com>*
8738 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8740 * Disable renegotiation completely - this fixes a severe security
8741 problem [CVE-2009-3555][] at the cost of breaking all
8742 renegotiation. Renegotiation can be re-enabled by setting
8743 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8744 run-time. This is really not recommended unless you know what
8749 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8751 * Don't set val to NULL when freeing up structures, it is freed up by
8752 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8753 zeroing past the valid field. [CVE-2009-0789][]
8755 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8757 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8758 checked correctly. This would allow some invalid signed attributes to
8759 appear to verify correctly. [CVE-2009-0591][]
8761 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8763 * Reject UniversalString and BMPString types with invalid lengths. This
8764 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8765 a legal length. [CVE-2009-0590][]
8769 * Set S/MIME signing as the default purpose rather than setting it
8770 unconditionally. This allows applications to override it at the store
8775 * Permit restricted recursion of ASN1 strings. This is needed in practice
8776 to handle some structures.
8780 * Improve efficiency of mem_gets: don't search whole buffer each time
8783 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8785 * New -hex option for openssl rand.
8789 * Print out UTF8String and NumericString when parsing ASN1.
8793 * Support NumericString type for name components.
8797 * Allow CC in the environment to override the automatically chosen
8798 compiler. Note that nothing is done to ensure flags work with the
8803 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8805 * Properly check EVP_VerifyFinal() and similar return values
8808 *Ben Laurie, Bodo Moeller, Google Security Team*
8810 * Enable TLS extensions by default.
8814 * Allow the CHIL engine to be loaded, whether the application is
8815 multithreaded or not. (This does not release the developer from the
8816 obligation to set up the dynamic locking callbacks.)
8818 *Sander Temme <sander@temme.net>*
8820 * Use correct exit code if there is an error in dgst command.
8822 *Steve Henson; problem pointed out by Roland Dirlewanger*
8824 * Tweak Configure so that you need to say "experimental-jpake" to enable
8825 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8829 * Add experimental JPAKE support, including demo authentication in
8830 s_client and s_server.
8834 * Set the comparison function in v3_addr_canonize().
8836 *Rob Austein <sra@hactrn.net>*
8838 * Add support for XMPP STARTTLS in s_client.
8840 *Philip Paeps <philip@freebsd.org>*
8842 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8843 to ensure that even with this option, only ciphersuites in the
8844 server's preference list will be accepted. (Note that the option
8845 applies only when resuming a session, so the earlier behavior was
8846 just about the algorithm choice for symmetric cryptography.)
8850 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8852 * Fix NULL pointer dereference if a DTLS server received
8853 ChangeCipherSpec as first record [CVE-2009-1386][].
8857 * Fix a state transition in s3_srvr.c and d1_srvr.c
8858 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8862 * The fix in 0.9.8c that supposedly got rid of unsafe
8863 double-checked locking was incomplete for RSA blinding,
8864 addressing just one layer of what turns out to have been
8865 doubly unsafe triple-checked locking.
8867 So now fix this for real by retiring the MONT_HELPER macro
8868 in crypto/rsa/rsa_eay.c.
8870 *Bodo Moeller; problem pointed out by Marius Schilder*
8872 * Various precautionary measures:
8874 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8876 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8877 (NB: This would require knowledge of the secret session ticket key
8878 to exploit, in which case you'd be SOL either way.)
8880 - Change bn_nist.c so that it will properly handle input BIGNUMs
8881 outside the expected range.
8883 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8886 *Neel Mehta, Bodo Moeller*
8888 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8889 the load fails. Useful for distros.
8891 *Ben Laurie and the FreeBSD team*
8893 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8897 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8901 * Expand ENGINE to support engine supplied SSL client certificate functions.
8903 This work was sponsored by Logica.
8907 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8908 keystores. Support for SSL/TLS client authentication too.
8909 Not compiled unless enable-capieng specified to Configure.
8911 This work was sponsored by Logica.
8915 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8916 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8917 attribute creation routines such as certificate requests and PKCS#12
8922 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8924 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8925 handshake which could lead to a client crash as found using the
8926 Codenomicon TLS test suite [CVE-2008-1672][]
8928 *Steve Henson, Mark Cox*
8930 * Fix double free in TLS server name extensions which could lead to
8931 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8935 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8937 Clear the error queue to ensure that error entries left from
8938 older function calls do not interfere with the correct operation.
8940 *Lutz Jaenicke, Erik de Castro Lopo*
8942 * Remove root CA certificates of commercial CAs:
8944 The OpenSSL project does not recommend any specific CA and does not
8945 have any policy with respect to including or excluding any CA.
8946 Therefore it does not make any sense to ship an arbitrary selection
8947 of root CA certificates with the OpenSSL software.
8951 * RSA OAEP patches to fix two separate invalid memory reads.
8952 The first one involves inputs when 'lzero' is greater than
8953 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8954 before the beginning of from). The second one involves inputs where
8955 the 'db' section contains nothing but zeroes (there is a one-byte
8956 invalid read after the end of 'db').
8958 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8960 * Partial backport from 0.9.9-dev:
8962 Introduce bn_mul_mont (dedicated Montgomery multiplication
8963 procedure) as a candidate for BIGNUM assembler implementation.
8964 While 0.9.9-dev uses assembler for various architectures, only
8965 x86_64 is available by default here in the 0.9.8 branch, and
8966 32-bit x86 is available through a compile-time setting.
8968 To try the 32-bit x86 assembler implementation, use Configure
8969 option "enable-montasm" (which exists only for this backport).
8971 As "enable-montasm" for 32-bit x86 disclaims code stability
8972 anyway, in this constellation we activate additional code
8973 backported from 0.9.9-dev for further performance improvements,
8974 namely BN_from_montgomery_word. (To enable this otherwise,
8975 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8977 *Andy Polyakov (backport partially by Bodo Moeller)*
8979 * Add TLS session ticket callback. This allows an application to set
8980 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8981 values. This is useful for key rollover for example where several key
8982 sets may exist with different names.
8986 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8987 This was broken until now in 0.9.8 releases, such that the only way
8988 a registered ENGINE could be used (assuming it initialises
8989 successfully on the host) was to explicitly set it as the default
8990 for the relevant algorithms. This is in contradiction with 0.9.7
8991 behaviour and the documentation. With this fix, when an ENGINE is
8992 registered into a given algorithm's table of implementations, the
8993 'uptodate' flag is reset so that auto-discovery will be used next
8994 time a new context for that algorithm attempts to select an
8997 *Ian Lister (tweaked by Geoff Thorpe)*
8999 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9000 implementation in the following ways:
9002 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9005 Lack of BER streaming support means one pass streaming processing is
9006 only supported if data is detached: setting the streaming flag is
9007 ignored for embedded content.
9009 CMS support is disabled by default and must be explicitly enabled
9010 with the enable-cms configuration option.
9014 * Update the GMP engine glue to do direct copies between BIGNUM and
9015 mpz_t when openssl and GMP use the same limb size. Otherwise the
9016 existing "conversion via a text string export" trick is still used.
9018 *Paul Sheer <paulsheer@gmail.com>*
9020 * Zlib compression BIO. This is a filter BIO which compressed and
9021 uncompresses any data passed through it.
9025 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9026 RFC3394 compatible AES key wrapping.
9030 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9031 sets string data without copying. X509_ALGOR_set0() and
9032 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9033 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9034 from an X509_ATTRIBUTE structure optionally checking it occurs only
9035 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9040 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9041 to get the expected BN_FLG_CONSTTIME behavior.
9043 *Bodo Moeller (Google)*
9047 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9048 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9049 - added some more tests to do_tests.pl
9050 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9051 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9052 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9053 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9054 - various changes to netware.pl to enable gcc-cross builds on Win32
9056 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9057 - various changes to fix missing prototype warnings
9058 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9059 - added AES, WHIRLPOOL and CPUID assembler code to build files
9060 - added missing AES assembler make rules to mk1mf.pl
9061 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9063 *Guenter Knauf <eflash@gmx.net>*
9065 * Implement certificate status request TLS extension defined in RFC3546.
9066 A client can set the appropriate parameters and receive the encoded
9067 OCSP response via a callback. A server can query the supplied parameters
9068 and set the encoded OCSP response in the callback. Add simplified examples
9069 to s_client and s_server.
9073 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9076 + Binary incompatibility of ssl_ctx_st structure
9077 + DTLS interoperation with non-compliant servers
9078 + Don't call get_session_cb() without proposed session
9079 + Fix ia64 assembler code
9081 *Andy Polyakov, Steve Henson*
9083 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9085 * DTLS Handshake overhaul. There were longstanding issues with
9086 OpenSSL DTLS implementation, which were making it impossible for
9087 RFC 4347 compliant client to communicate with OpenSSL server.
9088 Unfortunately just fixing these incompatibilities would "cut off"
9089 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9090 server keeps tolerating non RFC compliant syntax. The opposite is
9091 not true, 0.9.8f client can not communicate with earlier server.
9092 This update even addresses CVE-2007-4995.
9096 * Changes to avoid need for function casts in OpenSSL: some compilers
9097 (gcc 4.2 and later) reject their use.
9098 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9101 * Add RFC4507 support to OpenSSL. This includes the corrections in
9102 RFC4507bis. The encrypted ticket format is an encrypted encoded
9103 SSL_SESSION structure, that way new session features are automatically
9106 If a client application caches session in an SSL_SESSION structure
9107 support is transparent because tickets are now stored in the encoded
9110 The SSL_CTX structure automatically generates keys for ticket
9111 protection in servers so again support should be possible
9112 with no application modification.
9114 If a client or server wishes to disable RFC4507 support then the option
9115 SSL_OP_NO_TICKET can be set.
9117 Add a TLS extension debugging callback to allow the contents of any client
9118 or server extensions to be examined.
9120 This work was sponsored by Google.
9124 * Add initial support for TLS extensions, specifically for the server_name
9125 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9126 have new members for a host name. The SSL data structure has an
9127 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9128 stored in that context to allow for session resumption, even after the
9129 SSL has been switched to a new SSL_CTX in reaction to a client's
9130 server_name extension.
9132 New functions (subject to change):
9134 SSL_get_servername()
9135 SSL_get_servername_type()
9138 New CTRL codes and macros (subject to change):
9140 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9141 - SSL_CTX_set_tlsext_servername_callback()
9142 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9143 - SSL_CTX_set_tlsext_servername_arg()
9144 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9146 openssl s_client has a new '-servername ...' option.
9148 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9149 '-key2 ...', '-servername_fatal' (subject to change). This allows
9150 testing the HostName extension for a specific single host name ('-cert'
9151 and '-key' remain fallbacks for handshakes without HostName
9152 negotiation). If the unrecognized_name alert has to be sent, this by
9153 default is a warning; it becomes fatal with the '-servername_fatal'
9156 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9158 * Add AES and SSE2 assembly language support to VC++ build.
9162 * Mitigate attack on final subtraction in Montgomery reduction.
9166 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9167 (which previously caused an internal error).
9171 * Squeeze another 10% out of IGE mode when in != out.
9175 * AES IGE mode speedup.
9177 *Dean Gaudet (Google)*
9179 * Add the Korean symmetric 128-bit cipher SEED (see
9180 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9181 add SEED ciphersuites from RFC 4162:
9183 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9184 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9185 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9186 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9188 To minimize changes between patchlevels in the OpenSSL 0.9.8
9189 series, SEED remains excluded from compilation unless OpenSSL
9190 is configured with 'enable-seed'.
9192 *KISA, Bodo Moeller*
9194 * Mitigate branch prediction attacks, which can be practical if a
9195 single processor is shared, allowing a spy process to extract
9196 information. For detailed background information, see
9197 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9198 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9199 and Necessary Software Countermeasures"). The core of the change
9200 are new versions BN_div_no_branch() and
9201 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9202 respectively, which are slower, but avoid the security-relevant
9203 conditional branches. These are automatically called by BN_div()
9204 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9205 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9206 remove a conditional branch.
9208 BN_FLG_CONSTTIME is the new name for the previous
9209 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9210 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9211 in the exponent causes BN_mod_exp_mont() to use the alternative
9212 implementation in BN_mod_exp_mont_consttime().) The old name
9213 remains as a deprecated alias.
9215 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9216 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9217 constant-time implementations for more than just exponentiation.
9218 Here too the old name is kept as a deprecated alias.
9220 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9221 the BN_BLINDING structure gets an independent copy of the
9222 modulus. This means that the previous `BIGNUM *m` argument to
9223 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9224 essentially becomes `const BIGNUM *m`, although we can't actually
9225 change this in the header file before 0.9.9. It allows
9226 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9227 enable BN_FLG_CONSTTIME.
9229 *Matthew D Wood (Intel Corp)*
9231 * In the SSL/TLS server implementation, be strict about session ID
9232 context matching (which matters if an application uses a single
9233 external cache for different purposes). Previously,
9234 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9235 set. This did ensure strict client verification, but meant that,
9236 with applications using a single external cache for quite
9237 different requirements, clients could circumvent ciphersuite
9238 restrictions for a given session ID context by starting a session
9239 in a different context.
9243 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9244 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9245 authentication-only ciphersuites.
9249 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9250 not complete and could lead to a possible single byte overflow
9251 [CVE-2007-5135][] [Ben Laurie]
9253 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9255 * Since AES128 and AES256 (and similarly Camellia128 and
9256 Camellia256) share a single mask bit in the logic of
9257 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9258 kludge to work properly if AES128 is available and AES256 isn't
9259 (or if Camellia128 is available and Camellia256 isn't).
9263 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9264 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9265 When a point or a seed is encoded in a BIT STRING, we need to
9266 prevent the removal of trailing zero bits to get the proper DER
9267 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9268 of a NamedBitList, for which trailing 0 bits need to be removed.)
9272 * Have SSL/TLS server implementation tolerate "mismatched" record
9273 protocol version while receiving ClientHello even if the
9274 ClientHello is fragmented. (The server can't insist on the
9275 particular protocol version it has chosen before the ServerHello
9276 message has informed the client about his choice.)
9280 * Add RFC 3779 support.
9282 *Rob Austein for ARIN, Ben Laurie*
9284 * Load error codes if they are not already present instead of using a
9285 static variable. This allows them to be cleanly unloaded and reloaded.
9286 Improve header file function name parsing.
9290 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9291 or CAPABILITY handshake as required by RFCs.
9295 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9297 * Introduce limits to prevent malicious keys being able to
9298 cause a denial of service. [CVE-2006-2940][]
9300 *Steve Henson, Bodo Moeller*
9302 * Fix ASN.1 parsing of certain invalid structures that can result
9303 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9305 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9306 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9308 * Fix SSL client code which could crash if connecting to a
9309 malicious SSLv2 server. [CVE-2006-4343][]
9311 *Tavis Ormandy and Will Drewry, Google Security Team*
9313 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9314 match only those. Before that, "AES256-SHA" would be interpreted
9315 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9316 the same strength classification in 0.9.7h) as we currently only
9317 have a single AES bit in the ciphersuite description bitmap.
9318 That change, however, also applied to ciphersuite strings such as
9319 "RC4-MD5" that intentionally matched multiple ciphersuites --
9320 namely, SSL 2.0 ciphersuites in addition to the more common ones
9321 from SSL 3.0/TLS 1.0.
9323 So we change the selection algorithm again: Naming an explicit
9324 ciphersuite selects this one ciphersuite, and any other similar
9325 ciphersuite (same bitmap) from *other* protocol versions.
9326 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9327 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9329 Since SSL 2.0 does not have any ciphersuites for which the
9330 128/256 bit distinction would be relevant, this works for now.
9331 The proper fix will be to use different bits for AES128 and
9332 AES256, which would have avoided the problems from the beginning;
9333 however, bits are scarce, so we can only do this in a new release
9334 (not just a patchlevel) when we can change the SSL_CIPHER
9335 definition to split the single 'unsigned long mask' bitmap into
9336 multiple values to extend the available space.
9340 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9342 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9343 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9345 * Add AES IGE and biIGE modes.
9349 * Change the Unix randomness entropy gathering to use poll() when
9350 possible instead of select(), since the latter has some
9351 undesirable limitations.
9353 *Darryl Miles via Richard Levitte and Bodo Moeller*
9355 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9356 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9357 cannot be implicitly activated as part of, e.g., the "AES" alias.
9358 However, please upgrade to OpenSSL 0.9.9[-dev] for
9359 non-experimental use of the ECC ciphersuites to get TLS extension
9360 support, which is required for curve and point format negotiation
9361 to avoid potential handshake problems.
9365 * Disable rogue ciphersuites:
9367 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9368 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9369 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9371 The latter two were purportedly from
9372 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9375 Also deactivate the remaining ciphersuites from
9376 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9377 unofficial, and the ID has long expired.
9381 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9382 dual-core machines) and other potential thread-safety issues.
9386 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9387 versions), which is now available for royalty-free use
9388 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9389 Also, add Camellia TLS ciphersuites from RFC 4132.
9391 To minimize changes between patchlevels in the OpenSSL 0.9.8
9392 series, Camellia remains excluded from compilation unless OpenSSL
9393 is configured with 'enable-camellia'.
9397 * Disable the padding bug check when compression is in use. The padding
9398 bug check assumes the first packet is of even length, this is not
9399 necessarily true if compression is enabled and can result in false
9400 positives causing handshake failure. The actual bug test is ancient
9401 code so it is hoped that implementations will either have fixed it by
9402 now or any which still have the bug do not support compression.
9406 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9408 * When applying a cipher rule check to see if string match is an explicit
9409 cipher suite and only match that one cipher suite if it is.
9413 * Link in manifests for VC++ if needed.
9415 *Austin Ziegler <halostatue@gmail.com>*
9417 * Update support for ECC-based TLS ciphersuites according to
9418 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9419 TLS extensions, which are supported starting with the 0.9.9
9420 branch, not in the OpenSSL 0.9.8 branch).
9424 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9425 opaque EVP_CIPHER_CTX handling.
9429 * Fixes and enhancements to zlib compression code. We now only use
9430 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9431 to conform with the standards mentioned here:
9432 <http://www.zlib.net/DLL_FAQ.txt>
9433 Static zlib linking now works on Windows and the new --with-zlib-include
9434 --with-zlib-lib options to Configure can be used to supply the location
9435 of the headers and library. Gracefully handle case where zlib library
9440 * Several fixes and enhancements to the OID generation code. The old code
9441 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9442 handle numbers larger than ULONG_MAX, truncated printing and had a
9443 non standard OBJ_obj2txt() behaviour.
9447 * Add support for building of engines under engine/ as shared libraries
9448 under VC++ build system.
9452 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9453 Hopefully, we will not see any false combination of paths any more.
9457 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9459 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9460 (part of SSL_OP_ALL). This option used to disable the
9461 countermeasure against man-in-the-middle protocol-version
9462 rollback in the SSL 2.0 server implementation, which is a bad
9463 idea. [CVE-2005-2969][]
9465 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9466 for Information Security, National Institute of Advanced Industrial
9467 Science and Technology [AIST], Japan)*
9469 * Add two function to clear and return the verify parameter flags.
9473 * Keep cipherlists sorted in the source instead of sorting them at
9474 runtime, thus removing the need for a lock.
9478 * Avoid some small subgroup attacks in Diffie-Hellman.
9480 *Nick Mathewson and Ben Laurie*
9482 * Add functions for well-known primes.
9486 * Extended Windows CE support.
9488 *Satoshi Nakamura and Andy Polyakov*
9490 * Initialize SSL_METHOD structures at compile time instead of during
9491 runtime, thus removing the need for a lock.
9495 * Make PKCS7_decrypt() work even if no certificate is supplied by
9496 attempting to decrypt each encrypted key in turn. Add support to
9501 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9503 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9506 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9510 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9511 key into the same file any more.
9515 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9519 * Add -utf8 command line and config file option to 'ca'.
9521 *Stefan <stf@udoma.org*
9523 * Removed the macro des_crypt(), as it seems to conflict with some
9524 libraries. Use DES_crypt().
9528 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9529 involves renaming the source and generated shared-libs for
9530 both. The engines will accept the corrected or legacy ids
9531 ('ncipher' and '4758_cca' respectively) when binding. NB,
9532 this only applies when building 'shared'.
9534 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9536 * Add attribute functions to EVP_PKEY structure. Modify
9537 PKCS12_create() to recognize a CSP name attribute and
9538 use it. Make -CSP option work again in pkcs12 utility.
9542 * Add new functionality to the bn blinding code:
9543 - automatic re-creation of the BN_BLINDING parameters after
9544 a fixed number of uses (currently 32)
9545 - add new function for parameter creation
9546 - introduce flags to control the update behaviour of the
9547 BN_BLINDING parameters
9548 - hide BN_BLINDING structure
9549 Add a second BN_BLINDING slot to the RSA structure to improve
9550 performance when a single RSA object is shared among several
9555 * Add support for DTLS.
9557 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9559 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9560 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9564 * Remove buggy and incomplete DH cert support from
9565 ssl/ssl_rsa.c and ssl/s3_both.c
9569 * Use SHA-1 instead of MD5 as the default digest algorithm for
9570 the apps/openssl applications.
9574 * Compile clean with "-Wall -Wmissing-prototypes
9575 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9576 DEBUG_SAFESTACK must also be set.
9580 * Change ./Configure so that certain algorithms can be disabled by default.
9581 The new counterpiece to "no-xxx" is "enable-xxx".
9583 The patented RC5 and MDC2 algorithms will now be disabled unless
9584 "enable-rc5" and "enable-mdc2", respectively, are specified.
9586 (IDEA remains enabled despite being patented. This is because IDEA
9587 is frequently required for interoperability, and there is no license
9588 fee for non-commercial use. As before, "no-idea" can be used to
9589 avoid this algorithm.)
9593 * Add processing of proxy certificates (see RFC 3820). This work was
9594 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9595 EGEE (Enabling Grids for E-science in Europe).
9599 * RC4 performance overhaul on modern architectures/implementations, such
9600 as Intel P4, IA-64 and AMD64.
9604 * New utility extract-section.pl. This can be used specify an alternative
9605 section number in a pod file instead of having to treat each file as
9606 a separate case in Makefile. This can be done by adding two lines to the
9609 =for comment openssl_section:XXX
9611 The blank line is mandatory.
9615 * New arguments -certform, -keyform and -pass for s_client and s_server
9616 to allow alternative format key and certificate files and passphrase
9621 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9622 update associated structures and add various utility functions.
9624 Add new policy related verify parameters, include policy checking in
9625 standard verify code. Enhance 'smime' application with extra parameters
9626 to support policy checking and print out.
9630 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9631 Nehemiah processors. These extensions support AES encryption in hardware
9632 as well as RNG (though RNG support is currently disabled).
9634 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9636 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9640 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9642 *Andy Polyakov and a number of other people*
9644 * Improved PowerPC platform support. Most notably BIGNUM assembler
9645 implementation contributed by IBM.
9647 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9649 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9650 exponent rather than 'unsigned long'. There is a corresponding change to
9651 the new 'rsa_keygen' element of the RSA_METHOD structure.
9653 *Jelte Jansen, Geoff Thorpe*
9655 * Functionality for creating the initial serial number file is now
9656 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9658 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9659 number file to 1, which is bound to cause problems. To avoid
9660 the problems while respecting compatibility between different 0.9.7
9661 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9662 CA.pl for serial number initialization. With the new release 0.9.8,
9663 we can fix the problem directly in the 'ca' utility.)
9667 * Reduced header interdependencies by declaring more opaque objects in
9668 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9669 give fewer recursive includes, which could break lazy source code - so
9670 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9671 developers should define this symbol when building and using openssl to
9672 ensure they track the recommended behaviour, interfaces, [etc], but
9673 backwards-compatible behaviour prevails when this isn't defined.
9677 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9681 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9682 This will generate a random key of the appropriate length based on the
9683 cipher context. The EVP_CIPHER can provide its own random key generation
9684 routine to support keys of a specific form. This is used in the des and
9685 3des routines to generate a key of the correct parity. Update S/MIME
9686 code to use new functions and hence generate correct parity DES keys.
9687 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9688 valid (weak or incorrect parity).
9692 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9693 as looking them up. This is useful when the verified structure may contain
9694 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9695 present unless the new PKCS7_NO_CRL flag is asserted.
9699 * Extend ASN1 oid configuration module. It now additionally accepts the
9702 shortName = some long name, 1.2.3.4
9706 * Reimplemented the BN_CTX implementation. There is now no more static
9707 limitation on the number of variables it can handle nor the depth of the
9708 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9709 information can now expand as required, and rather than having a single
9710 static array of bignums, BN_CTX now uses a linked-list of such arrays
9711 allowing it to expand on demand whilst maintaining the usefulness of
9712 BN_CTX's "bundling".
9716 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9717 to allow all RSA operations to function using a single BN_CTX.
9721 * Preliminary support for certificate policy evaluation and checking. This
9722 is initially intended to pass the tests outlined in "Conformance Testing
9723 of Relying Party Client Certificate Path Processing Logic" v1.07.
9727 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9728 remained unused and not that useful. A variety of other little bignum
9729 tweaks and fixes have also been made continuing on from the audit (see
9734 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9735 associated ASN1, EVP and SSL functions and old ASN1 macros.
9739 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9740 and this should never fail. So the return value from the use of
9741 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9742 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9746 * BN_CTX_get() should return zero-valued bignums, providing the same
9747 initialised value as BN_new().
9749 *Geoff Thorpe, suggested by Ulf Möller*
9751 * Support for inhibitAnyPolicy certificate extension.
9755 * An audit of the BIGNUM code is underway, for which debugging code is
9756 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9757 is considered valid when processing BIGNUMs, and causes execution to
9758 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9759 further steps are taken to deliberately pollute unused data in BIGNUM
9760 structures to try and expose faulty code further on. For now, openssl will
9761 (in its default mode of operation) continue to tolerate the inconsistent
9762 forms that it has tolerated in the past, but authors and packagers should
9763 consider trying openssl and their own applications when compiled with
9764 these debugging symbols defined. It will help highlight potential bugs in
9765 their own code, and will improve the test coverage for OpenSSL itself. At
9766 some point, these tighter rules will become openssl's default to improve
9767 maintainability, though the assert()s and other overheads will remain only
9768 in debugging configurations. See bn.h for more details.
9770 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9772 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9773 that can only be obtained through BN_CTX_new() (which implicitly
9774 initialises it). The presence of this function only made it possible
9775 to overwrite an existing structure (and cause memory leaks).
9779 * Because of the callback-based approach for implementing LHASH as a
9780 template type, lh_insert() adds opaque objects to hash-tables and
9781 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9782 to clean up those corresponding objects before destroying the hash table
9783 (and losing the object pointers). So some over-zealous constifications in
9784 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9785 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9786 prototyped to have "const" restrictions on the object pointers they are
9787 given (and so aren't required to cast them away any more).
9791 * The tmdiff.h API was so ugly and minimal that our own timing utility
9792 (speed) prefers to use its own implementation. The two implementations
9793 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9794 its object type properly exposed (MS_TM) instead of casting to/from
9795 `char *`. This may still change yet if someone realises MS_TM and
9797 aren't necessarily the greatest nomenclatures - but this is what was used
9798 internally to the implementation so I've used that for now.
9802 * Ensure that deprecated functions do not get compiled when
9803 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9804 the self-tests were still using deprecated key-generation functions so
9805 these have been updated also.
9809 * Reorganise PKCS#7 code to separate the digest location functionality
9810 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9811 New function PKCS7_set_digest() to set the digest type for PKCS#7
9812 digestedData type. Add additional code to correctly generate the
9813 digestedData type and add support for this type in PKCS7 initialization
9818 * New function PKCS7_set0_type_other() this initializes a PKCS7
9819 structure of type "other".
9823 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9824 sure the loop does correctly stop and breaking ("division by zero")
9825 modulus operations are not performed. The (pre-generated) prime
9826 table crypto/bn/bn_prime.h was already correct, but it could not be
9827 re-generated on some platforms because of the "division by zero"
9828 situation in the script.
9830 *Ralf S. Engelschall*
9832 * Update support for ECC-based TLS ciphersuites according to
9833 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9834 SHA-1 now is only used for "small" curves (where the
9835 representation of a field element takes up to 24 bytes); for
9836 larger curves, the field element resulting from ECDH is directly
9837 used as premaster secret.
9839 *Douglas Stebila (Sun Microsystems Laboratories)*
9841 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9842 curve secp160r1 to the tests.
9844 *Douglas Stebila (Sun Microsystems Laboratories)*
9846 * Add the possibility to load symbols globally with DSO.
9848 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9850 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9851 control of the error stack.
9855 * Add support for STORE in ENGINE.
9859 * Add the STORE type. The intention is to provide a common interface
9860 to certificate and key stores, be they simple file-based stores, or
9861 HSM-type store, or LDAP stores, or...
9862 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9866 * Add a generic structure called OPENSSL_ITEM. This can be used to
9867 pass a list of arguments to any function as well as provide a way
9868 for a function to pass data back to the caller.
9872 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9873 works like BUF_strdup() but can be used to duplicate a portion of
9874 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9879 * Add the function sk_find_ex() which works like sk_find(), but will
9880 return an index to an element even if an exact match couldn't be
9881 found. The index is guaranteed to point at the element where the
9882 searched-for key would be inserted to preserve sorting order.
9886 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9887 takes an extra flags argument for optional functionality. Currently,
9888 the following flags are defined:
9890 OBJ_BSEARCH_VALUE_ON_NOMATCH
9891 This one gets OBJ_bsearch_ex() to return a pointer to the first
9892 element where the comparing function returns a negative or zero
9895 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9896 This one gets OBJ_bsearch_ex() to return a pointer to the first
9897 element where the comparing function returns zero. This is useful
9898 if there are more than one element where the comparing function
9903 * Make it possible to create self-signed certificates with 'openssl ca'
9904 in such a way that the self-signed certificate becomes part of the
9905 CA database and uses the same mechanisms for serial number generation
9906 as all other certificate signing. The new flag '-selfsign' enables
9907 this functionality. Adapt CA.sh and CA.pl.in.
9911 * Add functionality to check the public key of a certificate request
9912 against a given private. This is useful to check that a certificate
9913 request can be signed by that key (self-signing).
