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 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
27 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
31 * 'Configure' has been changed to figure out the configuration target if
32 none is given on the command line. Consequently, the 'config' script is
33 now only a mere wrapper. All documentation is changed to only mention
36 *Rich Salz and Richard Levitte*
38 * Added a library context that applications as well as other
39 libraries can use to form a separate context within which libcrypto
40 operations are performed.
42 There are two ways this can be used:
44 - Directly, by passing a library context to functions that take
45 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
47 - Indirectly, by creating a new library context and then assigning
48 it as the new default, with `OPENSSL_CTX_set0_default`.
50 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
51 apart from the functions directly related to `OPENSSL_CTX`, accept
52 NULL to indicate that the default library context should be used.
54 Library code that changes the default library context using
55 `OPENSSL_CTX_set0_default` should take care to restore it with a
56 second call before returning to the caller.
60 * Handshake now fails if Extended Master Secret extension is dropped
65 * Dropped interactive mode from the 'openssl' program. From now on,
66 the `openssl` command without arguments is equivalent to `openssl
71 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
72 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
73 While the old function names have been retained for backward compatibility
74 they should not be used in new developments
75 because their return values are confusing: Unlike other `_cmp()` functions
76 they do not return 0 in case their arguments are equal.
80 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
81 EC_GROUP_get_field_type().
85 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
86 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
87 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
88 Applications should rely on the library automatically assigning a suitable
89 EC_METHOD internally upon EC_GROUP construction.
93 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
94 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
95 assigned internally without application intervention.
96 Users of EC_GROUP_new() should switch to a different suitable constructor.
100 * Add CAdES-BES signature verification support, mostly derived
101 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
103 *Filipe Raimundo da Silva*
105 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
109 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
110 functions are not widely used and now OpenSSL automatically perform this
111 conversion when needed.
115 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
116 EC_KEY_precompute_mult(). These functions are not widely used and
117 applications should instead switch to named curves which OpenSSL has
118 hardcoded lookup tables for.
122 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
123 should instead use the L<EC_POINT_mul(3)> function.
127 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
128 that are not applicable to the new provider model. Applications should
129 instead use EVP_default_properties_is_fips_enabled() and
130 EVP_default_properties_enable_fips().
134 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
135 is set, an unexpected EOF is ignored, it pretends a close notify was received
136 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
140 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
141 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
142 used and applications should instead use the
143 L<EC_POINT_set_affine_coordinates(3)> and
144 L<EC_POINT_get_affine_coordinates(3)> functions.
148 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
149 arrays to be more easily constructed via a series of utility functions.
150 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
151 the various push functions and finally convert to a passable OSSL_PARAM
152 array using OSSL_PARAM_BLD_to_param().
156 * The security strength of SHA1 and MD5 based signatures in TLS has been
157 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
158 working at the default security level of 1 and instead requires security
159 level 0. The security level can be changed either using the cipher string
160 with @SECLEVEL, or calling SSL_CTX_set_security_level().
164 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
165 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
166 internal keys, if they correspond to one of those built in types.
170 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
171 contain a provider side internal key.
175 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
176 They are old functions that we don't use, and that you could disable with
177 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
181 * Project text documents not yet having a proper file name extension
182 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
183 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
187 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
188 have been converted to Markdown with the goal to produce documents
189 which not only look pretty when viewed online in the browser, but
190 remain well readable inside a plain text editor.
192 To achieve this goal, a 'minimalistic' Markdown style has been applied
193 which avoids formatting elements that interfere too much with the
194 reading flow in the text file. For example, it
196 * avoids [ATX headings][] and uses [setext headings][] instead
197 (which works for `<h1>` and `<h2>` headings only).
198 * avoids [inline links][] and uses [reference links][] instead.
199 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
201 [ATX headings]: https://github.github.com/gfm/#atx-headings
202 [setext headings]: https://github.github.com/gfm/#setext-headings
203 [inline links]: https://github.github.com/gfm/#inline-link
204 [reference links]: https://github.github.com/gfm/#reference-link
205 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
206 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
208 *Matthias St. Pierre*
210 * The test suite is changed to preserve results of each test recipe.
211 A new directory test-runs/ with subdirectories named like the
212 test recipes are created in the build tree for this purpose.
216 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
217 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
218 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
220 *David von Oheimb, Martin Peylo*
222 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
223 The legacy OCSP-focused and only partly documented API is retained.
224 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
228 * BIO_do_connect and BIO_do_handshake have been extended:
229 If domain name resolution yields multiple IP addresses all of them are tried
230 after connect() failures.
234 * All of the low level RSA functions have been deprecated including:
236 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
237 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
238 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
239 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
240 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
241