9917 * Make it possible to have multiple active certificates with the same
9918 subject in the CA index file. This is done only if the keyword
9919 'unique_subject' is set to 'no' in the main CA section (default
9920 if 'CA_default') of the configuration file. The value is saved
9921 with the database itself in a separate index attribute file,
9922 named like the index file with '.attr' appended to the name.
9926 * Generate multi-valued AVAs using '+' notation in config files for
9931 * Support for nameConstraints certificate extension.
9935 * Support for policyConstraints certificate extension.
9939 * Support for policyMappings certificate extension.
9943 * Make sure the default DSA_METHOD implementation only uses its
9944 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9945 and change its own handlers to be NULL so as to remove unnecessary
9946 indirection. This lets alternative implementations fallback to the
9947 default implementation more easily.
9951 * Support for directoryName in GeneralName related extensions
9956 * Make it possible to link applications using Makefile.shared.
9957 Make that possible even when linking against static libraries!
9961 * Support for single pass processing for S/MIME signing. This now
9962 means that S/MIME signing can be done from a pipe, in addition
9963 cleartext signing (multipart/signed type) is effectively streaming
9964 and the signed data does not need to be all held in memory.
9966 This is done with a new flag PKCS7_STREAM. When this flag is set
9967 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9968 is done after the data is output (and digests calculated) in
9969 SMIME_write_PKCS7().
9973 * Add full support for -rpath/-R, both in shared libraries and
9974 applications, at least on the platforms where it's known how
9979 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9980 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9981 will now compute a table of multiples of the generator that
9982 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9983 faster (notably in the case of a single point multiplication,
9984 scalar * generator).
9986 *Nils Larsch, Bodo Moeller*
9988 * IPv6 support for certificate extensions. The various extensions
9989 which use the IP:a.b.c.d can now take IPv6 addresses using the
9990 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9995 * Added an ENGINE that implements RSA by performing private key
9996 exponentiations with the GMP library. The conversions to and from
9997 GMP's mpz_t format aren't optimised nor are any montgomery forms
9998 cached, and on x86 it appears OpenSSL's own performance has caught up.
9999 However there are likely to be other architectures where GMP could
10000 provide a boost. This ENGINE is not built in by default, but it can be
10001 specified at Configure time and should be accompanied by the necessary
10002 linker additions, eg;
10003 ./config -DOPENSSL_USE_GMP -lgmp
10007 * "openssl engine" will not display ENGINE/DSO load failure errors when
10008 testing availability of engines with "-t" - the old behaviour is
10009 produced by increasing the feature's verbosity with "-tt".
10013 * ECDSA routines: under certain error conditions uninitialized BN objects
10014 could be freed. Solution: make sure initialization is performed early
10015 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10020 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10021 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10022 software implementations. For DSA and DH, parameter generation can
10023 also be overridden by providing the appropriate method callbacks.
10027 * Change the "progress" mechanism used in key-generation and
10028 primality testing to functions that take a new BN_GENCB pointer in
10029 place of callback/argument pairs. The new API functions have `_ex`
10030 postfixes and the older functions are reimplemented as wrappers for
10031 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10032 declarations of the old functions to help (graceful) attempts to
10033 migrate to the new functions. Also, the new key-generation API
10034 functions operate on a caller-supplied key-structure and return
10035 success/failure rather than returning a key or NULL - this is to
10036 help make "keygen" another member function of RSA_METHOD etc.
10038 Example for using the new callback interface:
10040 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10041 void *my_arg = ...;
10044 BN_GENCB_set(&my_cb, my_callback, my_arg);
10046 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10047 /* For the meaning of a, b in calls to my_callback(), see the
10048 * documentation of the function that calls the callback.
10049 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10050 * my_callback should return 1 if it wants BN_is_prime_ex()
10051 * to continue, or 0 to stop.
10056 * Change the ZLIB compression method to be stateful, and make it
10057 available to TLS with the number defined in
10058 draft-ietf-tls-compression-04.txt.
10062 * Add the ASN.1 structures and functions for CertificatePair, which
10063 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10065 CertificatePair ::= SEQUENCE {
10066 forward [0] Certificate OPTIONAL,
10067 reverse [1] Certificate OPTIONAL,
10068 -- at least one of the pair shall be present -- }
10070 Also implement the PEM functions to read and write certificate
10071 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10073 This needed to be defined, mostly for the sake of the LDAP
10074 attribute crossCertificatePair, but may prove useful elsewhere as
10079 * Make it possible to inhibit symlinking of shared libraries in
10080 Makefile.shared, for Cygwin's sake.
10084 * Extend the BIGNUM API by creating a function
10085 void BN_set_negative(BIGNUM *a, int neg);
10086 and a macro that behave like
10087 int BN_is_negative(const BIGNUM *a);
10089 to avoid the need to access 'a->neg' directly in applications.
10093 * Implement fast modular reduction for pseudo-Mersenne primes
10094 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10095 EC_GROUP_new_curve_GFp() will now automatically use this
10098 *Nils Larsch <nla@trustcenter.de>*
10100 * Add new lock type (CRYPTO_LOCK_BN).
10104 * Change the ENGINE framework to automatically load engines
10105 dynamically from specific directories unless they could be
10106 found to already be built in or loaded. Move all the
10107 current engines except for the cryptodev one to a new
10108 directory engines/.
10109 The engines in engines/ are built as shared libraries if
10110 the "shared" options was given to ./Configure or ./config.
10111 Otherwise, they are inserted in libcrypto.a.
10112 /usr/local/ssl/engines is the default directory for dynamic
10113 engines, but that can be overridden at configure time through
10114 the usual use of --prefix and/or --openssldir, and at run
10115 time with the environment variable OPENSSL_ENGINES.
10117 *Geoff Thorpe and Richard Levitte*
10119 * Add Makefile.shared, a helper makefile to build shared
10120 libraries. Adapt Makefile.org.
10124 * Add version info to Win32 DLLs.
10126 *Peter 'Luna' Runestig" <peter@runestig.com>*
10128 * Add new 'medium level' PKCS#12 API. Certificates and keys
10129 can be added using this API to created arbitrary PKCS#12
10130 files while avoiding the low level API.
10132 New options to PKCS12_create(), key or cert can be NULL and
10133 will then be omitted from the output file. The encryption
10134 algorithm NIDs can be set to -1 for no encryption, the mac
10135 iteration count can be set to 0 to omit the mac.
10137 Enhance pkcs12 utility by making the -nokeys and -nocerts
10138 options work when creating a PKCS#12 file. New option -nomac
10139 to omit the mac, NONE can be set for an encryption algorithm.
10140 New code is modified to use the enhanced PKCS12_create()
10141 instead of the low level API.
10145 * Extend ASN1 encoder to support indefinite length constructed
10146 encoding. This can output sequences tags and octet strings in
10147 this form. Modify pk7_asn1.c to support indefinite length
10148 encoding. This is experimental and needs additional code to
10149 be useful, such as an ASN1 bio and some enhanced streaming
10152 Extend template encode functionality so that tagging is passed
10153 down to the template encoder.
10157 * Let 'openssl req' fail if an argument to '-newkey' is not
10158 recognized instead of using RSA as a default.
10162 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10163 As these are not official, they are not included in "ALL";
10164 the "ECCdraft" ciphersuite group alias can be used to select them.
10166 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10168 * Add ECDH engine support.
10170 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10172 * Add ECDH in new directory crypto/ecdh/.
10174 *Douglas Stebila (Sun Microsystems Laboratories)*
10176 * Let BN_rand_range() abort with an error after 100 iterations
10177 without success (which indicates a broken PRNG).
10181 * Change BN_mod_sqrt() so that it verifies that the input value
10182 is really the square of the return value. (Previously,
10183 BN_mod_sqrt would show GIGO behaviour.)
10187 * Add named elliptic curves over binary fields from X9.62, SECG,
10188 and WAP/WTLS; add OIDs that were still missing.
10190 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10192 * Extend the EC library for elliptic curves over binary fields
10193 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10196 EC_GF2m_simple_method
10200 EC_GROUP_new_curve_GF2m
10201 EC_GROUP_set_curve_GF2m
10202 EC_GROUP_get_curve_GF2m
10203 EC_POINT_set_affine_coordinates_GF2m
10204 EC_POINT_get_affine_coordinates_GF2m
10205 EC_POINT_set_compressed_coordinates_GF2m
10207 Point compression for binary fields is disabled by default for
10208 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10211 As binary polynomials are represented as BIGNUMs, various members
10212 of the EC_GROUP and EC_POINT data structures can be shared
10213 between the implementations for prime fields and binary fields;
10214 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10215 are essentially identical to their `..._GFp` counterparts.
10216 (For simplicity, the `..._GFp` prefix has been dropped from
10217 various internal method names.)
10219 An internal 'field_div' method (similar to 'field_mul' and
10220 'field_sqr') has been added; this is used only for binary fields.
10222 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10224 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10225 through methods ('mul', 'precompute_mult').
10227 The generic implementations (now internally called 'ec_wNAF_mul'
10228 and 'ec_wNAF_precomputed_mult') remain the default if these
10229 methods are undefined.
10231 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10233 * New function EC_GROUP_get_degree, which is defined through
10234 EC_METHOD. For curves over prime fields, this returns the bit
10235 length of the modulus.
10237 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10239 * New functions EC_GROUP_dup, EC_POINT_dup.
10240 (These simply call ..._new and ..._copy).
10242 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10244 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10245 Polynomials are represented as BIGNUMs (where the sign bit is not
10246 used) in the following functions [macros]:
10249 BN_GF2m_sub [= BN_GF2m_add]
10250 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10251 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10252 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10254 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10255 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10256 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10257 BN_GF2m_cmp [= BN_ucmp]
10259 (Note that only the 'mod' functions are actually for fields GF(2^m).
10260 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10262 For some functions, an the irreducible polynomial defining a
10263 field can be given as an 'unsigned int[]' with strictly
10264 decreasing elements giving the indices of those bits that are set;
10265 i.e., p[] represents the polynomial
10266 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10268 p[0] > p[1] > ... > p[k] = 0.
10269 This applies to the following functions:
10272 BN_GF2m_mod_mul_arr
10273 BN_GF2m_mod_sqr_arr
10274 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10275 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10276 BN_GF2m_mod_exp_arr
10277 BN_GF2m_mod_sqrt_arr
10278 BN_GF2m_mod_solve_quad_arr
10282 Conversion can be performed by the following functions:
10287 bntest.c has additional tests for binary polynomial arithmetic.
10289 Two implementations for BN_GF2m_mod_div() are available.
10290 The default algorithm simply uses BN_GF2m_mod_inv() and
10291 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10292 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10293 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10295 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10297 * Add new error code 'ERR_R_DISABLED' that can be used when some
10298 functionality is disabled at compile-time.
10300 *Douglas Stebila <douglas.stebila@sun.com>*
10302 * Change default behaviour of 'openssl asn1parse' so that more
10303 information is visible when viewing, e.g., a certificate:
10305 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10306 mode the content of non-printable OCTET STRINGs is output in a
10307 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10308 avoid the appearance of a printable string.
10310 *Nils Larsch <nla@trustcenter.de>*
10312 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10314 EC_GROUP_set_asn1_flag()
10315 EC_GROUP_get_asn1_flag()
10316 EC_GROUP_set_point_conversion_form()
10317 EC_GROUP_get_point_conversion_form()
10318 These control ASN1 encoding details:
10319 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10320 has been set to OPENSSL_EC_NAMED_CURVE.
10321 - Points are encoded in uncompressed form by default; options for
10322 asn1_for are as for point2oct, namely
10323 POINT_CONVERSION_COMPRESSED
10324 POINT_CONVERSION_UNCOMPRESSED
10325 POINT_CONVERSION_HYBRID
10327 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10329 EC_GROUP_set_seed()
10330 EC_GROUP_get0_seed()
10331 EC_GROUP_get_seed_len()
10332 This is used only for ASN1 purposes (so far).
10334 *Nils Larsch <nla@trustcenter.de>*
10336 * Add 'field_type' member to EC_METHOD, which holds the NID
10337 of the appropriate field type OID. The new function
10338 EC_METHOD_get_field_type() returns this value.
10340 *Nils Larsch <nla@trustcenter.de>*
10343 EC_POINT_point2bn()
10344 EC_POINT_bn2point()
10345 EC_POINT_point2hex()
10346 EC_POINT_hex2point()
10347 providing useful interfaces to EC_POINT_point2oct() and
10348 EC_POINT_oct2point().
10350 *Nils Larsch <nla@trustcenter.de>*
10352 * Change internals of the EC library so that the functions
10353 EC_GROUP_set_generator()
10354 EC_GROUP_get_generator()
10355 EC_GROUP_get_order()
10356 EC_GROUP_get_cofactor()
10357 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10358 to methods, which would lead to unnecessary code duplication when
10359 adding different types of curves.
10361 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10363 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10364 arithmetic, and such that modified wNAFs are generated
10365 (which avoid length expansion in many cases).
10369 * Add a function EC_GROUP_check_discriminant() (defined via
10370 EC_METHOD) that verifies that the curve discriminant is non-zero.
10372 Add a function EC_GROUP_check() that makes some sanity tests
10373 on a EC_GROUP, its generator and order. This includes
10374 EC_GROUP_check_discriminant().
10376 *Nils Larsch <nla@trustcenter.de>*
10378 * Add ECDSA in new directory crypto/ecdsa/.
10380 Add applications 'openssl ecparam' and 'openssl ecdsa'
10381 (these are based on 'openssl dsaparam' and 'openssl dsa').
10383 ECDSA support is also included in various other files across the
10384 library. Most notably,
10385 - 'openssl req' now has a '-newkey ecdsa:file' option;
10386 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10387 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10388 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10389 them suitable for ECDSA where domain parameters must be
10390 extracted before the specific public key;
10391 - ECDSA engine support has been added.
10393 *Nils Larsch <nla@trustcenter.de>*
10395 * Include some named elliptic curves, and add OIDs from X9.62,
10396 SECG, and WAP/WTLS. Each curve can be obtained from the new
10398 EC_GROUP_new_by_curve_name(),
10399 and the list of available named curves can be obtained with
10400 EC_get_builtin_curves().
10401 Also add a 'curve_name' member to EC_GROUP objects, which can be
10403 EC_GROUP_set_curve_name()
10404 EC_GROUP_get_curve_name()
10406 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10408 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10409 was actually never needed) and in BN_mul(). The removal in BN_mul()
10410 required a small change in bn_mul_part_recursive() and the addition
10411 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10412 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10413 bn_sub_words() and bn_add_words() except they take arrays with
10418 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10420 * Cleanse PEM buffers before freeing them since they may contain
10423 *Benjamin Bennett <ben@psc.edu>*
10425 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10426 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10427 authentication-only ciphersuites.
10431 * Since AES128 and AES256 share a single mask bit in the logic of
10432 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10433 kludge to work properly if AES128 is available and AES256 isn't.
10437 * Expand security boundary to match 1.1.1 module.
10441 * Remove redundant features: hash file source, editing of test vectors
10442 modify fipsld to use external fips_premain.c signature.
10446 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10447 run algorithm test programs.
10451 * Make algorithm test programs more tolerant of whitespace.
10455 * Have SSL/TLS server implementation tolerate "mismatched" record
10456 protocol version while receiving ClientHello even if the
10457 ClientHello is fragmented. (The server can't insist on the
10458 particular protocol version it has chosen before the ServerHello
10459 message has informed the client about his choice.)
10463 * Load error codes if they are not already present instead of using a
10464 static variable. This allows them to be cleanly unloaded and reloaded.
10468 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10470 * Introduce limits to prevent malicious keys being able to
10471 cause a denial of service. [CVE-2006-2940][]
10473 *Steve Henson, Bodo Moeller*
10475 * Fix ASN.1 parsing of certain invalid structures that can result
10476 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10478 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10479 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10481 * Fix SSL client code which could crash if connecting to a
10482 malicious SSLv2 server. [CVE-2006-4343][]
10484 *Tavis Ormandy and Will Drewry, Google Security Team*
10486 * Change ciphersuite string processing so that an explicit
10487 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10488 will no longer include "AES128-SHA"), and any other similar
10489 ciphersuite (same bitmap) from *other* protocol versions (so that
10490 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10491 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10492 changes from 0.9.8b and 0.9.8d.
10496 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10498 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10499 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10501 * Change the Unix randomness entropy gathering to use poll() when
10502 possible instead of select(), since the latter has some
10503 undesirable limitations.
10505 *Darryl Miles via Richard Levitte and Bodo Moeller*
10507 * Disable rogue ciphersuites:
10509 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10510 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10511 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10513 The latter two were purportedly from
10514 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10517 Also deactivate the remaining ciphersuites from
10518 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10519 unofficial, and the ID has long expired.
10523 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10524 dual-core machines) and other potential thread-safety issues.
10528 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10530 * Adapt fipsld and the build system to link against the validated FIPS
10531 module in FIPS mode.
10535 * Fixes for VC++ 2005 build under Windows.
10539 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10540 from a Windows bash shell such as MSYS. It is autodetected from the
10541 "config" script when run from a VC++ environment. Modify standard VC++
10542 build to use fipscanister.o from the GNU make build.
10546 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10548 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10549 The value now differs depending on if you build for FIPS or not.
10550 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10551 safely run with a non-FIPSed libcrypto, as it may crash because of
10552 the difference induced by this change.
10556 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10558 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10559 (part of SSL_OP_ALL). This option used to disable the
10560 countermeasure against man-in-the-middle protocol-version
10561 rollback in the SSL 2.0 server implementation, which is a bad
10562 idea. [CVE-2005-2969][]
10564 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10565 for Information Security, National Institute of Advanced Industrial
10566 Science and Technology [AIST, Japan)]*
10568 * Minimal support for X9.31 signatures and PSS padding modes. This is
10569 mainly for FIPS compliance and not fully integrated at this stage.
10573 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10574 the exponentiation using a fixed-length exponent. (Otherwise,
10575 the information leaked through timing could expose the secret key
10576 after many signatures; cf. Bleichenbacher's attack on DSA with
10581 * Make a new fixed-window mod_exp implementation the default for
10582 RSA, DSA, and DH private-key operations so that the sequence of
10583 squares and multiplies and the memory access pattern are
10584 independent of the particular secret key. This will mitigate
10585 cache-timing and potential related attacks.
10587 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10588 and this is automatically used by BN_mod_exp_mont() if the new flag
10589 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10590 will use this BN flag for private exponents unless the flag
10591 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10592 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10594 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10596 * Change the client implementation for SSLv23_method() and
10597 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10598 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10599 (Previously, the SSL 2.0 backwards compatible Client Hello
10600 message format would be used even with SSL_OP_NO_SSLv2.)
10604 * Add support for smime-type MIME parameter in S/MIME messages which some
10609 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10610 a threadsafe manner. Modify rsa code to use new function and add calls
10611 to dsa and dh code (which had race conditions before).
10615 * Include the fixed error library code in the C error file definitions
10616 instead of fixing them up at runtime. This keeps the error code
10617 structures constant.
10621 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10623 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10626 * Fixes for newer kerberos headers. NB: the casts are needed because
10627 the 'length' field is signed on one version and unsigned on another
10628 with no (?) obvious way to tell the difference, without these VC++
10629 complains. Also the "definition" of FAR (blank) is no longer included
10630 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10631 some needed definitions.
10635 * Undo Cygwin change.
10639 * Added support for proxy certificates according to RFC 3820.
10640 Because they may be a security thread to unaware applications,
10641 they must be explicitly allowed in run-time. See
10642 docs/HOWTO/proxy_certificates.txt for further information.
10646 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10648 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10649 server and client random values. Previously
10650 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10651 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10653 This change has negligible security impact because:
10655 1. Server and client random values still have 24 bytes of pseudo random
10658 2. Server and client random values are sent in the clear in the initial
10661 3. The master secret is derived using the premaster secret (48 bytes in
10662 size for static RSA ciphersuites) as well as client server and random
10665 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10668 *Stephen Henson, reported by UK NISCC*
10670 * Use Windows randomness collection on Cygwin.
10674 * Fix hang in EGD/PRNGD query when communication socket is closed
10675 prematurely by EGD/PRNGD.
10677 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10679 * Prompt for pass phrases when appropriate for PKCS12 input format.
10683 * Back-port of selected performance improvements from development
10684 branch, as well as improved support for PowerPC platforms.
10688 * Add lots of checks for memory allocation failure, error codes to indicate
10689 failure and freeing up memory if a failure occurs.
10691 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10693 * Add new -passin argument to dgst.
10697 * Perform some character comparisons of different types in X509_NAME_cmp:
10698 this is needed for some certificates that re-encode DNs into UTF8Strings
10699 (in violation of RFC3280) and can't or won't issue name rollover
10704 * Make an explicit check during certificate validation to see that
10705 the CA setting in each certificate on the chain is correct. As a
10706 side effect always do the following basic checks on extensions,
10707 not just when there's an associated purpose to the check:
10709 - if there is an unhandled critical extension (unless the user
10710 has chosen to ignore this fault)
10711 - if the path length has been exceeded (if one is set at all)
10712 - that certain extensions fit the associated purpose (if one has
10717 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10719 * Avoid a race condition when CRLs are checked in a multi threaded
10720 environment. This would happen due to the reordering of the revoked
10721 entries during signature checking and serial number lookup. Now the
10722 encoding is cached and the serial number sort performed under a lock.
10723 Add new STACK function sk_is_sorted().
10727 * Add Delta CRL to the extension code.
10731 * Various fixes to s3_pkt.c so alerts are sent properly.
10733 *David Holmes <d.holmes@f5.com>*
10735 * Reduce the chances of duplicate issuer name and serial numbers (in
10736 violation of RFC3280) using the OpenSSL certificate creation utilities.
10737 This is done by creating a random 64 bit value for the initial serial
10738 number when a serial number file is created or when a self signed
10739 certificate is created using 'openssl req -x509'. The initial serial
10740 number file is created using 'openssl x509 -next_serial' in CA.pl
10741 rather than being initialized to 1.
10745 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10747 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10748 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10750 *Joe Orton, Steve Henson*
10752 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10755 *Joe Orton, Steve Henson*
10757 * Make it possible to have multiple active certificates with the same
10758 subject in the CA index file. This is done only if the keyword
10759 'unique_subject' is set to 'no' in the main CA section (default
10760 if 'CA_default') of the configuration file. The value is saved
10761 with the database itself in a separate index attribute file,
10762 named like the index file with '.attr' appended to the name.
10766 * X509 verify fixes. Disable broken certificate workarounds when
10767 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10768 keyUsage extension present. Don't accept CRLs with unhandled critical
10769 extensions: since verify currently doesn't process CRL extensions this
10770 rejects a CRL with *any* critical extensions. Add new verify error codes
10775 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10776 A clarification of RFC2560 will require the use of OCTET STRINGs and
10777 some implementations cannot handle the current raw format. Since OpenSSL
10778 copies and compares OCSP nonces as opaque blobs without any attempt at
10779 parsing them this should not create any compatibility issues.
10783 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10784 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10785 this HMAC (and other) operations are several times slower than OpenSSL
10790 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10792 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10794 * Use the correct content when signing type "other".
10798 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10800 * Fix various bugs revealed by running the NISCC test suite:
10802 Stop out of bounds reads in the ASN1 code when presented with
10803 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10805 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10807 If verify callback ignores invalid public key errors don't try to check
10808 certificate signature with the NULL public key.
10812 * New -ignore_err option in ocsp application to stop the server
10813 exiting on the first error in a request.
10817 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10818 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10823 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10824 extra data after the compression methods not only for TLS 1.0
10825 but also for SSL 3.0 (as required by the specification).
10827 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10829 * Change X509_certificate_type() to mark the key as exported/exportable
10830 when it's 512 *bits* long, not 512 bytes.
10834 * Change AES_cbc_encrypt() so it outputs exact multiple of
10835 blocks during encryption.
10839 * Various fixes to base64 BIO and non blocking I/O. On write
10840 flushes were not handled properly if the BIO retried. On read
10841 data was not being buffered properly and had various logic bugs.
10842 This also affects blocking I/O when the data being decoded is a
10847 * Various S/MIME bugfixes and compatibility changes:
10848 output correct application/pkcs7 MIME type if
10849 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10850 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10851 of files as .eml work). Correctly handle very long lines in MIME
10856 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10858 * Countermeasure against the Klima-Pokorny-Rosa extension of
10859 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10860 a protocol version number mismatch like a decryption error
10861 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10865 * Turn on RSA blinding by default in the default implementation
10866 to avoid a timing attack. Applications that don't want it can call
10867 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10868 They would be ill-advised to do so in most cases.
10870 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10872 * Change RSA blinding code so that it works when the PRNG is not
10873 seeded (in this case, the secret RSA exponent is abused as
10874 an unpredictable seed -- if it is not unpredictable, there
10875 is no point in blinding anyway). Make RSA blinding thread-safe
10876 by remembering the creator's thread ID in rsa->blinding and
10877 having all other threads use local one-time blinding factors
10878 (this requires more computation than sharing rsa->blinding, but
10879 avoids excessive locking; and if an RSA object is not shared
10880 between threads, blinding will still be very fast).
10884 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10885 ENGINE as defaults for all supported algorithms irrespective of
10886 the 'flags' parameter. 'flags' is now honoured, so applications
10887 should make sure they are passing it correctly.
10891 * Target "mingw" now allows native Windows code to be generated in
10892 the Cygwin environment as well as with the MinGW compiler.
10896 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10898 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10899 via timing by performing a MAC computation even if incorrect
10900 block cipher padding has been found. This is a countermeasure
10901 against active attacks where the attacker has to distinguish
10902 between bad padding and a MAC verification error. [CVE-2003-0078][]
10904 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10905 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10906 Martin Vuagnoux (EPFL, Ilion)*
10908 * Make the no-err option work as intended. The intention with no-err
10909 is not to have the whole error stack handling routines removed from
10910 libcrypto, it's only intended to remove all the function name and
10911 reason texts, thereby removing some of the footprint that may not
10912 be interesting if those errors aren't displayed anyway.
10914 NOTE: it's still possible for any application or module to have its
10915 own set of error texts inserted. The routines are there, just not
10916 used by default when no-err is given.
10920 * Add support for FreeBSD on IA64.
10922 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10924 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10925 Kerberos function mit_des_cbc_cksum(). Before this change,
10926 the value returned by DES_cbc_cksum() was like the one from
10927 mit_des_cbc_cksum(), except the bytes were swapped.
10929 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10931 * Allow an application to disable the automatic SSL chain building.
10932 Before this a rather primitive chain build was always performed in
10933 ssl3_output_cert_chain(): an application had no way to send the
10934 correct chain if the automatic operation produced an incorrect result.
10936 Now the chain builder is disabled if either:
10938 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10940 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10942 The reasoning behind this is that an application would not want the
10943 auto chain building to take place if extra chain certificates are
10944 present and it might also want a means of sending no additional
10945 certificates (for example the chain has two certificates and the
10950 * Add the possibility to build without the ENGINE framework.
10952 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10954 * Under Win32 gmtime() can return NULL: check return value in
10955 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10959 * DSA routines: under certain error conditions uninitialized BN objects
10960 could be freed. Solution: make sure initialization is performed early
10961 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10962 Nils Larsch <nla@trustcenter.de> via PR#459)
10966 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10967 checked on reconnect on the client side, therefore session resumption
10968 could still fail with a "ssl session id is different" error. This
10969 behaviour is masked when SSL_OP_ALL is used due to
10970 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10971 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10972 followup to PR #377.
10976 * IA-32 assembler support enhancements: unified ELF targets, support
10977 for SCO/Caldera platforms, fix for Cygwin shared build.
10981 * Add support for FreeBSD on sparc64. As a consequence, support for
10982 FreeBSD on non-x86 processors is separate from x86 processors on
10983 the config script, much like the NetBSD support.
10985 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10987 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
10989 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10992 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10993 code (06) was taken as the first octet of the session ID and the last
10994 octet was ignored consequently. As a result SSLv2 client side session
10995 caching could not have worked due to the session ID mismatch between
10997 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11002 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11003 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11008 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11009 seems that in spite of existing for more than a year, many application
11010 author have done nothing to provide the necessary callbacks, which
11011 means that this particular engine will not work properly anywhere.
11012 This is a very unfortunate situation which forces us, in the name
11013 of usability, to give the hw_ncipher.c a static lock, which is part
11015 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11016 appear in 0.9.8 or later. We EXPECT application authors to have
11017 dealt properly with this when 0.9.8 is released (unless we actually
11018 make such changes in the libcrypto locking code that changes will
11019 have to be made anyway).
11023 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11024 octets have been read, EOF or an error occurs. Without this change
11025 some truncated ASN1 structures will not produce an error.
11029 * Disable Heimdal support, since it hasn't been fully implemented.
11030 Still give the possibility to force the use of Heimdal, but with
11031 warnings and a request that patches get sent to openssl-dev.
11035 * Add the VC-CE target, introduce the WINCE sysname, and add
11036 INSTALL.WCE and appropriate conditionals to make it build.
11038 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11040 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11041 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11042 edit numbers of the version.
11044 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11046 * Introduce safe string copy and catenation functions
11047 (BUF_strlcpy() and BUF_strlcat()).
11049 *Ben Laurie (CHATS) and Richard Levitte*
11051 * Avoid using fixed-size buffers for one-line DNs.
11053 *Ben Laurie (CHATS)*
11055 * Add BUF_MEM_grow_clean() to avoid information leakage when
11056 resizing buffers containing secrets, and use where appropriate.
11058 *Ben Laurie (CHATS)*
11060 * Avoid using fixed size buffers for configuration file location.
11062 *Ben Laurie (CHATS)*
11064 * Avoid filename truncation for various CA files.
11066 *Ben Laurie (CHATS)*
11068 * Use sizeof in preference to magic numbers.
11070 *Ben Laurie (CHATS)*
11072 * Avoid filename truncation in cert requests.
11074 *Ben Laurie (CHATS)*
11076 * Add assertions to check for (supposedly impossible) buffer
11079 *Ben Laurie (CHATS)*
11081 * Don't cache truncated DNS entries in the local cache (this could
11082 potentially lead to a spoofing attack).
11084 *Ben Laurie (CHATS)*
11086 * Fix various buffers to be large enough for hex/decimal
11087 representations in a platform independent manner.
11089 *Ben Laurie (CHATS)*
11091 * Add CRYPTO_realloc_clean() to avoid information leakage when
11092 resizing buffers containing secrets, and use where appropriate.
11094 *Ben Laurie (CHATS)*
11096 * Add BIO_indent() to avoid much slightly worrying code to do
11099 *Ben Laurie (CHATS)*
11101 * Convert sprintf()/BIO_puts() to BIO_printf().
11103 *Ben Laurie (CHATS)*
11105 * buffer_gets() could terminate with the buffer only half
11108 *Ben Laurie (CHATS)*
11110 * Add assertions to prevent user-supplied crypto functions from
11111 overflowing internal buffers by having large block sizes, etc.
11113 *Ben Laurie (CHATS)*
11115 * New OPENSSL_assert() macro (similar to assert(), but enabled
11118 *Ben Laurie (CHATS)*
11120 * Eliminate unused copy of key in RC4.
11122 *Ben Laurie (CHATS)*
11124 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11126 *Ben Laurie (CHATS)*
11128 * Fix off-by-one error in EGD path.
11130 *Ben Laurie (CHATS)*
11132 * If RANDFILE path is too long, ignore instead of truncating.
11134 *Ben Laurie (CHATS)*
11136 * Eliminate unused and incorrectly sized X.509 structure
11139 *Ben Laurie (CHATS)*
11141 * Eliminate unused and dangerous function knumber().
11143 *Ben Laurie (CHATS)*
11145 * Eliminate unused and dangerous structure, KSSL_ERR.
11147 *Ben Laurie (CHATS)*
11149 * Protect against overlong session ID context length in an encoded
11150 session object. Since these are local, this does not appear to be
11153 *Ben Laurie (CHATS)*
11155 * Change from security patch (see 0.9.6e below) that did not affect
11156 the 0.9.6 release series:
11158 Remote buffer overflow in SSL3 protocol - an attacker could
11159 supply an oversized master key in Kerberos-enabled versions.
11162 *Ben Laurie (CHATS)*
11164 * Change the SSL kerb5 codes to match RFC 2712.
11168 * Make -nameopt work fully for req and add -reqopt switch.
11170 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11172 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11174 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11176 * Make sure tests can be performed even if the corresponding algorithms
11177 have been removed entirely. This was also the last step to make
11178 OpenSSL compilable with DJGPP under all reasonable conditions.
11180 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11182 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11183 to allow version independent disabling of normally unselected ciphers,
11184 which may be activated as a side-effect of selecting a single cipher.
11186 (E.g., cipher list string "RSA" enables ciphersuites that are left
11187 out of "ALL" because they do not provide symmetric encryption.
11188 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11190 *Lutz Jaenicke, Bodo Moeller*
11192 * Add appropriate support for separate platform-dependent build
11193 directories. The recommended way to make a platform-dependent
11194 build directory is the following (tested on Linux), maybe with
11197 # Place yourself outside of the OpenSSL source tree. In
11198 # this example, the environment variable OPENSSL_SOURCE
11199 # is assumed to contain the absolute OpenSSL source directory.
11200 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11201 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11202 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11203 mkdir -p `dirname $F`
11204 ln -s $OPENSSL_SOURCE/$F $F
11207 To be absolutely sure not to disturb the source tree, a "make clean"
11208 is a good thing. If it isn't successful, don't worry about it,
11209 it probably means the source directory is very clean.
11213 * Make sure any ENGINE control commands make local copies of string
11214 pointers passed to them whenever necessary. Otherwise it is possible
11215 the caller may have overwritten (or deallocated) the original string
11216 data when a later ENGINE operation tries to use the stored values.
11218 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11220 * Improve diagnostics in file reading and command-line digests.
11222 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11224 * Add AES modes CFB and OFB to the object database. Correct an
11225 error in AES-CFB decryption.
11229 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11230 allows existing EVP_CIPHER_CTX structures to be reused after
11231 calling `EVP_*Final()`. This behaviour is used by encryption
11232 BIOs and some applications. This has the side effect that
11233 applications must explicitly clean up cipher contexts with
11234 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11238 * Check the values of dna and dnb in bn_mul_recursive before calling
11239 bn_mul_comba (a non zero value means the a or b arrays do not contain
11240 n2 elements) and fallback to bn_mul_normal if either is not zero.
11244 * Fix escaping of non-ASCII characters when using the -subj option
11245 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11249 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11250 form for "surname", serialNumber has no short form.
11251 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11252 therefore remove "mail" short name for "internet 7".
11253 The OID for unique identifiers in X509 certificates is
11254 x500UniqueIdentifier, not uniqueIdentifier.
11255 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11259 * Add an "init" command to the ENGINE config module and auto initialize
11260 ENGINEs. Without any "init" command the ENGINE will be initialized
11261 after all ctrl commands have been executed on it. If init=1 the
11262 ENGINE is initialized at that point (ctrls before that point are run
11263 on the uninitialized ENGINE and after on the initialized one). If
11264 init=0 then the ENGINE will not be initialized at all.
11268 * Fix the 'app_verify_callback' interface so that the user-defined
11269 argument is actually passed to the callback: In the
11270 SSL_CTX_set_cert_verify_callback() prototype, the callback
11271 declaration has been changed from
11274 int (*cb)(X509_STORE_CTX *,void *);
11275 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11276 i=s->ctx->app_verify_callback(&ctx)
11277 has been changed into
11278 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11280 To update applications using SSL_CTX_set_cert_verify_callback(),
11281 a dummy argument can be added to their callback functions.
11283 *D. K. Smetters <smetters@parc.xerox.com>*
11285 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11287 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11289 * Add and OPENSSL_LOAD_CONF define which will cause
11290 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11291 This allows older applications to transparently support certain
11292 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11293 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11294 load the config file and OPENSSL_add_all_algorithms_conf() which will
11295 always load it have also been added.
11299 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11300 Adjust NIDs and EVP layer.
11302 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11304 * Config modules support in openssl utility.
11306 Most commands now load modules from the config file,
11307 though in a few (such as version) this isn't done
11308 because it couldn't be used for anything.
11310 In the case of ca and req the config file used is
11311 the same as the utility itself: that is the -config
11312 command line option can be used to specify an
11317 * Move default behaviour from OPENSSL_config(). If appname is NULL
11318 use "openssl_conf" if filename is NULL use default openssl config file.
11322 * Add an argument to OPENSSL_config() to allow the use of an alternative
11323 config section name. Add a new flag to tolerate a missing config file
11324 and move code to CONF_modules_load_file().
11328 * Support for crypto accelerator cards from Accelerated Encryption
11329 Processing, www.aep.ie. (Use engine 'aep')
11330 The support was copied from 0.9.6c [engine] and adapted/corrected
11331 to work with the new engine framework.
11333 *AEP Inc. and Richard Levitte*
11335 * Support for SureWare crypto accelerator cards from Baltimore
11336 Technologies. (Use engine 'sureware')
11337 The support was copied from 0.9.6c [engine] and adapted
11338 to work with the new engine framework.
11342 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11343 make the newer ENGINE framework commands for the CHIL engine work.
11345 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11347 * Make it possible to produce shared libraries on ReliantUNIX.
11349 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11351 * Add the configuration target debug-linux-ppro.
11352 Make 'openssl rsa' use the general key loading routines
11353 implemented in apps.c, and make those routines able to
11354 handle the key format FORMAT_NETSCAPE and the variant
11357 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11359 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11361 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11363 * Add -keyform to rsautl, and document -engine.
11365 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11367 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11368 BIO_R_NO_SUCH_FILE error code rather than the generic
11369 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11373 * Add new functions
11374 ERR_peek_last_error
11375 ERR_peek_last_error_line
11376 ERR_peek_last_error_line_data.
11377 These are similar to
11379 ERR_peek_error_line
11380 ERR_peek_error_line_data,
11381 but report on the latest error recorded rather than the first one
11382 still in the error queue.
11384 *Ben Laurie, Bodo Moeller*
11386 * default_algorithms option in ENGINE config module. This allows things
11388 default_algorithms = ALL
11389 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11393 * Preliminary ENGINE config module.
11397 * New experimental application configuration code.
11401 * Change the AES code to follow the same name structure as all other
11402 symmetric ciphers, and behave the same way. Move everything to
11403 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11405 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11407 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11409 *Ben Laurie and Theo de Raadt*
11411 * Add option to output public keys in req command.
11413 *Massimiliano Pala madwolf@openca.org*
11415 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11416 (up to about 10% better than before for P-192 and P-224).
11420 * New functions/macros
11422 SSL_CTX_set_msg_callback(ctx, cb)
11423 SSL_CTX_set_msg_callback_arg(ctx, arg)
11424 SSL_set_msg_callback(ssl, cb)
11425 SSL_set_msg_callback_arg(ssl, arg)
11427 to request calling a callback function
11429 void cb(int write_p, int version, int content_type,
11430 const void *buf, size_t len, SSL *ssl, void *arg)
11432 whenever a protocol message has been completely received
11433 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11434 protocol version according to which the SSL library interprets
11435 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11436 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11437 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11438 specification (change_cipher_spec(20), alert(21), handshake(22)).
11439 'buf' and 'len' point to the actual message, 'ssl' to the
11440 SSL object, and 'arg' is the application-defined value set by
11441 SSL[_CTX]_set_msg_callback_arg().
11443 'openssl s_client' and 'openssl s_server' have new '-msg' options
11444 to enable a callback that displays all protocol messages.
11448 * Change the shared library support so shared libraries are built as
11449 soon as the corresponding static library is finished, and thereby get
11450 openssl and the test programs linked against the shared library.
11451 This still only happens when the keyword "shard" has been given to
11452 the configuration scripts.
11454 NOTE: shared library support is still an experimental thing, and
11455 backward binary compatibility is still not guaranteed.
11457 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11459 * Add support for Subject Information Access extension.
11461 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11463 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11464 additional bytes when new memory had to be allocated, not just
11465 when reusing an existing buffer.
11469 * New command line and configuration option 'utf8' for the req command.
11470 This allows field values to be specified as UTF8 strings.
11474 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11475 runs for the former and machine-readable output for the latter.
11479 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11480 of the e-mail address in the DN (i.e., it will go into a certificate
11481 extension only). The new configuration file option 'email_in_dn = no'
11482 has the same effect.
11484 *Massimiliano Pala madwolf@openca.org*
11486 * Change all functions with names starting with `des_` to be starting
11487 with `DES_` instead. Add wrappers that are compatible with libdes,
11488 but are named `_ossl_old_des_*`. Finally, add macros that map the
11489 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11490 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11491 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11494 Since we provide two compatibility mappings, the user needs to
11495 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11496 compatibility is desired. The default (i.e., when that macro
11497 isn't defined) is OpenSSL 0.9.6c compatibility.
11499 There are also macros that enable and disable the support of old
11500 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11501 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11502 are defined, the default will apply: to support the old des routines.
11504 In either case, one must include openssl/des.h to get the correct
11505 definitions. Do not try to just include openssl/des_old.h, that
11508 NOTE: This is a major break of an old API into a new one. Software
11509 authors are encouraged to switch to the `DES_` style functions. Some
11510 time in the future, des_old.h and the libdes compatibility functions
11511 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11512 default), and then completely removed.
11516 * Test for certificates which contain unsupported critical extensions.
11517 If such a certificate is found during a verify operation it is
11518 rejected by default: this behaviour can be overridden by either
11519 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11520 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11521 X509_supported_extension() has also been added which returns 1 if a
11522 particular extension is supported.
11526 * Modify the behaviour of EVP cipher functions in similar way to digests
11527 to retain compatibility with existing code.
11531 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11532 compatibility with existing code. In particular the 'ctx' parameter does
11533 not have to be to be initialized before the call to EVP_DigestInit() and
11534 it is tidied up after a call to EVP_DigestFinal(). New function
11535 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11536 EVP_MD_CTX_copy() changed to not require the destination to be
11537 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11538 requires the destination to be valid.
11540 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11541 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11545 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11546 so that complete 'Handshake' protocol structures are kept in memory
11547 instead of overwriting 'msg_type' and 'length' with 'body' data.
11551 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11553 *Massimo Santin via Richard Levitte*
11555 * Major restructuring to the underlying ENGINE code. This includes
11556 reduction of linker bloat, separation of pure "ENGINE" manipulation
11557 (initialisation, etc) from functionality dealing with implementations
11558 of specific crypto interfaces. This change also introduces integrated
11559 support for symmetric ciphers and digest implementations - so ENGINEs
11560 can now accelerate these by providing EVP_CIPHER and EVP_MD
11561 implementations of their own. This is detailed in crypto/engine/README
11562 as it couldn't be adequately described here. However, there are a few
11563 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11564 were changed in the original introduction of ENGINE code have now
11565 reverted back - the hooking from this code to ENGINE is now a good
11566 deal more passive and at run-time, operations deal directly with
11567 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11568 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11569 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11570 they were not being used by the framework as there is no concept of a
11571 BIGNUM_METHOD and they could not be generalised to the new
11572 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11573 ENGINE_cpy() has been removed as it cannot be consistently defined in
11578 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11582 * Change mkdef.pl to sort symbols that get the same entry number,
11583 and make sure the automatically generated functions `ERR_load_*`
11584 become part of libeay.num as well.
11588 * New function SSL_renegotiate_pending(). This returns true once
11589 renegotiation has been requested (either SSL_renegotiate() call
11590 or HelloRequest/ClientHello received from the peer) and becomes
11591 false once a handshake has been completed.
11592 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11593 sends a HelloRequest, but does not ensure that a handshake takes
11594 place. SSL_renegotiate_pending() is useful for checking if the
11595 client has followed the request.)
11599 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11600 By default, clients may request session resumption even during
11601 renegotiation (if session ID contexts permit); with this option,
11602 session resumption is possible only in the first handshake.
11604 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11605 more bits available for options that should not be part of
11606 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11610 * Add some demos for certificate and certificate request creation.
11614 * Make maximum certificate chain size accepted from the peer application
11615 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11616 "Douglas E. Engert" <deengert@anl.gov>.
11620 * Add support for shared libraries for Unixware-7
11621 (Boyd Lynn Gerber <gerberb@zenez.com>).
11625 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11626 be done prior to destruction. Use this to unload error strings from
11627 ENGINEs that load their own error strings. NB: This adds two new API
11628 functions to "get" and "set" this destroy handler in an ENGINE.
11632 * Alter all existing ENGINE implementations (except "openssl" and
11633 "openbsd") to dynamically instantiate their own error strings. This
11634 makes them more flexible to be built both as statically-linked ENGINEs
11635 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11636 Also, add stub code to each that makes building them as self-contained
11637 shared-libraries easier (see README.ENGINE).
11641 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11642 implementations into applications that are completely implemented in
11643 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11644 commands that can be used to configure what shared-library to load and
11645 to control aspects of the way it is handled. Also, made an update to
11646 the README.ENGINE file that brings its information up-to-date and
11647 provides some information and instructions on the "dynamic" ENGINE
11648 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11652 * Make it possible to unload ranges of ERR strings with a new
11653 "ERR_unload_strings" function.
11657 * Add a copy() function to EVP_MD.
11661 * Make EVP_MD routines take a context pointer instead of just the
11662 md_data void pointer.
11666 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11667 that the digest can only process a single chunk of data
11668 (typically because it is provided by a piece of
11669 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11670 is only going to provide a single chunk of data, and hence the
11671 framework needn't accumulate the data for oneshot drivers.
11675 * As with "ERR", make it possible to replace the underlying "ex_data"
11676 functions. This change also alters the storage and management of global
11677 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11678 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11679 index counters. The API functions that use this state have been changed
11680 to take a "class_index" rather than pointers to the class's local STACK
11681 and counter, and there is now an API function to dynamically create new
11682 classes. This centralisation allows us to (a) plug a lot of the
11683 thread-safety problems that existed, and (b) makes it possible to clean
11684 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11685 such data would previously have always leaked in application code and
11686 workarounds were in place to make the memory debugging turn a blind eye
11687 to it. Application code that doesn't use this new function will still
11688 leak as before, but their memory debugging output will announce it now
11689 rather than letting it slide.
11691 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11692 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11693 has a return value to indicate success or failure.
11697 * Make it possible to replace the underlying "ERR" functions such that the
11698 global state (2 LHASH tables and 2 locks) is only used by the "default"
11699 implementation. This change also adds two functions to "get" and "set"
11700 the implementation prior to it being automatically set the first time
11701 any other ERR function takes place. Ie. an application can call "get",
11702 pass the return value to a module it has just loaded, and that module
11703 can call its own "set" function using that value. This means the
11704 module's "ERR" operations will use (and modify) the error state in the
11705 application and not in its own statically linked copy of OpenSSL code.
11709 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11710 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11711 the operation, and provides a more encapsulated way for external code
11712 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11713 to use these functions rather than manually incrementing the counts.
11715 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11719 * Add EVP test program.
11723 * Add symmetric cipher support to ENGINE. Expect the API to change!
11727 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11728 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11729 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11730 These allow a CRL to be built without having to access X509_CRL fields
11731 directly. Modify 'ca' application to use new functions.
11735 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11736 bug workarounds. Rollback attack detection is a security feature.
11737 The problem will only arise on OpenSSL servers when TLSv1 is not
11738 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11739 Software authors not wanting to support TLSv1 will have special reasons
11740 for their choice and can explicitly enable this option.
11742 *Bodo Moeller, Lutz Jaenicke*
11744 * Rationalise EVP so it can be extended: don't include a union of
11745 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11746 (similar to those existing for EVP_CIPHER_CTX).
11751 EVP_MD_CTX_init(&md); /* new function call */
11752 EVP_DigestInit(&md, EVP_sha1());
11753 EVP_DigestUpdate(&md, in, len);
11754 EVP_DigestFinal(&md, out, NULL);
11755 EVP_MD_CTX_cleanup(&md); /* new function call */
11759 * Make DES key schedule conform to the usual scheme, as well as
11760 correcting its structure. This means that calls to DES functions
11761 now have to pass a pointer to a des_key_schedule instead of a
11762 plain des_key_schedule (which was actually always a pointer
11765 des_key_schedule ks;
11767 des_set_key_checked(..., &ks);
11768 des_ncbc_encrypt(..., &ks, ...);
11770 (Note that a later change renames 'des_...' into 'DES_...'.)
11774 * Initial reduction of linker bloat: the use of some functions, such as
11775 PEM causes large amounts of unused functions to be linked in due to
11776 poor organisation. For example pem_all.c contains every PEM function
11777 which has a knock on effect of linking in large amounts of (unused)
11778 ASN1 code. Grouping together similar functions and splitting unrelated
11779 functions prevents this.
11783 * Cleanup of EVP macros.
11787 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11788 correct `_ecb suffix`.
11792 * Add initial OCSP responder support to ocsp application. The
11793 revocation information is handled using the text based index
11794 use by the ca application. The responder can either handle
11795 requests generated internally, supplied in files (for example
11796 via a CGI script) or using an internal minimal server.
11800 * Add configuration choices to get zlib compression for TLS.
11804 * Changes to Kerberos SSL for RFC 2712 compliance:
11805 1. Implemented real KerberosWrapper, instead of just using
11806 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11807 2. Implemented optional authenticator field of KerberosWrapper.
11809 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11810 and authenticator structs; see crypto/krb5/.
11812 Generalized Kerberos calls to support multiple Kerberos libraries.
11813 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11814 via Richard Levitte*
11816 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11817 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11818 values for each of the key sizes rather than having just
11819 parameters (and 'speed' generating keys each time).
11823 * Speed up EVP routines.
11826 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11827 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11828 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11829 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11831 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11832 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11833 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11836 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11838 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11842 * Added the OS2-EMX target.
11844 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11846 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11847 to support NCONF routines in extension code. New function CONF_set_nconf()
11848 to allow functions which take an NCONF to also handle the old LHASH
11849 structure: this means that the old CONF compatible routines can be
11850 retained (in particular wrt extensions) without having to duplicate the
11851 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11855 * Enhance the general user interface with mechanisms for inner control
11856 and with possibilities to have yes/no kind of prompts.
11860 * Change all calls to low level digest routines in the library and
11861 applications to use EVP. Add missing calls to HMAC_cleanup() and
11862 don't assume HMAC_CTX can be copied using memcpy().
11864 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11866 * Add the possibility to control engines through control names but with
11867 arbitrary arguments instead of just a string.
11868 Change the key loaders to take a UI_METHOD instead of a callback
11869 function pointer. NOTE: this breaks binary compatibility with earlier
11870 versions of OpenSSL [engine].
11871 Adapt the nCipher code for these new conditions and add a card insertion
11876 * Enhance the general user interface with mechanisms to better support
11877 dialog box interfaces, application-defined prompts, the possibility
11878 to use defaults (for example default passwords from somewhere else)
11879 and interrupts/cancellations.
11883 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11884 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11888 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11889 tidy up some unnecessarily weird code in 'sk_new()').
11891 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11893 * Change the key loading routines for ENGINEs to use the same kind
11894 callback (pem_password_cb) as all other routines that need this
11899 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11900 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11901 than this minimum value is recommended.
11905 * New random seeder for OpenVMS, using the system process statistics
11906 that are easily reachable.
11910 * Windows apparently can't transparently handle global
11911 variables defined in DLLs. Initialisations such as:
11913 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11915 won't compile. This is used by the any applications that need to
11916 declare their own ASN1 modules. This was fixed by adding the option
11917 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11918 needed for static libraries under Win32.
11922 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11923 setting of purpose and trust fields. New X509_STORE trust and
11924 purpose functions and tidy up setting in other SSL functions.
11928 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11929 structure. These are inherited by X509_STORE_CTX when it is
11930 initialised. This allows various defaults to be set in the
11931 X509_STORE structure (such as flags for CRL checking and custom
11932 purpose or trust settings) for functions which only use X509_STORE_CTX
11933 internally such as S/MIME.
11935 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11936 trust settings if they are not set in X509_STORE. This allows X509_STORE
11937 purposes and trust (in S/MIME for example) to override any set by default.
11939 Add command line options for CRL checking to smime, s_client and s_server
11944 * Initial CRL based revocation checking. If the CRL checking flag(s)
11945 are set then the CRL is looked up in the X509_STORE structure and
11946 its validity and signature checked, then if the certificate is found
11947 in the CRL the verify fails with a revoked error.
11949 Various new CRL related callbacks added to X509_STORE_CTX structure.
11951 Command line options added to 'verify' application to support this.
11953 This needs some additional work, such as being able to handle multiple
11954 CRLs with different times, extension based lookup (rather than just
11955 by subject name) and ultimately more complete V2 CRL extension
11960 * Add a general user interface API (crypto/ui/). This is designed
11961 to replace things like des_read_password and friends (backward
11962 compatibility functions using this new API are provided).
11963 The purpose is to remove prompting functions from the DES code
11964 section as well as provide for prompting through dialog boxes in
11965 a window system and the like.
11969 * Add "ex_data" support to ENGINE so implementations can add state at a
11970 per-structure level rather than having to store it globally.
11974 * Make it possible for ENGINE structures to be copied when retrieved by
11975 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11976 This causes the "original" ENGINE structure to act like a template,
11977 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11978 operational state can be localised to each ENGINE structure, despite the
11979 fact they all share the same "methods". New ENGINE structures returned in
11980 this case have no functional references and the return value is the single
11981 structural reference. This matches the single structural reference returned
11982 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11987 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11988 needs to match any other type at all we need to manually clear the
11993 * Changes to the "openssl engine" utility to include;
11994 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11995 about an ENGINE's available control commands.
11996 - executing control commands from command line arguments using the
11997 '-pre' and '-post' switches. '-post' is only used if '-t' is
11998 specified and the ENGINE is successfully initialised. The syntax for
11999 the individual commands are colon-separated, for example;
12000 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12004 * New dynamic control command support for ENGINEs. ENGINEs can now
12005 declare their own commands (numbers), names (strings), descriptions,
12006 and input types for run-time discovery by calling applications. A
12007 subset of these commands are implicitly classed as "executable"
12008 depending on their input type, and only these can be invoked through
12009 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12010 can be based on user input, config files, etc). The distinction is
12011 that "executable" commands cannot return anything other than a boolean
12012 result and can only support numeric or string input, whereas some
12013 discoverable commands may only be for direct use through
12014 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12015 pointers, or other custom uses. The "executable" commands are to
12016 support parameterisations of ENGINE behaviour that can be
12017 unambiguously defined by ENGINEs and used consistently across any
12018 OpenSSL-based application. Commands have been added to all the
12019 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12020 control over shared-library paths without source code alterations.
12024 * Changed all ENGINE implementations to dynamically allocate their
12025 ENGINEs rather than declaring them statically. Apart from this being
12026 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12027 this also allows the implementations to compile without using the
12028 internal engine_int.h header.
12032 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12033 'const' value. Any code that should be able to modify a RAND_METHOD
12034 should already have non-const pointers to it (ie. they should only
12035 modify their own ones).
12039 * Made a variety of little tweaks to the ENGINE code.
12040 - "atalla" and "ubsec" string definitions were moved from header files
12041 to C code. "nuron" string definitions were placed in variables
12042 rather than hard-coded - allowing parameterisation of these values
12043 later on via ctrl() commands.
12044 - Removed unused "#if 0"'d code.
12045 - Fixed engine list iteration code so it uses ENGINE_free() to release
12046 structural references.
12047 - Constified the RAND_METHOD element of ENGINE structures.
12048 - Constified various get/set functions as appropriate and added
12049 missing functions (including a catch-all ENGINE_cpy that duplicates
12050 all ENGINE values onto a new ENGINE except reference counts/state).
12051 - Removed NULL parameter checks in get/set functions. Setting a method
12052 or function to NULL is a way of cancelling out a previously set
12053 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12054 and doesn't justify the extra error symbols and code.
12055 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12056 flags from engine_int.h to engine.h.
12057 - Changed prototypes for ENGINE handler functions (init(), finish(),
12058 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12062 * Implement binary inversion algorithm for BN_mod_inverse in addition
12063 to the algorithm using long division. The binary algorithm can be
12064 used only if the modulus is odd. On 32-bit systems, it is faster
12065 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12066 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12067 up to 450 bits. In 64-bit environments, the binary algorithm
12068 appears to be advantageous for much longer moduli; here we use it
12069 for moduli up to 2048 bits.
12073 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12074 could not support the combine flag in choice fields.
12078 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12079 extensions from a certificate request to the certificate.
12083 * Allow multiple 'certopt' and 'nameopt' options to be separated
12084 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12085 file: this allows the display of the certificate about to be
12086 signed to be customised, to allow certain fields to be included
12087 or excluded and extension details. The old system didn't display
12088 multicharacter strings properly, omitted fields not in the policy
12089 and couldn't display additional details such as extensions.
12093 * Function EC_POINTs_mul for multiple scalar multiplication
12094 of an arbitrary number of elliptic curve points
12095 \sum scalars[i]*points[i],
12096 optionally including the generator defined for the EC_GROUP:
12097 scalar*generator + \sum scalars[i]*points[i].
12099 EC_POINT_mul is a simple wrapper function for the typical case
12100 that the point list has just one item (besides the optional
12105 * First EC_METHODs for curves over GF(p):
12107 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12108 operations and provides various method functions that can also
12109 operate with faster implementations of modular arithmetic.
12111 EC_GFp_mont_method() reuses most functions that are part of
12112 EC_GFp_simple_method, but uses Montgomery arithmetic.
12114 *Bodo Moeller; point addition and point doubling
12115 implementation directly derived from source code provided by
12116 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12118 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12119 crypto/ec/ec_lib.c):
12121 Curves are EC_GROUP objects (with an optional group generator)
12122 based on EC_METHODs that are built into the library.
12124 Points are EC_POINT objects based on EC_GROUP objects.
12126 Most of the framework would be able to handle curves over arbitrary
12127 finite fields, but as there are no obvious types for fields other
12128 than GF(p), some functions are limited to that for now.
12132 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12133 that the file contains a complete HTTP response.
12137 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12138 change the def and num file printf format specifier from "%-40sXXX"
12139 to "%-39s XXX". The latter will always guarantee a space after the
12140 field while the former will cause them to run together if the field
12141 is 40 of more characters long.
12145 * Constify the cipher and digest 'method' functions and structures
12146 and modify related functions to take constant EVP_MD and EVP_CIPHER
12151 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12152 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12156 * Modify `EVP_Digest*()` routines so they now return values. Although the
12157 internal software routines can never fail additional hardware versions
12162 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12164 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12165 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12168 ERR_R_NESTED_ASN1_ERROR
12170 ERR_R_MISSING_ASN1_EOS
12171 were 4 .. 9, conflicting with
12172 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12174 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12175 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12177 Add new error code 'ERR_R_INTERNAL_ERROR'.
12181 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12186 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12187 sets the subject name for a new request or supersedes the
12188 subject name in a given request. Formats that can be parsed are
12189 'CN=Some Name, OU=myOU, C=IT'
12191 'CN=Some Name/OU=myOU/C=IT'.
12193 Add options '-batch' and '-verbose' to 'openssl req'.
12195 *Massimiliano Pala <madwolf@hackmasters.net>*
12197 * Introduce the possibility to access global variables through
12198 functions on platform were that's the best way to handle exporting
12199 global variables in shared libraries. To enable this functionality,
12200 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12201 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12202 is normally done by Configure or something similar).
12204 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12205 in the source file (foo.c) like this:
12207 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12208 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12210 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12211 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12213 OPENSSL_DECLARE_GLOBAL(int,foo);
12214 #define foo OPENSSL_GLOBAL_REF(foo)
12215 OPENSSL_DECLARE_GLOBAL(double,bar);
12216 #define bar OPENSSL_GLOBAL_REF(bar)
12218 The #defines are very important, and therefore so is including the
12219 header file everywhere where the defined globals are used.
12221 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12222 of ASN.1 items, but that structure is a bit different.
12224 The largest change is in util/mkdef.pl which has been enhanced with
12225 better and easier to understand logic to choose which symbols should
12226 go into the Windows .def files as well as a number of fixes and code
12227 cleanup (among others, algorithm keywords are now sorted
12228 lexicographically to avoid constant rewrites).
12232 * In BN_div() keep a copy of the sign of 'num' before writing the
12233 result to 'rm' because if rm==num the value will be overwritten
12234 and produce the wrong result if 'num' is negative: this caused
12235 problems with BN_mod() and BN_nnmod().
12239 * Function OCSP_request_verify(). This checks the signature on an
12240 OCSP request and verifies the signer certificate. The signer
12241 certificate is just checked for a generic purpose and OCSP request
12246 * Add OCSP_check_validity() function to check the validity of OCSP
12247 responses. OCSP responses are prepared in real time and may only
12248 be a few seconds old. Simply checking that the current time lies
12249 between thisUpdate and nextUpdate max reject otherwise valid responses
12250 caused by either OCSP responder or client clock inaccuracy. Instead
12251 we allow thisUpdate and nextUpdate to fall within a certain period of
12252 the current time. The age of the response can also optionally be
12253 checked. Two new options -validity_period and -status_age added to
12258 * If signature or public key algorithm is unrecognized print out its
12259 OID rather that just UNKNOWN.
12263 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12264 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12265 ID to be generated from the issuer certificate alone which can then be
12266 passed to OCSP_id_issuer_cmp().
12270 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12271 ASN1 modules to export functions returning ASN1_ITEM pointers
12272 instead of the ASN1_ITEM structures themselves. This adds several
12273 new macros which allow the underlying ASN1 function/structure to
12274 be accessed transparently. As a result code should not use ASN1_ITEM
12275 references directly (such as &X509_it) but instead use the relevant
12276 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12277 use of the new ASN1 code on platforms where exporting structures
12278 is problematical (for example in shared libraries) but exporting
12279 functions returning pointers to structures is not.
12283 * Add support for overriding the generation of SSL/TLS session IDs.
12284 These callbacks can be registered either in an SSL_CTX or per SSL.
12285 The purpose of this is to allow applications to control, if they wish,
12286 the arbitrary values chosen for use as session IDs, particularly as it
12287 can be useful for session caching in multiple-server environments. A
12288 command-line switch for testing this (and any client code that wishes
12289 to use such a feature) has been added to "s_server".
12291 *Geoff Thorpe, Lutz Jaenicke*
12293 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12294 of the form '#if defined(...) || defined(...) || ...' and
12295 '#if !defined(...) && !defined(...) && ...'. This also avoids
12296 the growing number of special cases it was previously handling.
12300 * Make all configuration macros available for application by making
12301 sure they are available in opensslconf.h, by giving them names starting
12302 with `OPENSSL_` to avoid conflicts with other packages and by making
12303 sure e_os2.h will cover all platform-specific cases together with
12305 Additionally, it is now possible to define configuration/platform-
12306 specific names (called "system identities"). In the C code, these
12307 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12308 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12309 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12314 * New option -set_serial to 'req' and 'x509' this allows the serial
12315 number to use to be specified on the command line. Previously self
12316 signed certificates were hard coded with serial number 0 and the
12317 CA options of 'x509' had to use a serial number in a file which was
12322 * New options to 'ca' utility to support V2 CRL entry extensions.
12323 Currently CRL reason, invalidity date and hold instruction are
12324 supported. Add new CRL extensions to V3 code and some new objects.
12328 * New function EVP_CIPHER_CTX_set_padding() this is used to
12329 disable standard block padding (aka PKCS#5 padding) in the EVP
12330 API, which was previously mandatory. This means that the data is
12331 not padded in any way and so the total length much be a multiple
12332 of the block size, otherwise an error occurs.
12336 * Initial (incomplete) OCSP SSL support.
12340 * New function OCSP_parse_url(). This splits up a URL into its host,
12341 port and path components: primarily to parse OCSP URLs. New -url
12342 option to ocsp utility.
12346 * New nonce behavior. The return value of OCSP_check_nonce() now
12347 reflects the various checks performed. Applications can decide
12348 whether to tolerate certain situations such as an absent nonce
12349 in a response when one was present in a request: the ocsp application
12350 just prints out a warning. New function OCSP_add1_basic_nonce()
12351 this is to allow responders to include a nonce in a response even if
12352 the request is nonce-less.
12356 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12357 skipped when using openssl x509 multiple times on a single input file,
12358 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12362 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12363 set string type: to handle setting ASN1_TIME structures. Fix ca
12364 utility to correctly initialize revocation date of CRLs.
12368 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12369 the clients preferred ciphersuites and rather use its own preferences.
12370 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12371 Internet Explorer by ensuring unchanged hash method during stepup.
12372 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12376 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12377 to aes and add a new 'exist' option to print out symbols that don't
12382 * Additional options to ocsp utility to allow flags to be set and
12383 additional certificates supplied.
12387 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12388 OCSP client a number of certificate to only verify the response
12393 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12394 handle the new API. Currently only ECB, CBC modes supported. Add new
12397 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12398 Encryption Standard (AES) Ciphersuites for Transport Layer
12399 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12400 not enabled by default and were not part of the "ALL" ciphersuite
12401 alias because they were not yet official; they could be
12402 explicitly requested by specifying the "AESdraft" ciphersuite
12403 group alias. In the final release of OpenSSL 0.9.7, the group
12404 alias is called "AES" and is part of "ALL".)
12406 *Ben Laurie, Steve Henson, Bodo Moeller*
12408 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12409 request to response.
12413 * Functions for OCSP responders. OCSP_request_onereq_count(),
12414 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12415 extract information from a certificate request. OCSP_response_create()
12416 creates a response and optionally adds a basic response structure.
12417 OCSP_basic_add1_status() adds a complete single response to a basic
12418 response and returns the OCSP_SINGLERESP structure just added (to allow
12419 extensions to be included for example). OCSP_basic_add1_cert() adds a
12420 certificate to a basic response and OCSP_basic_sign() signs a basic
12421 response with various flags. New helper functions ASN1_TIME_check()
12422 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12423 (converts ASN1_TIME to GeneralizedTime).
12427 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12428 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12429 structure from a certificate. X509_pubkey_digest() digests the public_key
12430 contents: this is used in various key identifiers.
12434 * Make sk_sort() tolerate a NULL argument.
12436 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12438 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12439 passed by the function are trusted implicitly. If any of them signed the
12440 response then it is assumed to be valid and is not verified.
12444 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12445 to data. This was previously part of the PKCS7 ASN1 code. This
12446 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12447 *Steve Henson, reported by Kenneth R. Robinette
12448 <support@securenetterm.com>*
12450 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12451 routines: without these tracing memory leaks is very painful.
12452 Fix leaks in PKCS12 and PKCS7 routines.
12456 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12457 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12458 effectively meant GeneralizedTime would never be used. Now it
12459 is initialised to -1 but X509_time_adj() now has to check the value
12460 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12461 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12462 *Steve Henson, reported by Kenneth R. Robinette
12463 <support@securenetterm.com>*
12465 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12466 result in a zero length in the ASN1_INTEGER structure which was
12467 not consistent with the structure when d2i_ASN1_INTEGER() was used
12468 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12469 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12470 where it did not print out a minus for negative ASN1_INTEGER.
12474 * Add summary printout to ocsp utility. The various functions which
12475 convert status values to strings have been renamed to:
12476 OCSP_response_status_str(), OCSP_cert_status_str() and
12477 OCSP_crl_reason_str() and are no longer static. New options
12478 to verify nonce values and to disable verification. OCSP response
12479 printout format cleaned up.
12483 * Add additional OCSP certificate checks. These are those specified
12484 in RFC2560. This consists of two separate checks: the CA of the
12485 certificate being checked must either be the OCSP signer certificate
12486 or the issuer of the OCSP signer certificate. In the latter case the
12487 OCSP signer certificate must contain the OCSP signing extended key
12488 usage. This check is performed by attempting to match the OCSP
12489 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12490 in the OCSP_CERTID structures of the response.
12494 * Initial OCSP certificate verification added to OCSP_basic_verify()
12495 and related routines. This uses the standard OpenSSL certificate
12496 verify routines to perform initial checks (just CA validity) and
12497 to obtain the certificate chain. Then additional checks will be
12498 performed on the chain. Currently the root CA is checked to see
12499 if it is explicitly trusted for OCSP signing. This is used to set
12500 a root CA as a global signing root: that is any certificate that
12501 chains to that CA is an acceptable OCSP signing certificate.
12505 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12506 extensions from a separate configuration file.
12507 As when reading extensions from the main configuration file,
12508 the '-extensions ...' option may be used for specifying the
12511 *Massimiliano Pala <madwolf@comune.modena.it>*
12513 * New OCSP utility. Allows OCSP requests to be generated or
12514 read. The request can be sent to a responder and the output
12515 parsed, outputted or printed in text form. Not complete yet:
12516 still needs to check the OCSP response validity.
12520 * New subcommands for 'openssl ca':
12521 `openssl ca -status <serial>` prints the status of the cert with
12522 the given serial number (according to the index file).
12523 `openssl ca -updatedb` updates the expiry status of certificates
12526 *Massimiliano Pala <madwolf@comune.modena.it>*
12528 * New '-newreq-nodes' command option to CA.pl. This is like
12529 '-newreq', but calls 'openssl req' with the '-nodes' option
12530 so that the resulting key is not encrypted.
12532 *Damien Miller <djm@mindrot.org>*
12534 * New configuration for the GNU Hurd.
12536 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12538 * Initial code to implement OCSP basic response verify. This
12539 is currently incomplete. Currently just finds the signer's
12540 certificate and verifies the signature on the response.
12544 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12545 value of OPENSSLDIR. This is available via the new '-d' option
12546 to 'openssl version', and is also included in 'openssl version -a'.
12550 * Allowing defining memory allocation callbacks that will be given
12551 file name and line number information in additional arguments
12552 (a `const char*` and an int). The basic functionality remains, as
12553 well as the original possibility to just replace malloc(),
12554 realloc() and free() by functions that do not know about these
12555 additional arguments. To register and find out the current
12556 settings for extended allocation functions, the following
12557 functions are provided:
12559 CRYPTO_set_mem_ex_functions
12560 CRYPTO_set_locked_mem_ex_functions
12561 CRYPTO_get_mem_ex_functions
12562 CRYPTO_get_locked_mem_ex_functions
12564 These work the same way as CRYPTO_set_mem_functions and friends.
12565 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12566 extended allocation function is enabled.
12567 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12568 a conventional allocation function is enabled.
12570 *Richard Levitte, Bodo Moeller*
12572 * Finish off removing the remaining LHASH function pointer casts.
12573 There should no longer be any prototype-casting required when using
12574 the LHASH abstraction, and any casts that remain are "bugs". See
12575 the callback types and macros at the head of lhash.h for details
12576 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12580 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12581 If /dev/[u]random devices are not available or do not return enough
12582 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12584 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12585 /etc/entropy will be queried once each in this sequence, querying stops
12586 when enough entropy was collected without querying more sockets.
12590 * Change the Unix RAND_poll() variant to be able to poll several
12591 random devices, as specified by DEVRANDOM, until a sufficient amount
12592 of data has been collected. We spend at most 10 ms on each file
12593 (select timeout) and read in non-blocking mode. DEVRANDOM now
12594 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12595 (previously it was just the string "/dev/urandom"), so on typical
12596 platforms the 10 ms delay will never occur.
12597 Also separate out the Unix variant to its own file, rand_unix.c.
12598 For VMS, there's a currently-empty rand_vms.c.
12602 * Move OCSP client related routines to ocsp_cl.c. These
12603 provide utility functions which an application needing
12604 to issue a request to an OCSP responder and analyse the
12605 response will typically need: as opposed to those which an
12606 OCSP responder itself would need which will be added later.
12608 OCSP_request_sign() signs an OCSP request with an API similar
12609 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12610 response. OCSP_response_get1_basic() extracts basic response
12611 from response. OCSP_resp_find_status(): finds and extracts status
12612 information from an OCSP_CERTID structure (which will be created
12613 when the request structure is built). These are built from lower
12614 level functions which work on OCSP_SINGLERESP structures but
12615 won't normally be used unless the application wishes to examine
12616 extensions in the OCSP response for example.
12618 Replace nonce routines with a pair of functions.
12619 OCSP_request_add1_nonce() adds a nonce value and optionally
12620 generates a random value. OCSP_check_nonce() checks the
12621 validity of the nonce in an OCSP response.
12625 * Change function OCSP_request_add() to OCSP_request_add0_id().
12626 This doesn't copy the supplied OCSP_CERTID and avoids the
12627 need to free up the newly created id. Change return type
12628 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12629 This can then be used to add extensions to the request.
12630 Deleted OCSP_request_new(), since most of its functionality
12631 is now in OCSP_REQUEST_new() (and the case insensitive name
12632 clash) apart from the ability to set the request name which
12633 will be added elsewhere.
12637 * Update OCSP API. Remove obsolete extensions argument from
12638 various functions. Extensions are now handled using the new
12639 OCSP extension code. New simple OCSP HTTP function which
12640 can be used to send requests and parse the response.
12644 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12645 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12646 uses the special reorder version of SET OF to sort the attributes
12647 and reorder them to match the encoded order. This resolves a long
12648 standing problem: a verify on a PKCS7 structure just after signing
12649 it used to fail because the attribute order did not match the
12650 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12651 it uses the received order. This is necessary to tolerate some broken
12652 software that does not order SET OF. This is handled by encoding
12653 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12654 to produce the required SET OF.
12658 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12659 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12660 files to get correct declarations of the ASN.1 item variables.
12664 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12665 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12666 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12667 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12668 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12669 ASN1_ITEM and no wrapper functions.
12673 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12674 replace the old function pointer based I/O routines. Change most of
12675 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12679 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12680 lines, recognize more "algorithms" that can be deselected, and make
12681 it complain about algorithm deselection that isn't recognised.
12685 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12686 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12687 to use new functions. Add NO_ASN1_OLD which can be set to remove
12688 some old style ASN1 functions: this can be used to determine if old
12689 code will still work when these eventually go away.
12693 * New extension functions for OCSP structures, these follow the
12694 same conventions as certificates and CRLs.
12698 * New function X509V3_add1_i2d(). This automatically encodes and
12699 adds an extension. Its behaviour can be customised with various
12700 flags to append, replace or delete. Various wrappers added for
12701 certificates and CRLs.
12705 * Fix to avoid calling the underlying ASN1 print routine when
12706 an extension cannot be parsed. Correct a typo in the
12707 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12711 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12712 entries for variables.
12716 * Add functionality to apps/openssl.c for detecting locking
12717 problems: As the program is single-threaded, all we have
12718 to do is register a locking callback using an array for
12719 storing which locks are currently held by the program.
12723 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12724 SSL_get_ex_data_X509_STORE_idx(), which is used in
12725 ssl_verify_cert_chain() and thus can be called at any time
12726 during TLS/SSL handshakes so that thread-safety is essential.
12727 Unfortunately, the ex_data design is not at all suited
12728 for multi-threaded use, so it probably should be abolished.
12732 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12734 *Broadcom, tweaked and integrated by Geoff Thorpe*
12736 * Move common extension printing code to new function
12737 X509V3_print_extensions(). Reorganise OCSP print routines and
12738 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12742 * New function X509_signature_print() to remove duplication in some
12747 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12748 set (this was treated exactly the same as SET OF previously). This
12749 is used to reorder the STACK representing the structure to match the
12750 encoding. This will be used to get round a problem where a PKCS7
12751 structure which was signed could not be verified because the STACK
12752 order did not reflect the encoded order.
12756 * Reimplement the OCSP ASN1 module using the new code.
12760 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12761 for its ASN1 operations. The old style function pointers still exist
12762 for now but they will eventually go away.
12766 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12767 completely replaces the old ASN1 functionality with a table driven
12768 encoder and decoder which interprets an ASN1_ITEM structure describing
12769 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12770 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12771 has also been converted to the new form.
12775 * Change BN_mod_exp_recp so that negative moduli are tolerated
12776 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12777 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12778 for negative moduli.
12782 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12783 of not touching the result's sign bit.
12787 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12792 * Changed the LHASH code to use prototypes for callbacks, and created
12793 macros to declare and implement thin (optionally static) functions
12794 that provide type-safety and avoid function pointer casting for the
12795 type-specific callbacks.
12799 * Added Kerberos Cipher Suites to be used with TLS, as written in
12801 *Veers Staats <staatsvr@asc.hpc.mil>,
12802 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12804 * Reformat the FAQ so the different questions and answers can be divided
12805 in sections depending on the subject.
12809 * Have the zlib compression code load ZLIB.DLL dynamically under
12814 * New function BN_mod_sqrt for computing square roots modulo a prime
12815 (using the probabilistic Tonelli-Shanks algorithm unless
12816 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12817 be handled deterministically).
12819 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12821 * Make BN_mod_inverse faster by explicitly handling small quotients
12822 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12823 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12827 * New function BN_kronecker.
12831 * Fix BN_gcd so that it works on negative inputs; the result is
12832 positive unless both parameters are zero.
12833 Previously something reasonably close to an infinite loop was
12834 possible because numbers could be growing instead of shrinking
12835 in the implementation of Euclid's algorithm.
12839 * Fix BN_is_word() and BN_is_one() macros to take into account the
12840 sign of the number in question.
12842 Fix BN_is_word(a,w) to work correctly for w == 0.
12844 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12845 because its test if the absolute value of 'a' equals 'w'.
12846 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12847 it exists mostly for use in the implementations of BN_is_zero(),
12848 BN_is_one(), and BN_is_word().
12852 * New function BN_swap.
12856 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12857 the exponentiation functions are more likely to produce reasonable
12858 results on negative inputs.
12862 * Change BN_mod_mul so that the result is always non-negative.
12863 Previously, it could be negative if one of the factors was negative;
12864 I don't think anyone really wanted that behaviour.
12868 * Move `BN_mod_...` functions into new file crypto/bn/bn_mod.c
12869 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12870 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12871 and add new functions:
12880 BN_mod_lshift1_quick
12882 BN_mod_lshift_quick
12884 These functions always generate non-negative results.
12886 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12887 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12889 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12890 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12891 be reduced modulo m.
12893 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12896 The following entry accidentally appeared in the CHANGES file
12897 distributed with OpenSSL 0.9.7. The modifications described in
12898 it do *not* apply to OpenSSL 0.9.7.
12900 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12901 was actually never needed) and in BN_mul(). The removal in BN_mul()
12902 required a small change in bn_mul_part_recursive() and the addition
12903 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12904 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12905 bn_sub_words() and bn_add_words() except they take arrays with
12911 * In 'openssl passwd', verify passwords read from the terminal
12912 unless the '-salt' option is used (which usually means that
12913 verification would just waste user's time since the resulting
12914 hash is going to be compared with some given password hash)
12915 or the new '-noverify' option is used.
12917 This is an incompatible change, but it does not affect
12918 non-interactive use of 'openssl passwd' (passwords on the command
12919 line, '-stdin' option, '-in ...' option) and thus should not
12920 cause any problems.
12924 * Remove all references to RSAref, since there's no more need for it.
12928 * Make DSO load along a path given through an environment variable
12929 (SHLIB_PATH) with shl_load().
12933 * Constify the ENGINE code as a result of BIGNUM constification.
12934 Also constify the RSA code and most things related to it. In a
12935 few places, most notable in the depth of the ASN.1 code, ugly
12936 casts back to non-const were required (to be solved at a later
12941 * Make it so the openssl application has all engines loaded by default.
12945 * Constify the BIGNUM routines a little more.
12949 * Add the following functions:
12951 ENGINE_load_cswift()
12953 ENGINE_load_atalla()
12954 ENGINE_load_nuron()
12955 ENGINE_load_builtin_engines()
12957 That way, an application can itself choose if external engines that
12958 are built-in in OpenSSL shall ever be used or not. The benefit is
12959 that applications won't have to be linked with libdl or other dso
12960 libraries unless it's really needed.
12962 Changed 'openssl engine' to load all engines on demand.
12963 Changed the engine header files to avoid the duplication of some
12964 declarations (they differed!).
12968 * 'openssl engine' can now list capabilities.
12972 * Better error reporting in 'openssl engine'.
12976 * Never call load_dh_param(NULL) in s_server.
12980 * Add engine application. It can currently list engines by name and
12981 identity, and test if they are actually available.
12985 * Improve RPM specification file by forcing symbolic linking and making
12986 sure the installed documentation is also owned by root.root.
12988 *Damien Miller <djm@mindrot.org>*
12990 * Give the OpenSSL applications more possibilities to make use of
12991 keys (public as well as private) handled by engines.
12995 * Add OCSP code that comes from CertCo.
12999 * Add VMS support for the Rijndael code.
13003 * Added untested support for Nuron crypto accelerator.
13007 * Add support for external cryptographic devices. This code was
13008 previously distributed separately as the "engine" branch.
13010 *Geoff Thorpe, Richard Levitte*
13012 * Rework the filename-translation in the DSO code. It is now possible to
13013 have far greater control over how a "name" is turned into a filename
13014 depending on the operating environment and any oddities about the
13015 different shared library filenames on each system.
13019 * Support threads on FreeBSD-elf in Configure.
13023 * Fix for SHA1 assembly problem with MASM: it produces
13024 warnings about corrupt line number information when assembling
13025 with debugging information. This is caused by the overlapping
13028 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13031 NCONF_get_number() has no error checking at all. As a replacement,
13032 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13033 promoted strongly. The old NCONF_get_number is kept around for
13034 binary backward compatibility.
13035 Make it possible for methods to load from something other than a BIO,
13036 by providing a function pointer that is given a name instead of a BIO.
13037 For example, this could be used to load configuration data from an
13042 * Fix for non blocking accept BIOs. Added new I/O special reason
13043 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13044 with non blocking I/O was not possible because no retry code was
13045 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13050 * Added the beginnings of Rijndael support.
13054 * Fix for bug in DirectoryString mask setting. Add support for
13055 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13056 to allow certificate printing to more controllable, additional
13057 'certopt' option to 'x509' to allow new printing options to be
13062 * Clean old EAY MD5 hack from e_os.h.
13066 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13068 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13069 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13071 *Joe Orton, Steve Henson*
13073 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13075 * Fix additional bug revealed by the NISCC test suite:
13077 Stop bug triggering large recursion when presented with
13078 certain ASN.1 tags [CVE-2003-0851][]
13082 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13084 * Fix various bugs revealed by running the NISCC test suite:
13086 Stop out of bounds reads in the ASN1 code when presented with
13087 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13089 If verify callback ignores invalid public key errors don't try to check
13090 certificate signature with the NULL public key.
13094 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13095 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13100 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13101 extra data after the compression methods not only for TLS 1.0
13102 but also for SSL 3.0 (as required by the specification).
13104 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13106 * Change X509_certificate_type() to mark the key as exported/exportable
13107 when it's 512 *bits* long, not 512 bytes.
13111 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13113 * Countermeasure against the Klima-Pokorny-Rosa extension of
13114 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13115 a protocol version number mismatch like a decryption error
13116 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13120 * Turn on RSA blinding by default in the default implementation
13121 to avoid a timing attack. Applications that don't want it can call
13122 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13123 They would be ill-advised to do so in most cases.
13125 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13127 * Change RSA blinding code so that it works when the PRNG is not
13128 seeded (in this case, the secret RSA exponent is abused as
13129 an unpredictable seed -- if it is not unpredictable, there
13130 is no point in blinding anyway). Make RSA blinding thread-safe
13131 by remembering the creator's thread ID in rsa->blinding and
13132 having all other threads use local one-time blinding factors
13133 (this requires more computation than sharing rsa->blinding, but
13134 avoids excessive locking; and if an RSA object is not shared
13135 between threads, blinding will still be very fast).
13139 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13141 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13142 via timing by performing a MAC computation even if incorrect
13143 block cipher padding has been found. This is a countermeasure
13144 against active attacks where the attacker has to distinguish
13145 between bad padding and a MAC verification error. [CVE-2003-0078][]
13147 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13148 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13149 Martin Vuagnoux (EPFL, Ilion)*
13151 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13153 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13154 memory from its contents. This is done with a counter that will
13155 place alternating values in each byte. This can be used to solve
13156 two issues: 1) the removal of calls to memset() by highly optimizing
13157 compilers, and 2) cleansing with other values than 0, since those can
13158 be read through on certain media, for example a swap space on disk.
13162 * Bugfix: client side session caching did not work with external caching,
13163 because the session->cipher setting was not restored when reloading
13164 from the external cache. This problem was masked, when
13165 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13166 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13170 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13171 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13173 *Zeev Lieber <zeev-l@yahoo.com>*
13175 * Undo an undocumented change introduced in 0.9.6e which caused
13176 repeated calls to OpenSSL_add_all_ciphers() and
13177 OpenSSL_add_all_digests() to be ignored, even after calling
13182 * Change the default configuration reader to deal with last line not
13183 being properly terminated.
13187 * Change X509_NAME_cmp() so it applies the special rules on handling
13188 DN values that are of type PrintableString, as well as RDNs of type
13189 emailAddress where the value has the type ia5String.
13191 *stefank@valicert.com via Richard Levitte*
13193 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13194 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13195 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13196 the bitwise-OR of the two for use by the majority of applications
13197 wanting this behaviour, and update the docs. The documented
13198 behaviour and actual behaviour were inconsistent and had been
13199 changing anyway, so this is more a bug-fix than a behavioural
13202 *Geoff Thorpe, diagnosed by Nadav Har'El*
13204 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13205 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13209 * Fix initialization code race conditions in
13210 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13211 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13212 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13213 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13214 ssl2_get_cipher_by_char(),
13215 ssl3_get_cipher_by_char().
13217 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13219 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13220 the cached sessions are flushed, as the remove_cb() might use ex_data
13221 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13222 (see [openssl.org #212]).
13224 *Geoff Thorpe, Lutz Jaenicke*
13226 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13227 length, instead of the encoding length to d2i_ASN1_OBJECT.
13231 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13233 * [In 0.9.6g-engine release:]
13234 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13236 *Lynn Gazis <lgazis@rainbow.com>*
13238 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13240 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13241 and get fix the header length calculation.
13242 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13243 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13245 * Use proper error handling instead of 'assertions' in buffer
13246 overflow checks added in 0.9.6e. This prevents DoS (the
13247 assertions could call abort()).
13249 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13251 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13253 * Add various sanity checks to asn1_get_length() to reject
13254 the ASN1 length bytes if they exceed sizeof(long), will appear
13255 negative or the content length exceeds the length of the
13258 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13260 * Fix cipher selection routines: ciphers without encryption had no flags
13261 for the cipher strength set and where therefore not handled correctly
13262 by the selection routines (PR #130).
13266 * Fix EVP_dsa_sha macro.
13271 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13272 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13273 that was added in OpenSSL 0.9.6d.
13275 As the countermeasure turned out to be incompatible with some
13276 broken SSL implementations, the new option is part of SSL_OP_ALL.
13277 SSL_OP_ALL is usually employed when compatibility with weird SSL
13278 implementations is desired (e.g. '-bugs' option to 's_client' and
13279 's_server'), so the new option is automatically set in many
13284 * Changes in security patch:
13286 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13287 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13288 Air Force Materiel Command, USAF, under agreement number
13291 * Add various sanity checks to asn1_get_length() to reject
13292 the ASN1 length bytes if they exceed sizeof(long), will appear
13293 negative or the content length exceeds the length of the
13294 supplied buffer. [CVE-2002-0659][]
13296 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13298 * Assertions for various potential buffer overflows, not known to
13299 happen in practice.
13301 *Ben Laurie (CHATS)*
13303 * Various temporary buffers to hold ASCII versions of integers were
13304 too small for 64 bit platforms. [CVE-2002-0655][]
13305 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13307 * Remote buffer overflow in SSL3 protocol - an attacker could
13308 supply an oversized session ID to a client. [CVE-2002-0656][]
13310 *Ben Laurie (CHATS)*
13312 * Remote buffer overflow in SSL2 protocol - an attacker could
13313 supply an oversized client master key. [CVE-2002-0656][]
13315 *Ben Laurie (CHATS)*
13317 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13319 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13320 encoded as NULL) with id-dsa-with-sha1.
13322 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13324 * Check various `X509_...()` return values in apps/req.c.
13326 *Nils Larsch <nla@trustcenter.de>*
13328 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13329 an end-of-file condition would erroneously be flagged, when the CRLF
13330 was just at the end of a processed block. The bug was discovered when
13331 processing data through a buffering memory BIO handing the data to a
13332 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13333 <ptsekov@syntrex.com> and Nedelcho Stanev.
13337 * Implement a countermeasure against a vulnerability recently found
13338 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13339 before application data chunks to avoid the use of known IVs
13340 with data potentially chosen by the attacker.
13344 * Fix length checks in ssl3_get_client_hello().
13348 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13349 to prevent ssl3_read_internal() from incorrectly assuming that
13350 ssl3_read_bytes() found application data while handshake
13351 processing was enabled when in fact s->s3->in_read_app_data was
13352 merely automatically cleared during the initial handshake.
13354 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13356 * Fix object definitions for Private and Enterprise: they were not
13357 recognized in their shortname (=lowercase) representation. Extend
13358 obj_dat.pl to issue an error when using undefined keywords instead
13359 of silently ignoring the problem (Svenning Sorensen
13360 <sss@sss.dnsalias.net>).
13364 * Fix DH_generate_parameters() so that it works for 'non-standard'
13365 generators, i.e. generators other than 2 and 5. (Previously, the
13366 code did not properly initialise the 'add' and 'rem' values to
13367 BN_generate_prime().)
13369 In the new general case, we do not insist that 'generator' is
13370 actually a primitive root: This requirement is rather pointless;
13371 a generator of the order-q subgroup is just as good, if not
13376 * Map new X509 verification errors to alerts. Discovered and submitted by
13377 Tom Wu <tom@arcot.com>.
13381 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13382 returning non-zero before the data has been completely received
13383 when using non-blocking I/O.
13385 *Bodo Moeller; problem pointed out by John Hughes*
13387 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13389 *Ben Laurie, Lutz Jaenicke*
13391 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13392 Yoram Zahavi <YoramZ@gilian.com>).
13396 * Add information about CygWin 1.3 and on, and preserve proper
13397 configuration for the versions before that.
13399 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13401 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13402 check whether we deal with a copy of a session and do not delete from
13403 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13404 <izhar@checkpoint.com>.
13408 * Do not store session data into the internal session cache, if it
13409 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13410 flag is set). Proposed by Aslam <aslam@funk.com>.
13414 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13419 * [In 0.9.6d-engine release:]
13420 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13422 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13424 * Add the configuration target linux-s390x.
13426 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13428 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13429 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13430 variable as an indication that a ClientHello message has been
13431 received. As the flag value will be lost between multiple
13432 invocations of ssl3_accept when using non-blocking I/O, the
13433 function may not be aware that a handshake has actually taken
13434 place, thus preventing a new session from being added to the
13437 To avoid this problem, we now set s->new_session to 2 instead of
13438 using a local variable.
13440 *Lutz Jaenicke, Bodo Moeller*
13442 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13443 if the SSL_R_LENGTH_MISMATCH error is detected.
13445 *Geoff Thorpe, Bodo Moeller*
13447 * New 'shared_ldflag' column in Configure platform table.
13451 * Fix EVP_CIPHER_mode macro.
13453 *"Dan S. Camper" <dan@bti.net>*
13455 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13456 type, we must throw them away by setting rr->length to 0.
13458 *D P Chang <dpc@qualys.com>*
13460 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13462 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13463 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13464 worked incorrectly for those cases where range = `10..._2` and
13465 `3*range` is two bits longer than range.)
13469 * Only add signing time to PKCS7 structures if it is not already
13474 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13475 OBJ_ld_ce should be OBJ_id_ce.
13476 Also some ip-pda OIDs in crypto/objects/objects.txt were
13477 incorrect (cf. RFC 3039).
13479 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13481 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13482 returns early because it has nothing to do.
13484 *Andy Schneider <andy.schneider@bjss.co.uk>*
13486 * [In 0.9.6c-engine release:]
13487 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13489 *Andy Schneider <andy.schneider@bjss.co.uk>*
13491 * [In 0.9.6c-engine release:]
13492 Add support for Cryptographic Appliance's keyserver technology.
13493 (Use engine 'keyclient')
13495 *Cryptographic Appliances and Geoff Thorpe*
13497 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13498 is called via tools/c89.sh because arguments have to be
13499 rearranged (all '-L' options must appear before the first object
13502 *Richard Shapiro <rshapiro@abinitio.com>*
13504 * [In 0.9.6c-engine release:]
13505 Add support for Broadcom crypto accelerator cards, backported
13508 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13510 * [In 0.9.6c-engine release:]
13511 Add support for SureWare crypto accelerator cards from
13512 Baltimore Technologies. (Use engine 'sureware')
13514 *Baltimore Technologies and Mark Cox*
13516 * [In 0.9.6c-engine release:]
13517 Add support for crypto accelerator cards from Accelerated
13518 Encryption Processing, www.aep.ie. (Use engine 'aep')
13520 *AEP Inc. and Mark Cox*
13522 * Add a configuration entry for gcc on UnixWare.
13524 *Gary Benson <gbenson@redhat.com>*
13526 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13527 messages are stored in a single piece (fixed-length part and
13528 variable-length part combined) and fix various bugs found on the way.
13532 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13533 instead. BIO_gethostbyname() does not know what timeouts are
13534 appropriate, so entries would stay in cache even when they have
13536 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13538 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13539 faced with a pathologically small ClientHello fragment that does
13540 not contain client_version: Instead of aborting with an error,
13541 simply choose the highest available protocol version (i.e.,
13542 TLS 1.0 unless it is disabled). In practice, ClientHello
13543 messages are never sent like this, but this change gives us
13544 strictly correct behaviour at least for TLS.
13548 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13549 never resets s->method to s->ctx->method when called from within
13550 one of the SSL handshake functions.
13552 *Bodo Moeller; problem pointed out by Niko Baric*
13554 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13555 (sent using the client's version number) if client_version is
13556 smaller than the protocol version in use. Also change
13557 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13558 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13559 the client will at least see that alert.
13563 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13568 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13569 client receives HelloRequest while in a handshake.
13571 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13573 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13574 should end in 'break', not 'goto end' which circumvents various
13575 cleanups done in state SSL_ST_OK. But session related stuff
13576 must be disabled for SSL_ST_OK in the case that we just sent a
13579 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13580 before just sending a HelloRequest.
13582 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13584 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13585 reveal whether illegal block cipher padding was found or a MAC
13586 verification error occurred. (Neither SSLerr() codes nor alerts
13587 are directly visible to potential attackers, but the information
13588 may leak via logfiles.)
13590 Similar changes are not required for the SSL 2.0 implementation
13591 because the number of padding bytes is sent in clear for SSL 2.0,
13592 and the extra bytes are just ignored. However ssl/s2_pkt.c
13593 failed to verify that the purported number of padding bytes is in
13598 * Add OpenUNIX-8 support including shared libraries
13599 (Boyd Lynn Gerber <gerberb@zenez.com>).
13603 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13604 'wristwatch attack' using huge encoding parameters (cf.
13605 James H. Manger's CRYPTO 2001 paper). Note that the
13606 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13607 encoding parameters and hence was not vulnerable.
13611 * BN_sqr() bug fix.
13613 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13615 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13616 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13617 followed by modular reduction.
13619 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13621 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13622 equivalent based on BN_pseudo_rand() instead of BN_rand().
13626 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13627 This function was broken, as the check for a new client hello message
13628 to handle SGC did not allow these large messages.
13629 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13633 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13637 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13638 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13642 * Rework the configuration and shared library support for Tru64 Unix.
13643 The configuration part makes use of modern compiler features and
13644 still retains old compiler behavior for those that run older versions
13645 of the OS. The shared library support part includes a variant that
13646 uses the RPATH feature, and is available through the special
13647 configuration target "alpha-cc-rpath", which will never be selected
13650 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13652 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13653 with the same message size as in ssl3_get_certificate_request().
13654 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13655 messages might inadvertently be reject as too long.
13657 *Petr Lampa <lampa@fee.vutbr.cz>*
13659 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13663 * Modified SSL library such that the verify_callback that has been set
13664 specifically for an SSL object with SSL_set_verify() is actually being
13665 used. Before the change, a verify_callback set with this function was
13666 ignored and the verify_callback() set in the SSL_CTX at the time of
13667 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13668 to allow the necessary settings.
13672 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13673 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13674 done automatically (in contradiction to the requirements of the C
13675 standard). This made problems when used from OpenSSH.
13679 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13680 dh->length and always used
13682 BN_rand_range(priv_key, dh->p).
13684 BN_rand_range() is not necessary for Diffie-Hellman, and this
13685 specific range makes Diffie-Hellman unnecessarily inefficient if
13686 dh->length (recommended exponent length) is much smaller than the
13687 length of dh->p. We could use BN_rand_range() if the order of
13688 the subgroup was stored in the DH structure, but we only have
13693 BN_rand(priv_key, l, ...)
13695 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13702 RSA_eay_public_encrypt
13703 RSA_eay_private_decrypt
13704 RSA_eay_private_encrypt (signing)
13705 RSA_eay_public_decrypt (signature verification)
13707 (default implementations for RSA_public_encrypt,
13708 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13709 always reject numbers >= n.
13713 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13714 to synchronize access to 'locking_thread'. This is necessary on
13715 systems where access to 'locking_thread' (an 'unsigned long'
13716 variable) is not atomic.
13720 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13721 *before* setting the 'crypto_lock_rand' flag. The previous code had
13722 a race condition if 0 is a valid thread ID.
13724 *Travis Vitek <vitek@roguewave.com>*
13726 * Add support for shared libraries under Irix.
13728 *Albert Chin-A-Young <china@thewrittenword.com>*
13730 * Add configuration option to build on Linux on both big-endian and
13731 little-endian MIPS.
13733 *Ralf Baechle <ralf@uni-koblenz.de>*
13735 * Add the possibility to create shared libraries on HP-UX.
13739 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13741 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13742 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13743 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13744 PRNG state recovery was possible based on the output of
13745 one PRNG request appropriately sized to gain knowledge on
13746 'md' followed by enough consecutive 1-byte PRNG requests
13747 to traverse all of 'state'.
13749 1. When updating 'md_local' (the current thread's copy of 'md')
13750 during PRNG output generation, hash all of the previous
13751 'md_local' value, not just the half used for PRNG output.
13753 2. Make the number of bytes from 'state' included into the hash
13754 independent from the number of PRNG bytes requested.
13756 The first measure alone would be sufficient to avoid
13757 Markku-Juhani's attack. (Actually it had never occurred
13758 to me that the half of 'md_local' used for chaining was the
13759 half from which PRNG output bytes were taken -- I had always
13760 assumed that the secret half would be used.) The second
13761 measure makes sure that additional data from 'state' is never
13762 mixed into 'md_local' in small portions; this heuristically
13763 further strengthens the PRNG.
13767 * Fix crypto/bn/asm/mips3.s.
13771 * When only the key is given to "enc", the IV is undefined. Print out
13772 an error message in this case.
13776 * Handle special case when X509_NAME is empty in X509 printing routines.
13780 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13781 positive and less than q.
13785 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13786 used: it isn't thread safe and the add_lock_callback should handle
13789 *Paul Rose <Paul.Rose@bridge.com>*
13791 * Verify that incoming data obeys the block size in
13792 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13798 *Ulf Möller, Bodo Möller*
13800 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13801 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13802 when fixing the server behaviour for backwards-compatible 'client
13803 hello' messages. (Note that the attack is impractical against
13804 SSL 3.0 and TLS 1.0 anyway because length and version checking
13805 means that the probability of guessing a valid ciphertext is
13806 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13809 Before 0.9.5, the countermeasure (hide the error by generating a
13810 random 'decryption result') did not work properly because
13811 ERR_clear_error() was missing, meaning that SSL_get_error() would
13812 detect the supposedly ignored error.
13814 Both problems are now fixed.
13818 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13819 (previously it was 1024).
13823 * Fix for compatibility mode trust settings: ignore trust settings
13824 unless some valid trust or reject settings are present.
13828 * Fix for blowfish EVP: its a variable length cipher.
13832 * Fix various bugs related to DSA S/MIME verification. Handle missing
13833 parameters in DSA public key structures and return an error in the
13834 DSA routines if parameters are absent.
13838 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13839 in the current directory if neither $RANDFILE nor $HOME was set.
13840 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13841 caused some confusion to Windows users who haven't defined $HOME.
13842 Thus RAND_file_name() is changed again: e_os.h can define a
13843 DEFAULT_HOME, which will be used if $HOME is not set.
13844 For Windows, we use "C:"; on other platforms, we still require
13845 environment variables.
13847 * Move 'if (!initialized) RAND_poll()' into regions protected by
13848 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13849 having multiple threads call RAND_poll() concurrently.
13853 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13854 combination of a flag and a thread ID variable.
13855 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13856 flag), *other* threads can enter ssleay_add_bytes without obeying
13857 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13858 that they do not hold after the first thread unsets add_do_not_lock).
13862 * Change bctest again: '-x' expressions are not available in all
13863 versions of 'test'.
13867 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13869 * Fix a couple of memory leaks in PKCS7_dataDecode()
13871 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13873 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13874 the default extension for executables, if any. Also, make the perl
13875 scripts that use symlink() to test if it really exists and use "cp"
13876 if it doesn't. All this made OpenSSL compilable and installable in
13881 * Fix for asn1_GetSequence() for indefinite length constructed data.
13882 If SEQUENCE is length is indefinite just set c->slen to the total
13883 amount of data available.
13885 *Steve Henson, reported by shige@FreeBSD.org*
13887 *This change does not apply to 0.9.7.*
13889 * Change bctest to avoid here-documents inside command substitution
13890 (workaround for FreeBSD /bin/sh bug).
13891 For compatibility with Ultrix, avoid shell functions (introduced
13892 in the bctest version that searches along $PATH).
13896 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13897 with des_encrypt() defined on some operating systems, like Solaris
13902 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13903 On the Importance of Eliminating Errors in Cryptographic
13904 Computations, J. Cryptology 14 (2001) 2, 101-119,
13905 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13909 * MIPS assembler BIGNUM division bug fix.
13913 * Disabled incorrect Alpha assembler code.
13917 * Fix PKCS#7 decode routines so they correctly update the length
13918 after reading an EOC for the EXPLICIT tag.
13922 *This change does not apply to 0.9.7.*
13924 * Fix bug in PKCS#12 key generation routines. This was triggered
13925 if a 3DES key was generated with a 0 initial byte. Include
13926 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13927 (but broken) behaviour.
13931 * Enhance bctest to search for a working bc along $PATH and print
13934 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13936 * Fix memory leaks in err.c: free err_data string if necessary;
13937 don't write to the wrong index in ERR_set_error_data.
13941 * Implement ssl23_peek (analogous to ssl23_read), which previously
13946 * Replace rdtsc with `_emit` statements for VC++ version 5.
13948 *Jeremy Cooper <jeremy@baymoo.org>*
13950 * Make it possible to reuse SSLv2 sessions.
13954 * In copy_email() check for >= 0 as a return value for
13955 X509_NAME_get_index_by_NID() since 0 is a valid index.
13957 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13959 * Avoid coredump with unsupported or invalid public keys by checking if
13960 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13961 PKCS7_verify() fails with non detached data.
13965 * Don't use getenv in library functions when run as setuid/setgid.
13966 New function OPENSSL_issetugid().
13970 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13971 due to incorrect handling of multi-threading:
13973 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13975 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13977 3. Count how many times MemCheck_off() has been called so that
13978 nested use can be treated correctly. This also avoids
13979 inband-signalling in the previous code (which relied on the
13980 assumption that thread ID 0 is impossible).
13984 * Add "-rand" option also to s_client and s_server.
13988 * Fix CPU detection on Irix 6.x.
13989 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13990 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13992 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13997 *This change does not apply to 0.9.7.*
13999 * Use the cached encoding of an X509_NAME structure rather than
14000 copying it. This is apparently the reason for the libsafe "errors"
14001 but the code is actually correct.
14005 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14006 Bleichenbacher's DSA attack.
14007 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14008 to be set and top=0 forces the highest bit to be set; top=-1 is new
14009 and leaves the highest bit random.
14011 *Ulf Moeller, Bodo Moeller*
14013 * In the `NCONF_...`-based implementations for `CONF_...` queries
14014 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14015 a temporary CONF structure with the data component set to NULL
14016 (which gives segmentation faults in lh_retrieve).
14017 Instead, use NULL for the CONF pointer in CONF_get_string and
14018 CONF_get_number (which may use environment variables) and directly
14019 return NULL from CONF_get_section.
14023 * Fix potential buffer overrun for EBCDIC.
14027 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14028 keyUsage if basicConstraints absent for a CA.
14032 * Make SMIME_write_PKCS7() write mail header values with a format that
14033 is more generally accepted (no spaces before the semicolon), since
14034 some programs can't parse those values properly otherwise. Also make
14035 sure BIO's that break lines after each write do not create invalid
14040 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14041 macros previously used would not encode an empty SEQUENCE OF
14042 and break the signature.
14046 *This change does not apply to 0.9.7.*
14048 * Zero the premaster secret after deriving the master secret in
14053 * Add some EVP_add_digest_alias registrations (as found in
14054 OpenSSL_add_all_digests()) to SSL_library_init()
14055 aka OpenSSL_add_ssl_algorithms(). This provides improved
14056 compatibility with peers using X.509 certificates
14057 with unconventional AlgorithmIdentifier OIDs.
14061 * Fix for Irix with NO_ASM.
14063 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14065 * ./config script fixes.
14067 *Ulf Moeller, Richard Levitte*
14069 * Fix 'openssl passwd -1'.
14073 * Change PKCS12_key_gen_asc() so it can cope with non null
14074 terminated strings whose length is passed in the passlen
14075 parameter, for example from PEM callbacks. This was done
14076 by adding an extra length parameter to asc2uni().
14078 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14080 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14081 call failed, free the DSA structure.
14085 * Fix to uni2asc() to cope with zero length Unicode strings.
14086 These are present in some PKCS#12 files.
14090 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14091 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14092 when writing a 32767 byte record.
14094 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14096 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14097 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14099 (RSA objects have a reference count access to which is protected
14100 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14101 so they are meant to be shared between threads.)
14102 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14103 "Reddie, Steven" <Steven.Reddie@ca.com>*
14105 * Fix a deadlock in CRYPTO_mem_leaks().
14109 * Use better test patterns in bntest.
14113 * rand_win.c fix for Borland C.
14117 * BN_rshift bugfix for n == 0.
14121 * Add a 'bctest' script that checks for some known 'bc' bugs
14122 so that 'make test' does not abort just because 'bc' is broken.
14126 * Store verify_result within SSL_SESSION also for client side to
14127 avoid potential security hole. (Re-used sessions on the client side
14128 always resulted in verify_result==X509_V_OK, not using the original
14129 result of the server certificate verification.)
14133 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14134 SSL3_RT_APPLICATION_DATA, return 0.
14135 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14140 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14141 releases, have been re-implemented by renaming the previous
14142 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14143 and ssl3_read_internal, respectively, and adding 'peek' parameters
14144 to them. The new ssl[23]_{read,peek} functions are calls to
14145 ssl[23]_read_internal with the 'peek' flag set appropriately.
14146 A 'peek' parameter has also been added to ssl3_read_bytes, which
14147 does the actual work for ssl3_read_internal.
14151 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14152 the method-specific "init()" handler. Also clean up ex_data after
14153 calling the method-specific "finish()" handler. Previously, this was
14154 happening the other way round.
14158 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14159 The previous value, 12, was not always sufficient for BN_mod_exp().
14163 * Make sure that shared libraries get the internal name engine with
14164 the full version number and not just 0. This should mark the
14165 shared libraries as not backward compatible. Of course, this should
14166 be changed again when we can guarantee backward binary compatibility.
14170 * Fix typo in get_cert_by_subject() in by_dir.c
14172 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14174 * Rework the system to generate shared libraries:
14176 - Make note of the expected extension for the shared libraries and
14177 if there is a need for symbolic links from for example libcrypto.so.0
14178 to libcrypto.so.0.9.7. There is extended info in Configure for
14181 - Make as few rebuilds of the shared libraries as possible.
14183 - Still avoid linking the OpenSSL programs with the shared libraries.
14185 - When installing, install the shared libraries separately from the
14190 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14192 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14193 and not in SSL_clear because the latter is also used by the
14194 accept/connect functions; previously, the settings made by
14195 SSL_set_read_ahead would be lost during the handshake.
14197 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14199 * Correct util/mkdef.pl to be selective about disabled algorithms.
14200 Previously, it would create entries for disabled algorithms no
14205 * Added several new manual pages for SSL_* function.
14209 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14211 * In ssl23_get_client_hello, generate an error message when faced
14212 with an initial SSL 3.0/TLS record that is too small to contain the
14213 first two bytes of the ClientHello message, i.e. client_version.
14214 (Note that this is a pathologic case that probably has never happened
14215 in real life.) The previous approach was to use the version number
14216 from the record header as a substitute; but our protocol choice
14217 should not depend on that one because it is not authenticated
14218 by the Finished messages.
14222 * More robust randomness gathering functions for Windows.
14224 *Jeffrey Altman <jaltman@columbia.edu>*
14226 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14227 not set then we don't setup the error code for issuer check errors
14228 to avoid possibly overwriting other errors which the callback does
14229 handle. If an application does set the flag then we assume it knows
14230 what it is doing and can handle the new informational codes
14235 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14236 a general "ANY" type, as such it should be able to decode anything
14237 including tagged types. However it didn't check the class so it would
14238 wrongly interpret tagged types in the same way as their universal
14239 counterpart and unknown types were just rejected. Changed so that the
14240 tagged and unknown types are handled in the same way as a SEQUENCE:
14241 that is the encoding is stored intact. There is also a new type
14242 "V_ASN1_OTHER" which is used when the class is not universal, in this
14243 case we have no idea what the actual type is so we just lump them all
14248 * On VMS, stdout may very well lead to a file that is written to
14249 in a record-oriented fashion. That means that every write() will
14250 write a separate record, which will be read separately by the
14251 programs trying to read from it. This can be very confusing.
14253 The solution is to put a BIO filter in the way that will buffer
14254 text until a linefeed is reached, and then write everything a
14255 line at a time, so every record written will be an actual line,
14256 not chunks of lines and not (usually doesn't happen, but I've
14257 seen it once) several lines in one record. BIO_f_linebuffer() is
14260 Currently, it's a VMS-only method, because that's where it has
14261 been tested well enough.
14265 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14266 it can return incorrect results.
14267 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14268 but it was in 0.9.6-beta[12].)
14272 * Disable the check for content being present when verifying detached
14273 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14274 include zero length content when signing messages.
14278 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14279 BIO_ctrl (for BIO pairs).
14283 * Add DSO method for VMS.
14287 * Bug fix: Montgomery multiplication could produce results with the
14292 * Add RPM specification openssl.spec and modify it to build three
14293 packages. The default package contains applications, application
14294 documentation and run-time libraries. The devel package contains
14295 include files, static libraries and function documentation. The
14296 doc package contains the contents of the doc directory. The original
14297 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14301 * Add a large number of documentation files for many SSL routines.
14303 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14305 * Add a configuration entry for Sony News 4.
14307 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14309 * Don't set the two most significant bits to one when generating a
14310 random number < q in the DSA library.
14314 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14315 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14316 the underlying transport is blocking) if a handshake took place.
14317 (The default behaviour is needed by applications such as s_client
14318 and s_server that use select() to determine when to use SSL_read;
14319 but for applications that know in advance when to expect data, it
14320 just makes things more complicated.)
14324 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14329 * Add a few more EBCDIC conditionals that make `req` and `x509`
14330 work better on such systems.
14332 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14334 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14335 Update PKCS12_parse() so it copies the friendlyName and the
14336 keyid to the certificates aux info.
14340 * Fix bug in PKCS7_verify() which caused an infinite loop
14341 if there was more than one signature.
14343 *Sven Uszpelkat <su@celocom.de>*
14345 * Major change in util/mkdef.pl to include extra information
14346 about each symbol, as well as presenting variables as well
14347 as functions. This change means that there's n more need
14348 to rebuild the .num files when some algorithms are excluded.
14352 * Allow the verify time to be set by an application,
14353 rather than always using the current time.
14357 * Phase 2 verify code reorganisation. The certificate
14358 verify code now looks up an issuer certificate by a
14359 number of criteria: subject name, authority key id
14360 and key usage. It also verifies self signed certificates
14361 by the same criteria. The main comparison function is
14362 X509_check_issued() which performs these checks.
14364 Lot of changes were necessary in order to support this
14365 without completely rewriting the lookup code.
14367 Authority and subject key identifier are now cached.
14369 The LHASH 'certs' is X509_STORE has now been replaced
14370 by a STACK_OF(X509_OBJECT). This is mainly because an
14371 LHASH can't store or retrieve multiple objects with
14372 the same hash value.
14374 As a result various functions (which were all internal
14375 use only) have changed to handle the new X509_STORE
14376 structure. This will break anything that messed round
14377 with X509_STORE internally.
14379 The functions X509_STORE_add_cert() now checks for an
14380 exact match, rather than just subject name.
14382 The X509_STORE API doesn't directly support the retrieval
14383 of multiple certificates matching a given criteria, however
14384 this can be worked round by performing a lookup first
14385 (which will fill the cache with candidate certificates)
14386 and then examining the cache for matches. This is probably
14387 the best we can do without throwing out X509_LOOKUP
14388 entirely (maybe later...).
14390 The X509_VERIFY_CTX structure has been enhanced considerably.
14392 All certificate lookup operations now go via a get_issuer()
14393 callback. Although this currently uses an X509_STORE it
14394 can be replaced by custom lookups. This is a simple way
14395 to bypass the X509_STORE hackery necessary to make this
14396 work and makes it possible to use more efficient techniques
14397 in future. A very simple version which uses a simple
14398 STACK for its trusted certificate store is also provided
14399 using X509_STORE_CTX_trusted_stack().
14401 The verify_cb() and verify() callbacks now have equivalents
14402 in the X509_STORE_CTX structure.
14404 X509_STORE_CTX also has a 'flags' field which can be used
14405 to customise the verify behaviour.
14409 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14410 excludes S/MIME capabilities.
14414 * When a certificate request is read in keep a copy of the
14415 original encoding of the signed data and use it when outputting
14416 again. Signatures then use the original encoding rather than
14417 a decoded, encoded version which may cause problems if the
14418 request is improperly encoded.
14422 * For consistency with other BIO_puts implementations, call
14423 buffer_write(b, ...) directly in buffer_puts instead of calling
14426 In BIO_puts, increment b->num_write as in BIO_write.
14428 *Peter.Sylvester@EdelWeb.fr*
14430 * Fix BN_mul_word for the case where the word is 0. (We have to use
14431 BN_zero, we may not return a BIGNUM with an array consisting of
14432 words set to zero.)
14436 * Avoid calling abort() from within the library when problems are
14437 detected, except if preprocessor symbols have been defined
14438 (such as REF_CHECK, BN_DEBUG etc.).
14442 * New openssl application 'rsautl'. This utility can be
14443 used for low level RSA operations. DER public key
14444 BIO/fp routines also added.
14448 * New Configure entry and patches for compiling on QNX 4.
14450 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14452 * A demo state-machine implementation was sponsored by
14453 Nuron (<http://www.nuron.com/>) and is now available in
14454 demos/state_machine.
14458 * New options added to the 'dgst' utility for signature
14459 generation and verification.
14463 * Unrecognized PKCS#7 content types are now handled via a
14464 catch all ASN1_TYPE structure. This allows unsupported
14465 types to be stored as a "blob" and an application can
14466 encode and decode it manually.
14470 * Fix various signed/unsigned issues to make a_strex.c
14471 compile under VC++.
14473 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14475 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14476 length if passed a buffer. ASN1_INTEGER_to_BN failed
14477 if passed a NULL BN and its argument was negative.
14479 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14481 * Modification to PKCS#7 encoding routines to output definite
14482 length encoding. Since currently the whole structures are in
14483 memory there's not real point in using indefinite length
14484 constructed encoding. However if OpenSSL is compiled with
14485 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14489 * Added BIO_vprintf() and BIO_vsnprintf().
14493 * Added more prefixes to parse for in the strings written
14494 through a logging bio, to cover all the levels that are available
14495 through syslog. The prefixes are now:
14497 PANIC, EMERG, EMR => LOG_EMERG
14498 ALERT, ALR => LOG_ALERT
14499 CRIT, CRI => LOG_CRIT
14500 ERROR, ERR => LOG_ERR
14501 WARNING, WARN, WAR => LOG_WARNING
14502 NOTICE, NOTE, NOT => LOG_NOTICE
14503 INFO, INF => LOG_INFO
14504 DEBUG, DBG => LOG_DEBUG
14506 and as before, if none of those prefixes are present at the
14507 beginning of the string, LOG_ERR is chosen.
14509 On Win32, the `LOG_*` levels are mapped according to this:
14511 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14512 LOG_WARNING => EVENTLOG_WARNING_TYPE
14513 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14517 * Made it possible to reconfigure with just the configuration
14518 argument "reconf" or "reconfigure". The command line arguments
14519 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14520 and are retrieved from there when reconfiguring.
14526 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14528 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14532 * The obj_dat.pl script was messing up the sorting of object
14533 names. The reason was that it compared the quoted version
14534 of strings as a result "OCSP" > "OCSP Signing" because
14535 " > SPACE. Changed script to store unquoted versions of
14536 names and add quotes on output. It was also omitting some
14537 names from the lookup table if they were given a default
14538 value (that is if SN is missing it is given the same
14539 value as LN and vice versa), these are now added on the
14540 grounds that if an object has a name we should be able to
14541 look it up. Finally added warning output when duplicate
14542 short or long names are found.
14546 * Changes needed for Tandem NSK.
14548 *Scott Uroff <scott@xypro.com>*
14550 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14551 RSA_padding_check_SSLv23(), special padding was never detected
14552 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14553 version rollback attacks was not effective.
14555 In s23_clnt.c, don't use special rollback-attack detection padding
14556 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14557 client; similarly, in s23_srvr.c, don't do the rollback check if
14558 SSL 2.0 is the only protocol enabled in the server.
14562 * Make it possible to get hexdumps of unprintable data with 'openssl
14563 asn1parse'. By implication, the functions ASN1_parse_dump() and
14564 BIO_dump_indent() are added.
14568 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14569 these print out strings and name structures based on various
14570 flags including RFC2253 support and proper handling of
14571 multibyte characters. Added options to the 'x509' utility
14572 to allow the various flags to be set.
14576 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14577 Also change the functions X509_cmp_current_time() and
14578 X509_gmtime_adj() work with an ASN1_TIME structure,
14579 this will enable certificates using GeneralizedTime in validity
14580 dates to be checked.
14584 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14585 negative public key encodings) on by default,
14586 NO_NEG_PUBKEY_BUG can be set to disable it.
14590 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14591 content octets. An i2c_ASN1_OBJECT is unnecessary because
14592 the encoding can be trivially obtained from the structure.
14596 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14597 not read locks (`CRYPTO_r_[un]lock`).
14601 * A first attempt at creating official support for shared
14602 libraries through configuration. I've kept it so the
14603 default is static libraries only, and the OpenSSL programs
14604 are always statically linked for now, but there are
14605 preparations for dynamic linking in place.
14606 This has been tested on Linux and Tru64.
14610 * Randomness polling function for Win9x, as described in:
14611 Peter Gutmann, Software Generation of Practically Strong
14616 * Fix so PRNG is seeded in req if using an already existing
14621 * New options to smime application. -inform and -outform
14622 allow alternative formats for the S/MIME message including
14623 PEM and DER. The -content option allows the content to be
14624 specified separately. This should allow things like Netscape
14625 form signing output easier to verify.
14629 * Fix the ASN1 encoding of tags using the 'long form'.
14633 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14634 STRING types. These convert content octets to and from the
14635 underlying type. The actual tag and length octets are
14636 already assumed to have been read in and checked. These
14637 are needed because all other string types have virtually
14638 identical handling apart from the tag. By having versions
14639 of the ASN1 functions that just operate on content octets
14640 IMPLICIT tagging can be handled properly. It also allows
14641 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14642 and ASN1_INTEGER are identical apart from the tag.
14646 * Change the handling of OID objects as follows:
14648 - New object identifiers are inserted in objects.txt, following
14649 the syntax given in objects.README.
14650 - objects.pl is used to process obj_mac.num and create a new
14652 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14655 This is currently kind of a hack, and the perl code in objects.pl
14656 isn't very elegant, but it works as I intended. The simplest way
14657 to check that it worked correctly is to look in obj_dat.h and
14658 check the array nid_objs and make sure the objects haven't moved
14659 around (this is important!). Additions are OK, as well as
14660 consistent name changes.
14664 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14668 * Addition of the command line parameter '-rand file' to 'openssl req'.
14669 The given file adds to whatever has already been seeded into the
14670 random pool through the RANDFILE configuration file option or
14671 environment variable, or the default random state file.
14675 * mkstack.pl now sorts each macro group into lexical order.
14676 Previously the output order depended on the order the files
14677 appeared in the directory, resulting in needless rewriting
14682 * Patches to make OpenSSL compile under Win32 again. Mostly
14683 work arounds for the VC++ problem that it treats func() as
14684 func(void). Also stripped out the parts of mkdef.pl that
14685 added extra typesafe functions: these no longer exist.
14689 * Reorganisation of the stack code. The macros are now all
14690 collected in safestack.h . Each macro is defined in terms of
14691 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14692 DEBUG_SAFESTACK is now handled in terms of function casts,
14693 this has the advantage of retaining type safety without the
14694 use of additional functions. If DEBUG_SAFESTACK is not defined
14695 then the non typesafe macros are used instead. Also modified the
14696 mkstack.pl script to handle the new form. Needs testing to see
14697 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14698 the default if no major problems. Similar behaviour for ASN1_SET_OF
14699 and PKCS12_STACK_OF.
14703 * When some versions of IIS use the 'NET' form of private key the
14704 key derivation algorithm is different. Normally MD5(password) is
14705 used as a 128 bit RC4 key. In the modified case
14706 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14707 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14708 as the old Netscape_RSA functions except they have an additional
14709 'sgckey' parameter which uses the modified algorithm. Also added
14710 an -sgckey command line option to the rsa utility. Thanks to
14711 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14712 algorithm to openssl-dev.
14716 * The evp_local.h macros were using 'c.##kname' which resulted in
14717 invalid expansion on some systems (SCO 5.0.5 for example).
14718 Corrected to 'c.kname'.
14720 *Phillip Porch <root@theporch.com>*
14722 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14723 a STACK of email addresses from a certificate or request, these look
14724 in the subject name and the subject alternative name extensions and
14725 omit any duplicate addresses.
14729 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14730 This makes DSA verification about 2 % faster.
14734 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14735 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14736 plus overhead for 1024 bit moduli).
14737 This makes exponentiations about 0.5 % faster for 1024 bit
14738 exponents (as measured by "openssl speed rsa2048").
14742 * Rename memory handling macros to avoid conflicts with other
14744 Malloc => OPENSSL_malloc
14745 Malloc_locked => OPENSSL_malloc_locked
14746 Realloc => OPENSSL_realloc
14747 Free => OPENSSL_free
14751 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14752 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14756 * CygWin32 support.
14758 *John Jarvie <jjarvie@newsguy.com>*
14760 * The type-safe stack code has been rejigged. It is now only compiled
14761 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14762 by default all type-specific stack functions are "#define"d back to
14763 standard stack functions. This results in more streamlined output
14764 but retains the type-safety checking possibilities of the original
14769 * The STACK code has been cleaned up, and certain type declarations
14770 that didn't make a lot of sense have been brought in line. This has
14771 also involved a cleanup of sorts in safestack.h to more correctly
14772 map type-safe stack functions onto their plain stack counterparts.
14773 This work has also resulted in a variety of "const"ifications of
14774 lots of the code, especially `_cmp` operations which should normally
14775 be prototyped with "const" parameters anyway.
14779 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14780 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14781 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14782 where all of 'md' is used each time the PRNG is used, but 'state'
14783 is used only indexed by a cyclic counter. As entropy may not be
14784 well distributed from the beginning, 'md' is important as a
14785 chaining variable. However, the output function chains only half
14786 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14787 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14788 in all of 'state' being rewritten, with the new values depending
14789 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14793 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14794 the handshake is continued after ssl_verify_cert_chain();
14795 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14796 can lead to 'unexplainable' connection aborts later.
14798 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14800 * Major EVP API cipher revision.
14801 Add hooks for extra EVP features. This allows various cipher
14802 parameters to be set in the EVP interface. Support added for variable
14803 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14804 setting of RC2 and RC5 parameters.
14806 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14809 Remove lots of duplicated code from the EVP library. For example *every*
14810 cipher init() function handles the 'iv' in the same way according to the
14811 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14812 for CFB and OFB modes they zero ctx->num.
14814 New functionality allows removal of S/MIME code RC2 hack.
14816 Most of the routines have the same form and so can be declared in terms
14819 By shifting this to the top level EVP_CipherInit() it can be removed from
14820 all individual ciphers. If the cipher wants to handle IVs or keys
14821 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14824 Change lots of functions like EVP_EncryptUpdate() to now return a
14825 value: although software versions of the algorithms cannot fail
14826 any installed hardware versions can.
14830 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14831 this option is set, tolerate broken clients that send the negotiated
14832 protocol version number instead of the requested protocol version
14837 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14838 i.e. non-zero for export ciphersuites, zero otherwise.
14839 Previous versions had this flag inverted, inconsistent with
14840 rsa_tmp_cb (..._TMP_RSA_CB).
14842 *Bodo Moeller; problem reported by Amit Chopra*
14844 * Add missing DSA library text string. Work around for some IIS
14845 key files with invalid SEQUENCE encoding.
14849 * Add a document (doc/standards.txt) that list all kinds of standards
14850 and so on that are implemented in OpenSSL.
14854 * Enhance c_rehash script. Old version would mishandle certificates
14855 with the same subject name hash and wouldn't handle CRLs at all.
14856 Added -fingerprint option to crl utility, to support new c_rehash
14861 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14865 * Fix for SSL server purpose checking. Server checking was
14866 rejecting certificates which had extended key usage present
14867 but no ssl client purpose.
14869 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14871 * Make PKCS#12 code work with no password. The PKCS#12 spec
14872 is a little unclear about how a blank password is handled.
14873 Since the password in encoded as a BMPString with terminating
14874 double NULL a zero length password would end up as just the
14875 double NULL. However no password at all is different and is
14876 handled differently in the PKCS#12 key generation code. NS
14877 treats a blank password as zero length. MSIE treats it as no
14878 password on export: but it will try both on import. We now do
14879 the same: PKCS12_parse() tries zero length and no password if
14880 the password is set to "" or NULL (NULL is now a valid password:
14881 it wasn't before) as does the pkcs12 application.
14885 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14886 perror when PEM_read_bio_X509_REQ fails, the error message must
14887 be obtained from the error queue.
14891 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14892 it in ERR_remove_state if appropriate, and change ERR_get_state
14893 accordingly to avoid race conditions (this is necessary because
14894 thread_hash is no longer constant once set).
14898 * Bugfix for linux-elf makefile.one.
14902 * RSA_get_default_method() will now cause a default
14903 RSA_METHOD to be chosen if one doesn't exist already.
14904 Previously this was only set during a call to RSA_new()
14905 or RSA_new_method(NULL) meaning it was possible for
14906 RSA_get_default_method() to return NULL.
14910 * Added native name translation to the existing DSO code
14911 that will convert (if the flag to do so is set) filenames
14912 that are sufficiently small and have no path information
14913 into a canonical native form. Eg. "blah" converted to
14914 "libblah.so" or "blah.dll" etc.
14918 * New function ERR_error_string_n(e, buf, len) which is like
14919 ERR_error_string(e, buf), but writes at most 'len' bytes
14920 including the 0 terminator. For ERR_error_string_n, 'buf'
14923 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14925 * CONF library reworked to become more general. A new CONF
14926 configuration file reader "class" is implemented as well as a
14927 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14928 old `CONF_*` functions are still there, but are reimplemented to
14929 work in terms of the new functions. Also, a set of functions
14930 to handle the internal storage of the configuration data is
14931 provided to make it easier to write new configuration file
14932 reader "classes" (I can definitely see something reading a
14933 configuration file in XML format, for example), called `_CONF_*`,
14934 or "the configuration storage API"...
14936 The new configuration file reading functions are:
14938 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14939 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14941 NCONF_default, NCONF_WIN32
14943 NCONF_dump_fp, NCONF_dump_bio
14945 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14946 NCONF_new creates a new CONF object. This works in the same way
14947 as other interfaces in OpenSSL, like the BIO interface.
14948 `NCONF_dump_*` dump the internal storage of the configuration file,
14949 which is useful for debugging. All other functions take the same
14950 arguments as the old `CONF_*` functions with the exception of the
14951 first that must be a `CONF *` instead of a `LHASH *`.
14953 To make it easier to use the new classes with the old `CONF_*` functions,
14954 the function CONF_set_default_method is provided.
14958 * Add '-tls1' option to 'openssl ciphers', which was already
14959 mentioned in the documentation but had not been implemented.
14960 (This option is not yet really useful because even the additional
14961 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14965 * Initial DSO code added into libcrypto for letting OpenSSL (and
14966 OpenSSL-based applications) load shared libraries and bind to
14967 them in a portable way.
14969 *Geoff Thorpe, with contributions from Richard Levitte*
14971 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
14973 * Make sure _lrotl and _lrotr are only used with MSVC.
14975 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14976 (the default implementation of RAND_status).
14978 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14979 to '-clrext' (= clear extensions), as intended and documented.
14980 *Bodo Moeller; inconsistency pointed out by Michael Attili
14981 <attili@amaxo.com>*
14983 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14984 was larger than the MD block size.
14986 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14988 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14989 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14990 using the passed key: if the passed key was a private key the result
14991 of X509_print(), for example, would be to print out all the private key
14996 * des_quad_cksum() byte order bug fix.
14997 *Ulf Möller, using the problem description in krb4-0.9.7, where
14998 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15000 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15003 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15005 * For easily testing in shell scripts whether some command
15006 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15007 returns with exit code 0 iff no command of the given name is available.
15008 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15009 the output goes to stdout and nothing is printed to stderr.
15010 Additional arguments are always ignored.
15012 Since for each cipher there is a command of the same name,
15013 the 'no-cipher' compilation switches can be tested this way.
15015 ('openssl no-XXX' is not able to detect pseudo-commands such
15016 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15020 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15024 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15025 is set; it will be thrown away anyway because each handshake creates
15027 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15028 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15029 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15030 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15034 * New s_client option -ign_eof: EOF at stdin is ignored, and
15035 'Q' and 'R' lose their special meanings (quit/renegotiate).
15036 This is part of what -quiet does; unlike -quiet, -ign_eof
15037 does not suppress any output.
15041 * Add compatibility options to the purpose and trust code. The
15042 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15043 accepts a certificate or CA, this was the previous behaviour,
15044 with all the associated security issues.
15046 X509_TRUST_COMPAT is the old trust behaviour: only and
15047 automatically trust self signed roots in certificate store. A
15048 new trust setting X509_TRUST_DEFAULT is used to specify that
15049 a purpose has no associated trust setting and it should instead
15050 use the value in the default purpose.
15054 * Fix the PKCS#8 DSA private key code so it decodes keys again
15055 and fix a memory leak.
15059 * In util/mkerr.pl (which implements 'make errors'), preserve
15060 reason strings from the previous version of the .c file, as
15061 the default to have only downcase letters (and digits) in
15062 automatically generated reasons codes is not always appropriate.
15066 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15067 using strerror. Previously, ERR_reason_error_string() returned
15068 library names as reason strings for SYSerr; but SYSerr is a special
15069 case where small numbers are errno values, not library numbers.
15073 * Add '-dsaparam' option to 'openssl dhparam' application. This
15074 converts DSA parameters into DH parameters. (When creating parameters,
15075 DSA_generate_parameters is used.)
15079 * Include 'length' (recommended exponent length) in C code generated
15080 by 'openssl dhparam -C'.
15084 * The second argument to set_label in perlasm was already being used
15085 so couldn't be used as a "file scope" flag. Moved to third argument
15090 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15091 instead of RAND_bytes for encryption IVs and salts.
15095 * Include RAND_status() into RAND_METHOD instead of implementing
15096 it only for md_rand.c Otherwise replacing the PRNG by calling
15097 RAND_set_rand_method would be impossible.
15101 * Don't let DSA_generate_key() enter an infinite loop if the random
15102 number generation fails.
15106 * New 'rand' application for creating pseudo-random output.
15110 * Added configuration support for Linux/IA64
15112 *Rolf Haberrecker <rolf@suse.de>*
15114 * Assembler module support for Mingw32.
15118 * Shared library support for HPUX (in shlib/).
15120 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15122 * Shared library support for Solaris gcc.
15124 *Lutz Behnke <behnke@trustcenter.de>*
15126 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15128 * PKCS7_encrypt() was adding text MIME headers twice because they
15129 were added manually and by SMIME_crlf_copy().
15133 * In bntest.c don't call BN_rand with zero bits argument.
15135 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15137 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15138 case was implemented. This caused BN_div_recp() to fail occasionally.
15142 * Add an optional second argument to the set_label() in the perl
15143 assembly language builder. If this argument exists and is set
15144 to 1 it signals that the assembler should use a symbol whose
15145 scope is the entire file, not just the current function. This
15146 is needed with MASM which uses the format label:: for this scope.
15148 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15150 * Change the ASN1 types so they are typedefs by default. Before
15151 almost all types were #define'd to ASN1_STRING which was causing
15152 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15157 * Change names of new functions to the new get1/get0 naming
15158 convention: After 'get1', the caller owns a reference count
15159 and has to call `..._free`; 'get0' returns a pointer to some
15160 data structure without incrementing reference counters.
15161 (Some of the existing 'get' functions increment a reference
15162 counter, some don't.)
15163 Similarly, 'set1' and 'add1' functions increase reference
15164 counters or duplicate objects.
15168 * Allow for the possibility of temp RSA key generation failure:
15169 the code used to assume it always worked and crashed on failure.
15173 * Fix potential buffer overrun problem in BIO_printf().
15174 *Ulf Möller, using public domain code by Patrick Powell; problem
15175 pointed out by David Sacerdote <das33@cornell.edu>*
15177 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15178 RAND_egd() and RAND_status(). In the command line application,
15179 the EGD socket can be specified like a seed file using RANDFILE
15184 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15185 Some CAs (e.g. Verisign) distribute certificates in this form.
15189 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15190 list to exclude them. This means that no special compilation option
15191 is needed to use anonymous DH: it just needs to be included in the
15196 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15197 EVP_MD_type. The old functionality is available in a new macro called
15198 EVP_MD_md(). Change code that uses it and update docs.
15202 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15203 where the `void *` argument is replaced by a function pointer argument.
15204 Previously `void *` was abused to point to functions, which works on
15205 many platforms, but is not correct. As these functions are usually
15206 called by macros defined in OpenSSL header files, most source code
15207 should work without changes.
15211 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15212 sections with information on -D... compiler switches used for
15213 compiling the library so that applications can see them. To enable
15214 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15215 must be defined. E.g.,
15216 #define OPENSSL_ALGORITHM_DEFINES
15217 #include <openssl/opensslconf.h>
15218 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15220 *Richard Levitte, Ulf and Bodo Möller*
15222 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15227 * Change the 'other' type in certificate aux info to a STACK_OF
15228 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15229 the required ASN1 format: arbitrary types determined by an OID.
15233 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15234 argument to 'req'. This is not because the function is newer or
15235 better than others it just uses the work 'NEW' in the certificate
15236 request header lines. Some software needs this.
15240 * Reorganise password command line arguments: now passwords can be
15241 obtained from various sources. Delete the PEM_cb function and make
15242 it the default behaviour: i.e. if the callback is NULL and the
15243 usrdata argument is not NULL interpret it as a null terminated pass
15244 phrase. If usrdata and the callback are NULL then the pass phrase
15245 is prompted for as usual.
15249 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15250 the support is automatically enabled. The resulting binaries will
15251 autodetect the card and use it if present.
15253 *Ben Laurie and Compaq Inc.*
15255 * Work around for Netscape hang bug. This sends certificate request
15256 and server done in one record. Since this is perfectly legal in the
15257 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15258 the bugs/SSLv3 entry for more info.
15262 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15266 * Add -rand argument to smime and pkcs12 applications and read/write
15271 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15275 * Add command line password options to the remaining applications.
15279 * Bug fix for BN_div_recp() for numerators with an even number of
15284 * More tests in bntest.c, and changed test_bn output.
15288 * ./config recognizes MacOS X now.
15292 * Bug fix for BN_div() when the first words of num and divisor are
15293 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15297 * Add support for various broken PKCS#8 formats, and command line
15298 options to produce them.
15302 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15303 get temporary BIGNUMs from a BN_CTX.
15307 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15312 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15313 include a #define from the old name to the new. The original intent
15314 was that statically linked binaries could for example just call
15315 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15316 link with digests. This never worked because SSLeay_add_all_digests()
15317 and SSLeay_add_all_ciphers() were in the same source file so calling
15318 one would link with the other. They are now in separate source files.
15322 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15326 * Use a less unusual form of the Miller-Rabin primality test (it used
15327 a binary algorithm for exponentiation integrated into the Miller-Rabin
15328 loop, our standard modexp algorithms are faster).
15332 * Support for the EBCDIC character set completed.
15334 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15336 * Source code cleanups: use const where appropriate, eliminate casts,
15337 use `void *` instead of `char *` in lhash.
15341 * Bugfix: ssl3_send_server_key_exchange was not restartable
15342 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15343 this the server could overwrite ephemeral keys that the client
15348 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15349 using 50 iterations of the Rabin-Miller test.
15351 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15352 iterations of the Rabin-Miller test as required by the appendix
15353 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15354 As BN_is_prime_fasttest includes trial division, DSA parameter
15355 generation becomes much faster.
15357 This implies a change for the callback functions in DSA_is_prime
15358 and DSA_generate_parameters: The callback function is called once
15359 for each positive witness in the Rabin-Miller test, not just
15360 occasionally in the inner loop; and the parameters to the
15361 callback function now provide an iteration count for the outer
15362 loop rather than for the current invocation of the inner loop.
15363 DSA_generate_parameters additionally can call the callback
15364 function with an 'iteration count' of -1, meaning that a
15365 candidate has passed the trial division test (when q is generated
15366 from an application-provided seed, trial division is skipped).
15370 * New function BN_is_prime_fasttest that optionally does trial
15371 division before starting the Rabin-Miller test and has
15372 an additional BN_CTX * argument (whereas BN_is_prime always
15373 has to allocate at least one BN_CTX).
15374 'callback(1, -1, cb_arg)' is called when a number has passed the
15375 trial division stage.
15379 * Fix for bug in CRL encoding. The validity dates weren't being handled
15384 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15388 * New function BN_pseudo_rand().
15392 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15393 bignum version of BN_from_montgomery() with the working code from
15394 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15399 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15400 made it impossible to use the same SSL_SESSION data structure in
15401 SSL2 clients in multiple threads.
15405 * The return value of RAND_load_file() no longer counts bytes obtained
15406 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15407 to seed the PRNG (previously an explicit byte count was required).
15409 *Ulf Möller, Bodo Möller*
15411 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15412 used `char *` instead of `void *` and had casts all over the place.
15416 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15420 * Retain source code compatibility for BN_prime_checks macro:
15421 BN_is_prime(..., BN_prime_checks, ...) now uses
15422 BN_prime_checks_for_size to determine the appropriate number of
15423 Rabin-Miller iterations.
15427 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15428 DH_CHECK_P_NOT_SAFE_PRIME.
15429 (Check if this is true? OpenPGP calls them "strong".)
15433 * Merge the functionality of "dh" and "gendh" programs into a new program
15434 "dhparam". The old programs are retained for now but will handle DH keys
15435 (instead of parameters) in future.
15439 * Make the ciphers, s_server and s_client programs check the return values
15440 when a new cipher list is set.
15444 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15445 ciphers. Before when the 56bit ciphers were enabled the sorting was
15448 The syntax for the cipher sorting has been extended to support sorting by
15449 cipher-strength (using the strength_bits hard coded in the tables).
15450 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15452 Fix a bug in the cipher-command parser: when supplying a cipher command
15453 string with an "undefined" symbol (neither command nor alphanumeric
15454 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15455 an error is flagged.
15457 Due to the strength-sorting extension, the code of the
15458 ssl_create_cipher_list() function was completely rearranged. I hope that
15459 the readability was also increased :-)
15461 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15463 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15464 for the first serial number and places 2 in the serial number file. This
15465 avoids problems when the root CA is created with serial number zero and
15466 the first user certificate has the same issuer name and serial number
15471 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15472 the new code. Add documentation for this stuff.
15476 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15477 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15478 structures and behave in an analogous way to the X509v3 functions:
15479 they shouldn't be called directly but wrapper functions should be used
15482 So we also now have some wrapper functions that call the X509at functions
15483 when passed certificate requests. (TO DO: similar things can be done with
15484 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15485 things. Some of these need some d2i or i2d and print functionality
15486 because they handle more complex structures.)
15490 * Add missing #ifndefs that caused missing symbols when building libssl
15491 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15492 NO_RSA in `ssl/s2*.c`.
15494 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15496 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15497 has a return value which indicates the quality of the random data
15498 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15499 error queue. New function RAND_pseudo_bytes() generates output that is
15500 guaranteed to be unique but not unpredictable. RAND_add is like
15501 RAND_seed, but takes an extra argument for an entropy estimate
15502 (RAND_seed always assumes full entropy).
15506 * Do more iterations of Rabin-Miller probable prime test (specifically,
15507 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15508 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15509 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15510 false-positive rate of at most 2^-80 for random input.
15514 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15518 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15519 in the 0.9.5 release), this returns the chain
15520 from an X509_CTX structure with a dup of the stack and all
15521 the X509 reference counts upped: so the stack will exist
15522 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15525 Also make SSL_SESSION_print() print out the verify return
15530 * Add manpage for the pkcs12 command. Also change the default
15531 behaviour so MAC iteration counts are used unless the new
15532 -nomaciter option is used. This improves file security and
15533 only older versions of MSIE (4.0 for example) need it.
15537 * Honor the no-xxx Configure options when creating .DEF files.
15541 * Add PKCS#10 attributes to field table: challengePassword,
15542 unstructuredName and unstructuredAddress. These are taken from
15543 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15544 international characters are used.
15546 More changes to X509_ATTRIBUTE code: allow the setting of types
15547 based on strings. Remove the 'loc' parameter when adding
15548 attributes because these will be a SET OF encoding which is sorted
15553 * Initial changes to the 'req' utility to allow request generation
15554 automation. This will allow an application to just generate a template
15555 file containing all the field values and have req construct the
15558 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15559 used all over the place including certificate requests and PKCS#7
15560 structures. They are currently handled manually where necessary with
15561 some primitive wrappers for PKCS#7. The new functions behave in a
15562 manner analogous to the X509 extension functions: they allow
15563 attributes to be looked up by NID and added.
15565 Later something similar to the X509V3 code would be desirable to
15566 automatically handle the encoding, decoding and printing of the
15567 more complex types. The string types like challengePassword can
15568 be handled by the string table functions.
15570 Also modified the multi byte string table handling. Now there is
15571 a 'global mask' which masks out certain types. The table itself
15572 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15573 is useful when for example there is only one permissible type
15574 (as in countryName) and using the mask might result in no valid
15579 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15580 SSL_get_peer_finished to allow applications to obtain the latest
15581 Finished messages sent to the peer or expected from the peer,
15582 respectively. (SSL_get_peer_finished is usually the Finished message
15583 actually received from the peer, otherwise the protocol will be aborted.)
15585 As the Finished message are message digests of the complete handshake
15586 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15587 be used for external authentication procedures when the authentication
15588 provided by SSL/TLS is not desired or is not enough.
15592 * Enhanced support for Alpha Linux is added. Now ./config checks if
15593 the host supports BWX extension and if Compaq C is present on the
15594 $PATH. Just exploiting of the BWX extension results in 20-30%
15595 performance kick for some algorithms, e.g. DES and RC4 to mention
15596 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15601 * Add support for MS "fast SGC". This is arguably a violation of the
15602 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15603 weak crypto and after checking the certificate is SGC a second one
15604 with strong crypto. MS SGC stops the first handshake after receiving
15605 the server certificate message and sends a second client hello. Since
15606 a server will typically do all the time consuming operations before
15607 expecting any further messages from the client (server key exchange
15608 is the most expensive) there is little difference between the two.
15610 To get OpenSSL to support MS SGC we have to permit a second client
15611 hello message after we have sent server done. In addition we have to
15612 reset the MAC if we do get this second client hello.
15616 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15617 if a DER encoded private key is RSA or DSA traditional format. Changed
15618 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15619 format DER encoded private key. Newer code should use PKCS#8 format which
15620 has the key type encoded in the ASN1 structure. Added DER private key
15621 support to pkcs8 application.
15625 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15626 ciphersuites has been selected (as required by the SSL 3/TLS 1
15627 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15628 is set, we interpret this as a request to violate the specification
15629 (the worst that can happen is a handshake failure, and 'correct'
15630 behaviour would result in a handshake failure anyway).
15634 * In SSL_CTX_add_session, take into account that there might be multiple
15635 SSL_SESSION structures with the same session ID (e.g. when two threads
15636 concurrently obtain them from an external cache).
15637 The internal cache can handle only one SSL_SESSION with a given ID,
15638 so if there's a conflict, we now throw out the old one to achieve
15643 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15644 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15645 some routines that use cipher OIDs: some ciphers do not have OIDs
15646 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15651 * Simplify the trust setting structure and code. Now we just have
15652 two sequences of OIDs for trusted and rejected settings. These will
15653 typically have values the same as the extended key usage extension
15654 and any application specific purposes.
15656 The trust checking code now has a default behaviour: it will just
15657 check for an object with the same NID as the passed id. Functions can
15658 be provided to override either the default behaviour or the behaviour
15659 for a given id. SSL client, server and email already have functions
15660 in place for compatibility: they check the NID and also return "trusted"
15661 if the certificate is self signed.
15665 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15666 traditional format into an EVP_PKEY structure.
15670 * Add a password callback function PEM_cb() which either prompts for
15671 a password if usr_data is NULL or otherwise assumes it is a null
15672 terminated password. Allow passwords to be passed on command line
15673 environment or config files in a few more utilities.
15677 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15678 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15679 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15680 Update documentation.
15684 * Support for ASN1 "NULL" type. This could be handled before by using
15685 ASN1_TYPE but there wasn't any function that would try to read a NULL
15686 and produce an error if it couldn't. For compatibility we also have
15687 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15688 don't allocate anything because they don't need to.
15692 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15695 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15697 * Rebuild of the memory allocation routines used by OpenSSL code and
15698 possibly others as well. The purpose is to make an interface that
15699 provide hooks so anyone can build a separate set of allocation and
15700 deallocation routines to be used by OpenSSL, for example memory
15701 pool implementations, or something else, which was previously hard
15702 since Malloc(), Realloc() and Free() were defined as macros having
15703 the values malloc, realloc and free, respectively (except for Win32
15704 compilations). The same is provided for memory debugging code.
15705 OpenSSL already comes with functionality to find memory leaks, but
15706 this gives people a chance to debug other memory problems.
15708 With these changes, a new set of functions and macros have appeared:
15710 CRYPTO_set_mem_debug_functions() [F]
15711 CRYPTO_get_mem_debug_functions() [F]
15712 CRYPTO_dbg_set_options() [F]
15713 CRYPTO_dbg_get_options() [F]
15714 CRYPTO_malloc_debug_init() [M]
15716 The memory debug functions are NULL by default, unless the library
15717 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15718 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15719 gives the standard debugging functions that come with OpenSSL) or
15720 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15721 provided by the library user) must be used. When the standard
15722 debugging functions are used, CRYPTO_dbg_set_options can be used to
15723 request additional information:
15724 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15725 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15727 Also, things like CRYPTO_set_mem_functions will always give the
15728 expected result (the new set of functions is used for allocation
15729 and deallocation) at all times, regardless of platform and compiler
15732 To finish it up, some functions that were never use in any other
15733 way than through macros have a new API and new semantic:
15735 CRYPTO_dbg_malloc()
15736 CRYPTO_dbg_realloc()
15739 All macros of value have retained their old syntax.
15741 *Richard Levitte and Bodo Moeller*
15743 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15744 ordering of SMIMECapabilities wasn't in "strength order" and there
15745 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15750 * Some ASN1 types with illegal zero length encoding (INTEGER,
15751 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15753 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15755 * Merge in my S/MIME library for OpenSSL. This provides a simple
15756 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15757 functionality to handle multipart/signed properly) and a utility
15758 called 'smime' to call all this stuff. This is based on code I
15759 originally wrote for Celo who have kindly allowed it to be
15760 included in OpenSSL.
15764 * Add variants des_set_key_checked and des_set_key_unchecked of
15765 des_set_key (aka des_key_sched). Global variable des_check_key
15766 decides which of these is called by des_set_key; this way
15767 des_check_key behaves as it always did, but applications and
15768 the library itself, which was buggy for des_check_key == 1,
15769 have a cleaner way to pick the version they need.
15773 * New function PKCS12_newpass() which changes the password of a
15778 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15779 dynamic mix. In both cases the ids can be used as an index into the
15780 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15781 functions so they accept a list of the field values and the
15782 application doesn't need to directly manipulate the X509_TRUST
15787 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15792 * Modify the way the V3 extension code looks up extensions. This now
15793 works in a similar way to the object code: we have some "standard"
15794 extensions in a static table which is searched with OBJ_bsearch()
15795 and the application can add dynamic ones if needed. The file
15796 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15797 updated whenever a new extension is added to the core code and kept
15798 in ext_nid order. There is a simple program 'tabtest.c' which checks
15799 this. New extensions are not added too often so this file can readily
15800 be maintained manually.
15802 There are two big advantages in doing things this way. The extensions
15803 can be looked up immediately and no longer need to be "added" using
15804 X509V3_add_standard_extensions(): this function now does nothing.
15805 Side note: I get *lots* of email saying the extension code doesn't
15806 work because people forget to call this function.
15807 Also no dynamic allocation is done unless new extensions are added:
15808 so if we don't add custom extensions there is no need to call
15809 X509V3_EXT_cleanup().
15813 * Modify enc utility's salting as follows: make salting the default. Add a
15814 magic header, so unsalted files fail gracefully instead of just decrypting
15815 to garbage. This is because not salting is a big security hole, so people
15816 should be discouraged from doing it.
15820 * Fixes and enhancements to the 'x509' utility. It allowed a message
15821 digest to be passed on the command line but it only used this
15822 parameter when signing a certificate. Modified so all relevant
15823 operations are affected by the digest parameter including the
15824 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15825 DSA key was used because it didn't fix the digest.
15829 * Initial certificate chain verify code. Currently tests the untrusted
15830 certificates for consistency with the verify purpose (which is set
15831 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15833 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15834 this is because it will reject chains with invalid extensions whereas
15835 every previous version of OpenSSL and SSLeay made no checks at all.
15837 Trust code: checks the root CA for the relevant trust settings. Trust
15838 settings have an initial value consistent with the verify purpose: e.g.
15839 if the verify purpose is for SSL client use it expects the CA to be
15840 trusted for SSL client use. However the default value can be changed to
15841 permit custom trust settings: one example of this would be to only trust
15842 certificates from a specific "secure" set of CAs.
15844 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15845 which should be used for version portability: especially since the
15846 verify structure is likely to change more often now.
15848 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15849 to set them. If not set then assume SSL clients will verify SSL servers
15852 Two new options to the verify program: -untrusted allows a set of
15853 untrusted certificates to be passed in and -purpose which sets the
15854 intended purpose of the certificate. If a purpose is set then the
15855 new chain verify code is used to check extension consistency.
15859 * Support for the authority information access extension.
15863 * Modify RSA and DSA PEM read routines to transparently handle
15864 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15865 public keys in a format compatible with certificate
15866 SubjectPublicKeyInfo structures. Unfortunately there were already
15867 functions called *_PublicKey_* which used various odd formats so
15868 these are retained for compatibility: however the DSA variants were
15869 never in a public release so they have been deleted. Changed dsa/rsa
15870 utilities to handle the new format: note no releases ever handled public
15871 keys so we should be OK.
15873 The primary motivation for this change is to avoid the same fiasco
15874 that dogs private keys: there are several incompatible private key
15875 formats some of which are standard and some OpenSSL specific and
15876 require various evil hacks to allow partial transparent handling and
15877 even then it doesn't work with DER formats. Given the option anything
15878 other than PKCS#8 should be dumped: but the other formats have to
15879 stay in the name of compatibility.
15881 With public keys and the benefit of hindsight one standard format
15882 is used which works with EVP_PKEY, RSA or DSA structures: though
15883 it clearly returns an error if you try to read the wrong kind of key.
15885 Added a -pubkey option to the 'x509' utility to output the public key.
15886 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15887 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15888 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15889 that do the same as the `EVP_PKEY_assign_*()` except they up the
15890 reference count of the added key (they don't "swallow" the
15895 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15896 CRLs would fail if the file contained no certificates or no CRLs:
15897 added a new function to read in both types and return the number
15898 read: this means that if none are read it will be an error. The
15899 DER versions of the certificate and CRL reader would always fail
15900 because it isn't possible to mix certificates and CRLs in DER format
15901 without choking one or the other routine. Changed this to just read
15902 a certificate: this is the best we can do. Also modified the code
15903 in apps/verify.c to take notice of return codes: it was previously
15904 attempting to read in certificates from NULL pointers and ignoring
15905 any errors: this is one reason why the cert and CRL reader seemed
15906 to work. It doesn't check return codes from the default certificate
15907 routines: these may well fail if the certificates aren't installed.
15911 * Code to support otherName option in GeneralName.
15915 * First update to verify code. Change the verify utility
15916 so it warns if it is passed a self signed certificate:
15917 for consistency with the normal behaviour. X509_verify
15918 has been modified to it will now verify a self signed
15919 certificate if *exactly* the same certificate appears
15920 in the store: it was previously impossible to trust a
15921 single self signed certificate. This means that:
15922 openssl verify ss.pem
15923 now gives a warning about a self signed certificate but
15924 openssl verify -CAfile ss.pem ss.pem
15929 * For servers, store verify_result in SSL_SESSION data structure
15930 (and add it to external session representation).
15931 This is needed when client certificate verifications fails,
15932 but an application-provided verification callback (set by
15933 SSL_CTX_set_cert_verify_callback) allows accepting the session
15934 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15935 but returns 1): When the session is reused, we have to set
15936 ssl->verify_result to the appropriate error code to avoid
15939 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15941 * Fix a bug in the new PKCS#7 code: it didn't consider the
15942 case in PKCS7_dataInit() where the signed PKCS7 structure
15943 didn't contain any existing data because it was being created.
15945 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15947 * Add a salt to the key derivation routines in enc.c. This
15948 forms the first 8 bytes of the encrypted file. Also add a
15949 -S option to allow a salt to be input on the command line.
15953 * New function X509_cmp(). Oddly enough there wasn't a function
15954 to compare two certificates. We do this by working out the SHA1
15955 hash and comparing that. X509_cmp() will be needed by the trust
15960 * SSL_get1_session() is like SSL_get_session(), but increments
15961 the reference count in the SSL_SESSION returned.
15963 *Geoff Thorpe <geoff@eu.c2.net>*
15965 * Fix for 'req': it was adding a null to request attributes.
15966 Also change the X509_LOOKUP and X509_INFO code to handle
15967 certificate auxiliary information.
15971 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15976 * Add the possibility to add extra information to the memory leak
15977 detecting output, to form tracebacks, showing from where each
15978 allocation was originated: CRYPTO_push_info("constant string") adds
15979 the string plus current file name and line number to a per-thread
15980 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15981 is like calling CYRPTO_pop_info() until the stack is empty.
15982 Also updated memory leak detection code to be multi-thread-safe.
15986 * Add options -text and -noout to pkcs7 utility and delete the
15987 encryption options which never did anything. Update docs.
15991 * Add options to some of the utilities to allow the pass phrase
15992 to be included on either the command line (not recommended on
15993 OSes like Unix) or read from the environment. Update the
15994 manpages and fix a few bugs.
15998 * Add a few manpages for some of the openssl commands.
16002 * Fix the -revoke option in ca. It was freeing up memory twice,
16003 leaking and not finding already revoked certificates.
16007 * Extensive changes to support certificate auxiliary information.
16008 This involves the use of X509_CERT_AUX structure and X509_AUX
16009 functions. An X509_AUX function such as PEM_read_X509_AUX()
16010 can still read in a certificate file in the usual way but it
16011 will also read in any additional "auxiliary information". By
16012 doing things this way a fair degree of compatibility can be
16013 retained: existing certificates can have this information added
16014 using the new 'x509' options.
16016 Current auxiliary information includes an "alias" and some trust
16017 settings. The trust settings will ultimately be used in enhanced
16018 certificate chain verification routines: currently a certificate
16019 can only be trusted if it is self signed and then it is trusted
16024 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16025 The problem was that one of the replacement routines had not been working
16026 since SSLeay releases. For now the offending routine has been replaced
16027 with non-optimised assembler. Even so, this now gives around 95%
16028 performance improvement for 1024 bit RSA signs.
16032 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16033 handling. Most clients have the effective key size in bits equal to
16034 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16035 A few however don't do this and instead use the size of the decrypted key
16036 to determine the RC2 key length and the AlgorithmIdentifier to determine
16037 the effective key length. In this case the effective key length can still
16038 be 40 bits but the key length can be 168 bits for example. This is fixed
16039 by manually forcing an RC2 key into the EVP_PKEY structure because the
16040 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16041 the key length and effective key length are equal.
16045 * Add a bunch of functions that should simplify the creation of
16046 X509_NAME structures. Now you should be able to do:
16047 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16048 and have it automatically work out the correct field type and fill in
16049 the structures. The more adventurous can try:
16050 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16051 and it will (hopefully) work out the correct multibyte encoding.
16055 * Change the 'req' utility to use the new field handling and multibyte
16056 copy routines. Before the DN field creation was handled in an ad hoc
16057 way in req, ca, and x509 which was rather broken and didn't support
16058 BMPStrings or UTF8Strings. Since some software doesn't implement
16059 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16060 using the dirstring_type option. See the new comment in the default
16061 openssl.cnf for more info.
16065 * Make crypto/rand/md_rand.c more robust:
16066 - Assure unique random numbers after fork().
16067 - Make sure that concurrent threads access the global counter and
16068 md serializably so that we never lose entropy in them
16069 or use exactly the same state in multiple threads.
16070 Access to the large state is not always serializable because
16071 the additional locking could be a performance killer, and
16072 md should be large enough anyway.
16076 * New file apps/app_rand.c with commonly needed functionality
16077 for handling the random seed file.
16079 Use the random seed file in some applications that previously did not:
16081 dsaparam -genkey (which also ignored its '-rand' option),
16084 x509 (when signing).
16085 Except on systems with /dev/urandom, it is crucial to have a random
16086 seed file at least for key creation, DSA signing, and for DH exchanges;
16087 for RSA signatures we could do without one.
16089 gendh and gendsa (unlike genrsa) used to read only the first byte
16090 of each file listed in the '-rand' option. The function as previously
16091 found in genrsa is now in app_rand.c and is used by all programs
16092 that support '-rand'.
16096 * In RAND_write_file, use mode 0600 for creating files;
16097 don't just chmod when it may be too late.
16101 * Report an error from X509_STORE_load_locations
16102 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16106 * New function ASN1_mbstring_copy() this copies a string in either
16107 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16108 into an ASN1_STRING type. A mask of permissible types is passed
16109 and it chooses the "minimal" type to use or an error if not type
16114 * Add function equivalents to the various macros in asn1.h. The old
16115 macros are retained with an `M_` prefix. Code inside the library can
16116 use the `M_` macros. External code (including the openssl utility)
16117 should *NOT* in order to be "shared library friendly".
16121 * Add various functions that can check a certificate's extensions
16122 to see if it usable for various purposes such as SSL client,
16123 server or S/MIME and CAs of these types. This is currently
16124 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16125 verification. Also added a -purpose flag to x509 utility to
16126 print out all the purposes.
16130 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16135 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16136 for, obtain and decode and extension and obtain its critical flag.
16137 This allows all the necessary extension code to be handled in a
16138 single function call.
16142 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16143 platforms. See crypto/rc4/rc4_enc.c for further details.
16147 * New -noout option to asn1parse. This causes no output to be produced
16148 its main use is when combined with -strparse and -out to extract data
16149 from a file (which may not be in ASN.1 format).
16153 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16154 when producing the local key id.
16156 *Richard Levitte <levitte@stacken.kth.se>*
16158 * New option -dhparam in s_server. This allows a DH parameter file to be
16159 stated explicitly. If it is not stated then it tries the first server
16160 certificate file. The previous behaviour hard coded the filename
16165 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16166 a public key to be input or output. For example:
16167 openssl rsa -in key.pem -pubout -out pubkey.pem
16168 Also added necessary DSA public key functions to handle this.
16172 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16173 in the message. This was handled by allowing
16174 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16176 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16178 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16179 to the end of the strings whereas this didn't. This would cause problems
16180 if strings read with d2i_ASN1_bytes() were later modified.
16182 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16184 * Fix for base64 decode bug. When a base64 bio reads only one line of
16185 data and it contains EOF it will end up returning an error. This is
16186 caused by input 46 bytes long. The cause is due to the way base64
16187 BIOs find the start of base64 encoded data. They do this by trying a
16188 trial decode on each line until they find one that works. When they
16189 do a flag is set and it starts again knowing it can pass all the
16190 data directly through the decoder. Unfortunately it doesn't reset
16191 the context it uses. This means that if EOF is reached an attempt
16192 is made to pass two EOFs through the context and this causes the
16193 resulting error. This can also cause other problems as well. As is
16194 usual with these problems it takes *ages* to find and the fix is
16195 trivial: move one line.
16197 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16199 * Ugly workaround to get s_client and s_server working under Windows. The
16200 old code wouldn't work because it needed to select() on sockets and the
16201 tty (for keypresses and to see if data could be written). Win32 only
16202 supports select() on sockets so we select() with a 1s timeout on the
16203 sockets and then see if any characters are waiting to be read, if none
16204 are present then we retry, we also assume we can always write data to
16205 the tty. This isn't nice because the code then blocks until we've
16206 received a complete line of data and it is effectively polling the
16207 keyboard at 1s intervals: however it's quite a bit better than not
16208 working at all :-) A dedicated Windows application might handle this
16209 with an event loop for example.
16213 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16214 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16215 will be called when RSA_sign() and RSA_verify() are used. This is useful
16216 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16217 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16218 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16219 This necessitated the support of an extra signature type NID_md5_sha1
16220 for SSL signatures and modifications to the SSL library to use it instead
16221 of calling RSA_public_decrypt() and RSA_private_encrypt().
16225 * Add new -verify -CAfile and -CApath options to the crl program, these
16226 will lookup a CRL issuers certificate and verify the signature in a
16227 similar way to the verify program. Tidy up the crl program so it
16228 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16229 less strict. It will now permit CRL extensions even if it is not
16230 a V2 CRL: this will allow it to tolerate some broken CRLs.
16234 * Initialize all non-automatic variables each time one of the openssl
16235 sub-programs is started (this is necessary as they may be started
16236 multiple times from the "OpenSSL>" prompt).
16238 *Lennart Bang, Bodo Moeller*
16240 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16241 removing all other RSA functionality (this is what NO_RSA does). This
16242 is so (for example) those in the US can disable those operations covered
16243 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16248 * Non-copying interface to BIO pairs.
16249 (still largely untested)
16253 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16254 ASCII string. This was handled independently in various places before.
16258 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16259 UTF8 strings a character at a time.
16263 * Use client_version from client hello to select the protocol
16264 (s23_srvr.c) and for RSA client key exchange verification
16265 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16269 * Add various utility functions to handle SPKACs, these were previously
16270 handled by poking round in the structure internals. Added new function
16271 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16272 print, verify and generate SPKACs. Based on an original idea from
16273 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16277 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16281 * Allow the config file extension section to be overwritten on the
16282 command line. Based on an original idea from Massimiliano Pala
16283 <madwolf@comune.modena.it>. The new option is called -extensions
16284 and can be applied to ca, req and x509. Also -reqexts to override
16285 the request extensions in req and -crlexts to override the crl extensions
16290 * Add new feature to the SPKAC handling in ca. Now you can include
16291 the same field multiple times by preceding it by "XXXX." for example:
16294 this is the same syntax as used in the req config file.
16298 * Allow certificate extensions to be added to certificate requests. These
16299 are specified in a 'req_extensions' option of the req section of the
16300 config file. They can be printed out with the -text option to req but
16301 are otherwise ignored at present.
16305 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16306 data read consists of only the final block it would not decrypted because
16307 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16308 A misplaced 'break' also meant the decrypted final block might not be
16309 copied until the next read.
16313 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16314 a few extra parameters to the DH structure: these will be useful if
16315 for example we want the value of 'q' or implement X9.42 DH.
16319 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16320 provides hooks that allow the default DSA functions or functions on a
16321 "per key" basis to be replaced. This allows hardware acceleration and
16322 hardware key storage to be handled without major modification to the
16323 library. Also added low level modexp hooks and CRYPTO_EX structure and
16324 associated functions.
16328 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16329 as "read only": it can't be written to and the buffer it points to will
16330 not be freed. Reading from a read only BIO is much more efficient than
16331 a normal memory BIO. This was added because there are several times when
16332 an area of memory needs to be read from a BIO. The previous method was
16333 to create a memory BIO and write the data to it, this results in two
16334 copies of the data and an O(n^2) reading algorithm. There is a new
16335 function BIO_new_mem_buf() which creates a read only memory BIO from
16336 an area of memory. Also modified the PKCS#7 routines to use read only
16341 * Bugfix: ssl23_get_client_hello did not work properly when called in
16342 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16343 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16344 but a retry condition occurred while trying to read the rest.
16348 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16349 NID_pkcs7_encrypted by default: this was wrong since this should almost
16350 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16351 the encrypted data type: this is a more sensible place to put it and it
16352 allows the PKCS#12 code to be tidied up that duplicated this
16357 * Changed obj_dat.pl script so it takes its input and output files on
16358 the command line. This should avoid shell escape redirection problems
16363 * Initial support for certificate extension requests, these are included
16364 in things like Xenroll certificate requests. Included functions to allow
16365 extensions to be obtained and added.
16369 * -crlf option to s_client and s_server for sending newlines as
16370 CRLF (as required by many protocols).
16374 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16376 * Install libRSAglue.a when OpenSSL is built with RSAref.
16378 *Ralf S. Engelschall*
16380 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16382 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16384 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16389 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16390 DH parameters/keys (q is lost during that conversion, but the resulting
16391 DH parameters contain its length).
16393 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16394 much faster than DH_generate_parameters (which creates parameters
16395 where `p = 2*q + 1`), and also the smaller q makes DH computations
16396 much more efficient (160-bit exponentiation instead of 1024-bit
16397 exponentiation); so this provides a convenient way to support DHE
16398 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16399 utter importance to use
16400 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16402 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16403 when such DH parameters are used, because otherwise small subgroup
16404 attacks may become possible!
16408 * Avoid memory leak in i2d_DHparams.
16412 * Allow the -k option to be used more than once in the enc program:
16413 this allows the same encrypted message to be read by multiple recipients.
16417 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16418 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16419 it will always use the numerical form of the OID, even if it has a short
16424 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16425 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16426 otherwise bn_mod_exp was called. In the case of hardware keys for example
16427 no private key components need be present and it might store extra data
16428 in the RSA structure, which cannot be accessed from bn_mod_exp.
16429 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16430 private key operations.
16434 * Added support for SPARC Linux.
16438 * pem_password_cb function type incompatibly changed from
16439 typedef int pem_password_cb(char *buf, int size, int rwflag);
16441 ....(char *buf, int size, int rwflag, void *userdata);
16442 so that applications can pass data to their callbacks:
16443 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16444 additional void * argument, which is just handed through whenever
16445 the password callback is called.
16447 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16449 New function SSL_CTX_set_default_passwd_cb_userdata.
16451 Compatibility note: As many C implementations push function arguments
16452 onto the stack in reverse order, the new library version is likely to
16453 interoperate with programs that have been compiled with the old
16454 pem_password_cb definition (PEM_whatever takes some data that
16455 happens to be on the stack as its last argument, and the callback
16456 just ignores this garbage); but there is no guarantee whatsoever that
16459 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16460 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16461 problems not only on Windows, but also on some Unix platforms.
16462 To avoid problematic command lines, these definitions are now in an
16463 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16464 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16468 * MIPS III/IV assembler module is reimplemented.
16472 * More DES library cleanups: remove references to srand/rand and
16473 delete an unused file.
16477 * Add support for the free Netwide assembler (NASM) under Win32,
16478 since not many people have MASM (ml) and it can be hard to obtain.
16479 This is currently experimental but it seems to work OK and pass all
16480 the tests. Check out INSTALL.W32 for info.
16484 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16485 without temporary keys kept an extra copy of the server key,
16486 and connections with temporary keys did not free everything in case
16491 * New function RSA_check_key and new openssl rsa option -check
16492 for verifying the consistency of RSA keys.
16494 *Ulf Moeller, Bodo Moeller*
16496 * Various changes to make Win32 compile work:
16497 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16498 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16499 comparison" warnings.
16500 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16504 * Add a debugging option to PKCS#5 v2 key generation function: when
16505 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16506 derived keys are printed to stderr.
16510 * Copy the flags in ASN1_STRING_dup().
16512 *Roman E. Pavlov <pre@mo.msk.ru>*
16514 * The x509 application mishandled signing requests containing DSA
16515 keys when the signing key was also DSA and the parameters didn't match.
16517 It was supposed to omit the parameters when they matched the signing key:
16518 the verifying software was then supposed to automatically use the CA's
16519 parameters if they were absent from the end user certificate.
16521 Omitting parameters is no longer recommended. The test was also
16522 the wrong way round! This was probably due to unusual behaviour in
16523 EVP_cmp_parameters() which returns 1 if the parameters match.
16524 This meant that parameters were omitted when they *didn't* match and
16525 the certificate was useless. Certificates signed with 'ca' didn't have
16528 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16530 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16531 The interface is as follows:
16532 Applications can use
16533 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16534 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16535 "off" is now the default.
16536 The library internally uses
16537 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16538 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16539 to disable memory-checking temporarily.
16541 Some inconsistent states that previously were possible (and were
16542 even the default) are now avoided.
16544 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16545 with each memory chunk allocated; this is occasionally more helpful
16546 than just having a counter.
16548 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16550 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16555 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16556 which largely parallels "options", but is for changing API behaviour,
16557 whereas "options" are about protocol behaviour.
16558 Initial "mode" flags are:
16560 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16561 a single record has been written.
16562 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16563 retries use the same buffer location.
16564 (But all of the contents must be
16569 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16572 * Fix problems with no-hmac etc.
16574 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16576 * New functions RSA_get_default_method(), RSA_set_method() and
16577 RSA_get_method(). These allows replacement of RSA_METHODs without having
16578 to mess around with the internals of an RSA structure.
16582 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16583 Also really enable memory leak checks in openssl.c and in some
16586 *Chad C. Mulligan, Bodo Moeller*
16588 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16589 up the length of negative integers. This has now been simplified to just
16590 store the length when it is first determined and use it later, rather
16591 than trying to keep track of where data is copied and updating it to
16593 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16595 * Add a new function PKCS7_signatureVerify. This allows the verification
16596 of a PKCS#7 signature but with the signing certificate passed to the
16597 function itself. This contrasts with PKCS7_dataVerify which assumes the
16598 certificate is present in the PKCS#7 structure. This isn't always the
16599 case: certificates can be omitted from a PKCS#7 structure and be
16600 distributed by "out of band" means (such as a certificate database).
16604 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16605 function prototypes in pem.h, also change util/mkdef.pl to add the
16606 necessary function names.
16610 * mk1mf.pl (used by Windows builds) did not properly read the
16611 options set by Configure in the top level Makefile, and Configure
16612 was not even able to write more than one option correctly.
16613 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16617 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16618 file to be loaded from a BIO or FILE pointer. The BIO version will
16619 for example allow memory BIOs to contain config info.
16623 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16624 Whoever hopes to achieve shared-library compatibility across versions
16625 must use this, not the compile-time macro.
16626 (Exercise 0.9.4: Which is the minimum library version required by
16628 Note: All this applies only to multi-threaded programs, others don't
16633 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16634 through a BIO pair triggered the default case, i.e.
16635 SSLerr(...,SSL_R_UNKNOWN_STATE).
16639 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16640 can use the SSL library even if none of the specific BIOs is
16645 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16646 for the encoded length.
16648 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16650 * Add initial documentation of the X509V3 functions.
16654 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16655 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16656 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16657 secure PKCS#8 private key format with a high iteration count.
16661 * Fix determination of Perl interpreter: A perl or perl5
16662 *directory* in $PATH was also accepted as the interpreter.
16664 *Ralf S. Engelschall*
16666 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16667 wrong with it but it was very old and did things like calling
16668 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16669 unusual formatting.
16673 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16674 to use the new extension code.
16678 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16679 with macros. This should make it easier to change their form, add extra
16680 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16685 * Add to configuration table a new entry that can specify an alternative
16686 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16687 according to Mark Crispin <MRC@Panda.COM>.
16691 * DES CBC did not update the IV. Weird.
16695 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16696 Changing the behaviour of the former might break existing programs --
16697 where IV updating is needed, des_ncbc_encrypt can be used.
16700 * When bntest is run from "make test" it drives bc to check its
16701 calculations, as well as internally checking them. If an internal check
16702 fails, it needs to cause bc to give a non-zero result or make test carries
16703 on without noticing the failure. Fixed.
16707 * DES library cleanups.
16711 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16712 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16713 ciphers. NOTE: although the key derivation function has been verified
16714 against some published test vectors it has not been extensively tested
16715 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16720 * Instead of "mkdir -p", which is not fully portable, use new
16721 Perl script "util/mkdir-p.pl".
16725 * Rewrite the way password based encryption (PBE) is handled. It used to
16726 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16727 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16728 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16729 the 'parameter' field of the AlgorithmIdentifier is passed to the
16730 underlying key generation function so it must do its own ASN1 parsing.
16731 This has also changed the EVP_PBE_CipherInit() function which now has a
16732 'parameter' argument instead of literal salt and iteration count values
16733 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16737 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16738 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16739 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16740 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16741 value was just used as a "magic string" and not used directly its
16742 value doesn't matter.
16746 * Introduce some semblance of const correctness to BN. Shame C doesn't
16751 * "linux-sparc64" configuration (ultrapenguin).
16753 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16754 "linux-sparc" configuration.
16756 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16758 * config now generates no-xxx options for missing ciphers.
16762 * Support the EBCDIC character set (work in progress).
16763 File ebcdic.c not yet included because it has a different license.
16765 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16767 * Support BS2000/OSD-POSIX.
16769 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16771 * Make callbacks for key generation use `void *` instead of `char *`.
16775 * Make S/MIME samples compile (not yet tested).
16779 * Additional typesafe stacks.
16783 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16787 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16789 * New configuration variant "sco5-gcc".
16791 * Updated some demos.
16793 *Sean O Riordain, Wade Scholine*
16795 * Add missing BIO_free at exit of pkcs12 application.
16799 * Fix memory leak in conf.c.
16803 * Updates for Win32 to assembler version of MD5.
16807 * Set #! path to perl in apps/der_chop to where we found it
16808 instead of using a fixed path.
16812 * SHA library changes for irix64-mips4-cc.
16816 * Improvements for VMS support.
16820 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16822 * Bignum library bug fix. IRIX 6 passes "make test" now!
16823 This also avoids the problems with SC4.2 and unpatched SC5.
16825 *Andy Polyakov <appro@fy.chalmers.se>*
16827 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16828 These are required because of the typesafe stack would otherwise break
16829 existing code. If old code used a structure member which used to be STACK
16830 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16831 sk_num or sk_value it would produce an error because the num, data members
16832 are not present in STACK_OF. Now it just produces a warning. sk_set
16833 replaces the old method of assigning a value to sk_value
16834 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16835 that does this will no longer work (and should use sk_set instead) but
16836 this could be regarded as a "questionable" behaviour anyway.
16840 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16841 correctly handle encrypted S/MIME data.
16845 * Change type of various DES function arguments from des_cblock
16846 (which means, in function argument declarations, pointer to char)
16847 to des_cblock * (meaning pointer to array with 8 char elements),
16848 which allows the compiler to do more typechecking; it was like
16849 that back in SSLeay, but with lots of ugly casts.
16851 Introduce new type const_des_cblock.
16855 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16856 problems: find RecipientInfo structure that matches recipient certificate
16857 and initialise the ASN1 structures properly based on passed cipher.
16861 * Belatedly make the BN tests actually check the results.
16865 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16866 to and from BNs: it was completely broken. New compilation option
16867 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16868 key elements as negative integers.
16872 * Reorganize and speed up MD5.
16874 *Andy Polyakov <appro@fy.chalmers.se>*
16878 *Richard Levitte <richard@levitte.org>*
16880 * New option -out to asn1parse to allow the parsed structure to be
16881 output to a file. This is most useful when combined with the -strparse
16882 option to examine the output of things like OCTET STRINGS.
16886 * Make SSL library a little more fool-proof by not requiring any longer
16887 that `SSL_set_{accept,connect}_state` be called before
16888 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16889 in many applications because usually everything *appeared* to work as
16890 intended anyway -- now it really works as intended).
16894 * Move openssl.cnf out of lib/.
16898 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16899 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16900 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16902 *Ralf S. Engelschall*
16904 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16905 handle PKCS#7 enveloped data properly.
16907 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16909 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16910 copying pointers. The cert_st handling is changed by this in
16911 various ways (and thus what used to be known as ctx->default_cert
16912 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16913 any longer when s->cert does not give us what we need).
16914 ssl_cert_instantiate becomes obsolete by this change.
16915 As soon as we've got the new code right (possibly it already is?),
16916 we have solved a couple of bugs of the earlier code where s->cert
16917 was used as if it could not have been shared with other SSL structures.
16919 Note that using the SSL API in certain dirty ways now will result
16920 in different behaviour than observed with earlier library versions:
16921 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16922 does not influence s as it used to.
16924 In order to clean up things more thoroughly, inside SSL_SESSION
16925 we don't use CERT any longer, but a new structure SESS_CERT
16926 that holds per-session data (if available); currently, this is
16927 the peer's certificate chain and, for clients, the server's certificate
16928 and temporary key. CERT holds only those values that can have
16929 meaningful defaults in an SSL_CTX.
16933 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16934 from the internal representation. Various PKCS#7 fixes: remove some
16935 evil casts and set the enc_dig_alg field properly based on the signing
16940 * Allow PKCS#12 password to be set from the command line or the
16941 environment. Let 'ca' get its config file name from the environment
16942 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16947 * Allow certificate policies extension to use an IA5STRING for the
16948 organization field. This is contrary to the PKIX definition but
16949 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16954 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16955 without disallowing inline assembler and the like for non-pedantic builds.
16959 * Support Borland C++ builder.
16961 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16967 * SHA-1 cleanups and performance enhancements.
16969 *Andy Polyakov <appro@fy.chalmers.se>*
16971 * Sparc v8plus assembler for the bignum library.
16973 *Andy Polyakov <appro@fy.chalmers.se>*
16975 * Accept any -xxx and +xxx compiler options in Configure.
16979 * Update HPUX configuration.
16983 * Add missing `sk_<type>_unshift()` function to safestack.h
16985 *Ralf S. Engelschall*
16987 * New function SSL_CTX_use_certificate_chain_file that sets the
16988 "extra_cert"s in addition to the certificate. (This makes sense
16989 only for "PEM" format files, as chains as a whole are not
16994 * Support verify_depth from the SSL API.
16995 x509_vfy.c had what can be considered an off-by-one-error:
16996 Its depth (which was not part of the external interface)
16997 was actually counting the number of certificates in a chain;
16998 now it really counts the depth.
17002 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17003 instead of X509err, which often resulted in confusing error
17004 messages since the error codes are not globally unique
17005 (e.g. an alleged error in ssl3_accept when a certificate
17006 didn't match the private key).
17008 * New function SSL_CTX_set_session_id_context that allows to set a default
17009 value (so that you don't need SSL_set_session_id_context for each
17010 connection using the SSL_CTX).
17014 * OAEP decoding bug fix.
17018 * Support INSTALL_PREFIX for package builders, as proposed by
17023 * New Configure options "threads" and "no-threads". For systems
17024 where the proper compiler options are known (currently Solaris
17025 and Linux), "threads" is the default.
17029 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17033 * Install various scripts to $(OPENSSLDIR)/misc, not to
17034 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17035 such as /usr/local/bin.
17039 * "make linux-shared" to build shared libraries.
17041 *Niels Poppe <niels@netbox.org>*
17043 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17047 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17048 extension adding in x509 utility.
17052 * Remove NOPROTO sections and error code comments.
17056 * Partial rewrite of the DEF file generator to now parse the ANSI
17061 * New Configure options --prefix=DIR and --openssldir=DIR.
17065 * Complete rewrite of the error code script(s). It is all now handled
17066 by one script at the top level which handles error code gathering,
17067 header rewriting and C source file generation. It should be much better
17068 than the old method: it now uses a modified version of Ulf's parser to
17069 read the ANSI prototypes in all header files (thus the old K&R definitions
17070 aren't needed for error creation any more) and do a better job of
17071 translating function codes into names. The old 'ASN1 error code embedded
17072 in a comment' is no longer necessary and it doesn't use .err files which
17073 have now been deleted. Also the error code call doesn't have to appear all
17074 on one line (which resulted in some large lines...).
17078 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17082 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17083 0 (which usually indicates a closed connection), but continue reading.
17087 * Fix some race conditions.
17091 * Add support for CRL distribution points extension. Add Certificate
17092 Policies and CRL distribution points documentation.
17096 * Move the autogenerated header file parts to crypto/opensslconf.h.
17100 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17101 8 of keying material. Merlin has also confirmed interop with this fix
17102 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17104 *Merlin Hughes <merlin@baltimore.ie>*
17106 * Fix lots of warnings.
17108 *Richard Levitte <levitte@stacken.kth.se>*
17110 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17111 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17113 *Richard Levitte <levitte@stacken.kth.se>*
17115 * Fix problems with sizeof(long) == 8.
17117 *Andy Polyakov <appro@fy.chalmers.se>*
17119 * Change functions to ANSI C.
17123 * Fix typos in error codes.
17125 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17127 * Remove defunct assembler files from Configure.
17131 * SPARC v8 assembler BIGNUM implementation.
17133 *Andy Polyakov <appro@fy.chalmers.se>*
17135 * Support for Certificate Policies extension: both print and set.
17136 Various additions to support the r2i method this uses.
17140 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17141 return a const string when you are expecting an allocated buffer.
17145 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17146 types DirectoryString and DisplayText.
17150 * Add code to allow r2i extensions to access the configuration database,
17151 add an LHASH database driver and add several ctx helper functions.
17155 * Fix an evil bug in bn_expand2() which caused various BN functions to
17156 fail when they extended the size of a BIGNUM.
17160 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17161 support typesafe stack.
17165 * Fix typo in SSL_[gs]et_options().
17167 *Nils Frostberg <nils@medcom.se>*
17169 * Delete various functions and files that belonged to the (now obsolete)
17170 old X509V3 handling code.
17174 * New Configure option "rsaref".
17178 * Don't auto-generate pem.h.
17182 * Introduce type-safe ASN.1 SETs.
17186 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17188 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17190 * Introduce type-safe STACKs. This will almost certainly break lots of code
17191 that links with OpenSSL (well at least cause lots of warnings), but fear
17192 not: the conversion is trivial, and it eliminates loads of evil casts. A
17193 few STACKed things have been converted already. Feel free to convert more.
17194 In the fullness of time, I'll do away with the STACK type altogether.
17198 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17199 specified in `<certfile>` by updating the entry in the index.txt file.
17200 This way one no longer has to edit the index.txt file manually for
17201 revoking a certificate. The -revoke option does the gory details now.
17203 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17205 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17206 `-text` option at all and this way the `-noout -text` combination was
17207 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17209 *Ralf S. Engelschall*
17211 * Make sure a corresponding plain text error message exists for the
17212 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17213 verify callback function determined that a certificate was revoked.
17215 *Ralf S. Engelschall*
17217 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17218 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17219 all available ciphers including rc5, which was forgotten until now.
17220 In order to let the testing shell script know which algorithms
17221 are available, a new (up to now undocumented) command
17222 `openssl list-cipher-commands` is used.
17226 * Bugfix: s_client occasionally would sleep in select() when
17227 it should have checked SSL_pending() first.
17231 * New functions DSA_do_sign and DSA_do_verify to provide access to
17232 the raw DSA values prior to ASN.1 encoding.
17236 * Tweaks to Configure
17238 *Niels Poppe <niels@netbox.org>*
17240 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17245 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17249 * New config option to avoid instructions that are illegal on the 80386.
17250 The default code is faster, but requires at least a 486.
17254 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17255 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17256 same as SSL2_VERSION anyway.
17260 * New "-showcerts" option for s_client.
17264 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17265 application. Various cleanups and fixes.
17269 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17270 modify error routines to work internally. Add error codes and PBE init
17271 to library startup routines.
17275 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17276 packing functions to asn1 and evp. Changed function names and error
17277 codes along the way.
17281 * PKCS12 integration: and so it begins... First of several patches to
17282 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17283 objects to objects.h
17287 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17288 and display support for Thawte strong extranet extension.
17292 * Add LinuxPPC support.
17294 *Jeff Dubrule <igor@pobox.org>*
17296 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17297 bn_div_words in alpha.s.
17299 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17301 * Make sure the RSA OAEP test is skipped under -DRSAref because
17302 OAEP isn't supported when OpenSSL is built with RSAref.
17304 *Ulf Moeller <ulf@fitug.de>*
17306 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17307 so they no longer are missing under -DNOPROTO.
17309 *Soren S. Jorvang <soren@t.dk>*
17311 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17313 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17314 doesn't work when the session is reused. Coming soon!
17318 * Fix a security hole, that allows sessions to be reused in the wrong
17319 context thus bypassing client cert protection! All software that uses
17320 client certs and session caches in multiple contexts NEEDS PATCHING to
17321 allow session reuse! A fuller solution is in the works.
17323 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17325 * Some more source tree cleanups (removed obsolete files
17326 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17327 permission on "config" script to be executable) and a fix for the INSTALL
17330 *Ulf Moeller <ulf@fitug.de>*
17332 * Remove some legacy and erroneous uses of malloc, free instead of
17335 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17337 * Make rsa_oaep_test return non-zero on error.
17339 *Ulf Moeller <ulf@fitug.de>*
17341 * Add support for native Solaris shared libraries. Configure
17342 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17343 if someone would make that last step automatic.
17345 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17347 * ctx_size was not built with the right compiler during "make links". Fixed.
17351 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17352 except NULL ciphers". This means the default cipher list will no longer
17353 enable NULL ciphers. They need to be specifically enabled e.g. with
17354 the string "DEFAULT:eNULL".
17358 * Fix to RSA private encryption routines: if p < q then it would
17359 occasionally produce an invalid result. This will only happen with
17360 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17364 * Be less restrictive and allow also `perl util/perlpath.pl
17365 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17366 because this way one can also use an interpreter named `perl5' (which is
17367 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17368 installed as `perl').
17370 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17372 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17374 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17376 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17377 advapi32.lib to Win32 build and change the pem test comparison
17378 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17379 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17380 and crypto/des/ede_cbcm_enc.c.
17384 * DES quad checksum was broken on big-endian architectures. Fixed.
17388 * Comment out two functions in bio.h that aren't implemented. Fix up the
17389 Win32 test batch file so it (might) work again. The Win32 test batch file
17390 is horrible: I feel ill....
17394 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17395 in e_os.h. Audit of header files to check ANSI and non ANSI
17396 sections: 10 functions were absent from non ANSI section and not exported
17397 from Windows DLLs. Fixed up libeay.num for new functions.
17401 * Make `openssl version' output lines consistent.
17403 *Ralf S. Engelschall*
17405 * Fix Win32 symbol export lists for BIO functions: Added
17406 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17407 to ms/libeay{16,32}.def.
17409 *Ralf S. Engelschall*
17411 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17412 fine under Unix and passes some trivial tests I've now added. But the
17413 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17414 added to make sure no one expects that this stuff really works in the
17415 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17416 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17419 *Ralf S. Engelschall*
17421 * Fix the generation of two part addresses in perl.
17423 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17425 * Add config entry for Linux on MIPS.
17427 *John Tobey <jtobey@channel1.com>*
17429 * Make links whenever Configure is run, unless we are on Windoze.
17433 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17434 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17439 * Add a useful kludge to allow package maintainers to specify compiler and
17440 other platforms details on the command line without having to patch the
17441 Configure script every time: One now can use
17442 `perl Configure <id>:<details>`,
17443 i.e. platform ids are allowed to have details appended
17444 to them (separated by colons). This is treated as there would be a static
17445 pre-configured entry in Configure's %table under key `<id>` with value
17446 `<details>` and `perl Configure <id>` is called. So, when you want to
17447 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17448 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17449 now, which overrides the FreeBSD-elf entry on-the-fly.
17451 *Ralf S. Engelschall*
17453 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17457 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17458 on the `perl Configure ...' command line. This way one can compile
17459 OpenSSL libraries with Position Independent Code (PIC) which is needed
17460 for linking it into DSOs.
17462 *Ralf S. Engelschall*
17464 * Remarkably, export ciphers were totally broken and no-one had noticed!
17469 * Cleaned up the LICENSE document: The official contact for any license
17470 questions now is the OpenSSL core team under openssl-core@openssl.org.
17471 And add a paragraph about the dual-license situation to make sure people
17472 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17473 to the OpenSSL toolkit.
17475 *Ralf S. Engelschall*
17477 * General source tree makefile cleanups: Made `making xxx in yyy...'
17478 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17479 Additionally cleaned up the `make links' target: Remove unnecessary
17480 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17481 to speed processing and no longer clutter the display with confusing
17482 stuff. Instead only the actually done links are displayed.
17484 *Ralf S. Engelschall*
17486 * Permit null encryption ciphersuites, used for authentication only. It used
17487 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17488 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17493 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17494 signed attributes when verifying signatures (this would break them),
17495 the detached data encoding was wrong and public keys obtained using
17496 X509_get_pubkey() weren't freed.
17500 * Add text documentation for the BUFFER functions. Also added a work around
17501 to a Win95 console bug. This was triggered by the password read stuff: the
17502 last character typed gets carried over to the next fread(). If you were
17503 generating a new cert request using 'req' for example then the last
17504 character of the passphrase would be CR which would then enter the first
17509 * Added the new 'Includes OpenSSL Cryptography Software' button as
17510 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17511 button and can be used by applications based on OpenSSL to show the
17512 relationship to the OpenSSL project.
17514 *Ralf S. Engelschall*
17516 * Remove confusing variables in function signatures in files
17517 ssl/ssl_lib.c and ssl/ssl.h.
17519 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17521 * Don't install bss_file.c under PREFIX/include/
17523 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17525 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17526 functions that return function pointers and has support for NT specific
17527 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17528 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17529 unsigned to signed types: this was killing the Win32 compile.
17533 * Add new certificate file to stack functions,
17534 SSL_add_dir_cert_subjects_to_stack() and
17535 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17536 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17537 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17538 This means that Apache-SSL and similar packages don't have to mess around
17539 to add as many CAs as they want to the preferred list.
17543 * Experiment with doxygen documentation. Currently only partially applied to
17545 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17546 openssl.doxy as the configuration file.
17550 * Get rid of remaining C++-style comments which strict C compilers hate.
17552 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17554 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17555 compiled in by default: it has problems with large keys.
17559 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17560 DH private keys and/or callback functions which directly correspond to
17561 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17562 is needed for applications which have to configure certificates on a
17563 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17565 For the RSA certificate situation is makes no difference, but
17566 for the DSA certificate situation this fixes the "no shared cipher"
17567 problem where the OpenSSL cipher selection procedure failed because the
17568 temporary keys were not overtaken from the context and the API provided
17569 no way to reconfigure them.
17570 The new functions now let applications reconfigure the stuff and they
17571 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17572 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17573 non-public-API function ssl_cert_instantiate() is used as a helper
17574 function and also to reduce code redundancy inside ssl_rsa.c.
17576 *Ralf S. Engelschall*
17578 * Move s_server -dcert and -dkey options out of the undocumented feature
17579 area because they are useful for the DSA situation and should be
17580 recognized by the users.
17582 *Ralf S. Engelschall*
17584 * Fix the cipher decision scheme for export ciphers: the export bits are
17585 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17586 SSL_EXP_MASK. So, the original variable has to be used instead of the
17587 already masked variable.
17589 *Richard Levitte <levitte@stacken.kth.se>*
17591 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17593 *Richard Levitte <levitte@stacken.kth.se>*
17595 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17596 from `int` to `unsigned int` because it is a length and initialized by
17597 EVP_DigestFinal() which expects an `unsigned int *`.
17599 *Richard Levitte <levitte@stacken.kth.se>*
17601 * Don't hard-code path to Perl interpreter on shebang line of Configure
17602 script. Instead use the usual Shell->Perl transition trick.
17604 *Ralf S. Engelschall*
17606 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17607 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17608 -noout -modulus' as it's already the case for `openssl rsa -noout
17609 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17610 currently the public key is printed (a decision which was already done by
17611 `openssl dsa -modulus' in the past) which serves a similar purpose.
17612 Additionally the NO_RSA no longer completely removes the whole -modulus
17613 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17616 *Ralf S. Engelschall*
17618 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17619 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17621 *Arne Ansper <arne@ats.cyber.ee>*
17623 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17624 to be added. Now both 'req' and 'ca' can use new objects defined in the
17629 * Add cool BIO that does syslog (or event log on NT).
17631 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17633 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17634 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17635 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17636 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17640 * Add preliminary config info for new extension code.
17644 * Make RSA_NO_PADDING really use no padding.
17646 *Ulf Moeller <ulf@fitug.de>*
17648 * Generate errors when private/public key check is done.
17652 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17653 for some CRL extensions and new objects added.
17657 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17658 key usage extension and fuller support for authority key id.
17662 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17663 padding method for RSA, which is recommended for new applications in PKCS
17664 #1 v2.0 (RFC 2437, October 1998).
17665 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17666 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17667 against Bleichbacher's attack on RSA.
17668 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17671 * Updates to the new SSL compression code
17673 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17675 * Fix so that the version number in the master secret, when passed
17676 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17677 (because the server will not accept higher), that the version number
17678 is 0x03,0x01, not 0x03,0x00
17680 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17682 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17683 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17684 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17688 * Support for RAW extensions where an arbitrary extension can be
17689 created by including its DER encoding. See apps/openssl.cnf for
17694 * Make sure latest Perl versions don't interpret some generated C array
17695 code as Perl array code in the crypto/err/err_genc.pl script.
17697 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17699 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17700 not many people have the assembler. Various Win32 compilation fixes and
17701 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17702 build instructions.
17706 * Modify configure script 'Configure' to automatically create crypto/date.h
17707 file under Win32 and also build pem.h from pem.org. New script
17708 util/mkfiles.pl to create the MINFO file on environments that can't do a
17709 'make files': perl util/mkfiles.pl >MINFO should work.
17713 * Major rework of DES function declarations, in the pursuit of correctness
17714 and purity. As a result, many evil casts evaporated, and some weirdness,
17715 too. You may find this causes warnings in your code. Zapping your evil
17716 casts will probably fix them. Mostly.
17720 * Fix for a typo in asn1.h. Bug fix to object creation script
17721 obj_dat.pl. It considered a zero in an object definition to mean
17722 "end of object": none of the objects in objects.h have any zeros
17723 so it wasn't spotted.
17725 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17727 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17728 Masking (CBCM). In the absence of test vectors, the best I have been able
17729 to do is check that the decrypt undoes the encrypt, so far. Send me test
17730 vectors if you have them.
17734 * Correct calculation of key length for export ciphers (too much space was
17735 allocated for null ciphers). This has not been tested!
17739 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17740 message is now correct (it understands "crypto" and "ssl" on its
17741 command line). There is also now an "update" option. This will update
17742 the util/ssleay.num and util/libeay.num files with any new functions.
17744 perl util/mkdef.pl crypto ssl update
17745 it will update them.
17749 * Overhauled the Perl interface:
17750 - ported BN stuff to OpenSSL's different BN library
17751 - made the perl/ source tree CVS-aware
17752 - renamed the package from SSLeay to OpenSSL (the files still contain
17753 their history because I've copied them in the repository)
17754 - removed obsolete files (the test scripts will be replaced
17755 by better Test::Harness variants in the future)
17757 *Ralf S. Engelschall*
17759 * First cut for a very conservative source tree cleanup:
17760 1. merge various obsolete readme texts into doc/ssleay.txt
17761 where we collect the old documents and readme texts.
17762 2. remove the first part of files where I'm already sure that we no
17763 longer need them because of three reasons: either they are just temporary
17764 files which were left by Eric or they are preserved original files where
17765 I've verified that the diff is also available in the CVS via "cvs diff
17766 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17767 the crypto/md/ stuff).
17769 *Ralf S. Engelschall*
17771 * More extension code. Incomplete support for subject and issuer alt
17772 name, issuer and authority key id. Change the i2v function parameters
17773 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17774 what that's for :-) Fix to ASN1 macro which messed up
17775 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17779 * Preliminary support for ENUMERATED type. This is largely copied from the
17784 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17786 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17788 * Make sure `make rehash` target really finds the `openssl` program.
17790 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17792 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17793 like to hear about it if this slows down other processors.
17797 * Add CygWin32 platform information to Configure script.
17799 *Alan Batie <batie@aahz.jf.intel.com>*
17801 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17803 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17805 * New program nseq to manipulate netscape certificate sequences
17809 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17814 * Fixes to BN code. Previously the default was to define BN_RECURSION
17815 but the BN code had some problems that would cause failures when
17816 doing certificate verification and some other functions.
17818 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17820 * Add ASN1 and PEM code to support netscape certificate sequences.
17824 * Add ASN1 and PEM code to support netscape certificate sequences.
17828 * Add several PKIX and private extended key usage OIDs.
17832 * Modify the 'ca' program to handle the new extension code. Modify
17833 openssl.cnf for new extension format, add comments.
17837 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17838 and add a sample to openssl.cnf so req -x509 now adds appropriate
17843 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17844 error code, add initial support to X509_print() and x509 application.
17848 * Takes a deep breath and start adding X509 V3 extension support code. Add
17849 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17850 stuff is currently isolated and isn't even compiled yet.
17854 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17855 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17856 Removed the versions check from X509 routines when loading extensions:
17857 this allows certain broken certificates that don't set the version
17858 properly to be processed.
17862 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17863 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17864 can still be regenerated with "make depend".
17868 * Spelling mistake in C version of CAST-128.
17870 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17872 * Changes to the error generation code. The perl script err-code.pl
17873 now reads in the old error codes and retains the old numbers, only
17874 adding new ones if necessary. It also only changes the .err files if new
17875 codes are added. The makefiles have been modified to only insert errors
17876 when needed (to avoid needlessly modifying header files). This is done
17877 by only inserting errors if the .err file is newer than the auto generated
17878 C file. To rebuild all the error codes from scratch (the old behaviour)
17879 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17880 or delete all the .err files.
17884 * CAST-128 was incorrectly implemented for short keys. The C version has
17885 been fixed, but is untested. The assembler versions are also fixed, but
17886 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17887 to regenerate it if needed.
17888 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17889 Hagino <itojun@kame.net>*
17891 * File was opened incorrectly in randfile.c.
17893 *Ulf Möller <ulf@fitug.de>*
17895 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17896 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17897 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17898 al: it's just almost always a UTCTime. Note this patch adds new error
17899 codes so do a "make errors" if there are problems.
17903 * Correct Linux 1 recognition in config.
17905 *Ulf Möller <ulf@fitug.de>*
17907 * Remove pointless MD5 hash when using DSA keys in ca.
17909 *Anonymous <nobody@replay.com>*
17911 * Generate an error if given an empty string as a cert directory. Also
17912 generate an error if handed NULL (previously returned 0 to indicate an
17913 error, but didn't set one).
17915 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17917 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17921 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17922 parameters. This was causing a warning which killed off the Win32 compile.
17926 * Remove C++ style comments from crypto/bn/bn_local.h.
17928 *Neil Costigan <neil.costigan@celocom.com>*
17930 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17931 based on a text string, looking up short and long names and finally
17932 "dot" format. The "dot" format stuff didn't work. Added new function
17933 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17934 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17935 OID is not part of the table.
17939 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17940 X509_LOOKUP_by_alias().
17944 * Sort openssl functions by name.
17948 * Get the gendsa program working (hopefully) and add it to app list. Remove
17949 encryption from sample DSA keys (in case anyone is interested the password
17954 * Make *all* `*_free` functions accept a NULL pointer.
17956 *Frans Heymans <fheymans@isaserver.be>*
17958 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17961 *Anonymous <nobody@replay.com>*
17963 * s_server should send the CAfile as acceptable CAs, not its own cert.
17965 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17967 * Don't blow it for numeric -newkey arguments to apps/req.
17969 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17971 * Temp key "for export" tests were wrong in s3_srvr.c.
17973 *Anonymous <nobody@replay.com>*
17975 * Add prototype for temp key callback functions
17976 SSL_CTX_set_tmp_{rsa,dh}_callback().
17980 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17981 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17985 * X509_name_add_entry() freed the wrong thing after an error.
17987 *Arne Ansper <arne@ats.cyber.ee>*
17989 * rsa_eay.c would attempt to free a NULL context.
17991 *Arne Ansper <arne@ats.cyber.ee>*
17993 * BIO_s_socket() had a broken should_retry() on Windoze.
17995 *Arne Ansper <arne@ats.cyber.ee>*
17997 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17999 *Arne Ansper <arne@ats.cyber.ee>*
18001 * Make sure the already existing X509_STORE->depth variable is initialized
18002 in X509_STORE_new(), but document the fact that this variable is still
18003 unused in the certificate verification process.
18005 *Ralf S. Engelschall*
18007 * Fix the various library and apps files to free up pkeys obtained from
18008 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18012 * Fix reference counting in X509_PUBKEY_get(). This makes
18013 demos/maurice/example2.c work, amongst others, probably.
18015 *Steve Henson and Ben Laurie*
18017 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18018 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18019 are no longer created. This way we have a single and consistent command
18020 line interface `openssl <command>`, similar to `cvs <command>`.
18022 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18024 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18025 BIT STRING wrapper always have zero unused bits.
18029 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18033 * Make the top-level INSTALL documentation easier to understand.
18037 * Makefiles updated to exit if an error occurs in a sub-directory
18038 make (including if user presses ^C) [Paul Sutton]
18040 * Make Montgomery context stuff explicit in RSA data structure.
18044 * Fix build order of pem and err to allow for generated pem.h.
18048 * Fix renumbering bug in X509_NAME_delete_entry().
18052 * Enhanced the err-ins.pl script so it makes the error library number
18053 global and can add a library name. This is needed for external ASN1 and
18054 other error libraries.
18058 * Fixed sk_insert which never worked properly.
18062 * Fix ASN1 macros so they can handle indefinite length constructed
18063 EXPLICIT tags. Some non standard certificates use these: they can now
18068 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18069 into a single doc/ssleay.txt bundle. This way the information is still
18070 preserved but no longer messes up this directory. Now it's new room for
18071 the new set of documentation files.
18073 *Ralf S. Engelschall*
18075 * SETs were incorrectly DER encoded. This was a major pain, because they
18076 shared code with SEQUENCEs, which aren't coded the same. This means that
18077 almost everything to do with SETs or SEQUENCEs has either changed name or
18078 number of arguments.
18080 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18082 * Fix test data to work with the above.
18086 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18087 was already fixed by Eric for 0.9.1 it seems.
18089 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18091 * Autodetect FreeBSD3.
18095 * Fix various bugs in Configure. This affects the following platforms:
18099 unixware-2.0-pentium
18104 * Eliminate generated files from CVS. Reorder tests to regenerate files
18105 before they are needed.
18109 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18113 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18115 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18116 changed SSLeay to OpenSSL in version strings.
18118 *Ralf S. Engelschall*
18120 * Some fixups to the top-level documents.
18124 * Fixed the nasty bug where rsaref.h was not found under compile-time
18125 because the symlink to include/ was missing.
18127 *Ralf S. Engelschall*
18129 * Incorporated the popular no-RSA/DSA-only patches
18130 which allow to compile a RSA-free SSLeay.
18132 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18134 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18135 when "ssleay" is still not found.
18137 *Ralf S. Engelschall*
18139 * Added more platforms to Configure: Cray T3E, HPUX 11,
18141 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18143 * Updated the README file.
18145 *Ralf S. Engelschall*
18147 * Added various .cvsignore files in the CVS repository subdirs
18148 to make a "cvs update" really silent.
18150 *Ralf S. Engelschall*
18152 * Recompiled the error-definition header files and added
18153 missing symbols to the Win32 linker tables.
18155 *Ralf S. Engelschall*
18157 * Cleaned up the top-level documents;
18158 o new files: CHANGES and LICENSE
18159 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18160 o merged COPYRIGHT into LICENSE
18161 o removed obsolete TODO file
18162 o renamed MICROSOFT to INSTALL.W32
18164 *Ralf S. Engelschall*
18166 * Removed dummy files from the 0.9.1b source tree:
18167 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18168 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18169 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18170 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18171 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18173 *Ralf S. Engelschall*
18175 * Added various platform portability fixes.
18179 * The Genesis of the OpenSSL rpject:
18180 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18181 Young and Tim J. Hudson created while they were working for C2Net until
18184 *The OpenSSL Project*
18186 ### Changes between 0.9.0b and 0.9.1b [not released]
18188 * Updated a few CA certificates under certs/
18192 * Changed some BIGNUM api stuff.
18196 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18197 DGUX x86, Linux Alpha, etc.
18201 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18202 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18207 * Add -strparse option to asn1pars program which parses nested
18210 *Dr Stephen Henson <shenson@bigfoot.com>*
18212 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18216 * DSA fix for "ca" program.
18220 * Added "-genkey" option to "dsaparam" program.
18224 * Added RIPE MD160 (rmd160) message digest.
18228 * Added -a (all) option to "ssleay version" command.
18232 * Added PLATFORM define which is the id given to Configure.
18236 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18240 * Extended the ASN.1 parser routines.
18244 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18248 * Added a BN_CTX to the BN library.
18252 * Fixed the weak key values in DES library
18256 * Changed API in EVP library for cipher aliases.
18260 * Added support for RC2/64bit cipher.
18264 * Converted the lhash library to the crypto/mem.c functions.
18268 * Added more recognized ASN.1 object ids.
18272 * Added more RSA padding checks for SSL/TLS.
18276 * Added BIO proxy/filter functionality.
18280 * Added extra_certs to SSL_CTX which can be used
18281 send extra CA certificates to the client in the CA cert chain sending
18282 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18286 * Now Fortezza is denied in the authentication phase because
18287 this is key exchange mechanism is not supported by SSLeay at all.
18291 * Additional PKCS1 checks.
18295 * Support the string "TLSv1" for all TLS v1 ciphers.
18299 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18300 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18304 * Fixed a few memory leaks.
18308 * Fixed various code and comment typos.
18312 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18313 bytes sent in the client random.
18315 *Edward Bishop <ebishop@spyglass.com>*
18319 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18320 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18321 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18322 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18323 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18324 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18325 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18326 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18327 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18328 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18329 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18330 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18331 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18332 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18333 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18334 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18335 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18336 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18337 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18338 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18339 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18340 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18341 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18342 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18343 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18344 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18345 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18346 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18347 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18348 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18349 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18350 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18351 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18352 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18353 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18354 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18355 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18356 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18357 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18358 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18359 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18360 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18361 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18362 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18363 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18364 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18365 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18366 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18367 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18368 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18369 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18370 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18371 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18372 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18373 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18374 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18375 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18376 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18377 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18378 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18379 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18380 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18381 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18382 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18383 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18384 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18385 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18386 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18387 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18388 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18389 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18390 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18391 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18392 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18393 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18394 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18395 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18396 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18397 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18398 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18399 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18400 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18401 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18402 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18403 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18404 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18405 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18406 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18407 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18408 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18409 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18410 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18411 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18412 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18413 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18414 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18415 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18416 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18417 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18418 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18419 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18420 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18421 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18422 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18423 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18424 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18425 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18426 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18427 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18428 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18429 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18430 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18431 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18432 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18433 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18434 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18435 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18436 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18437 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18438 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18439 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18440 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18441 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18442 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18443 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18444 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18445 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18446 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18447 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18448 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18449 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18450 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18451 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18452 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18453 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18454 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18455 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18456 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18457 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18458 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18459 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18460 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18461 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18462 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18463 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18464 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18465 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18466 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18467 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18468 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18469 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18470 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18471 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18472 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18473 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18474 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18475 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18476 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18477 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18478 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655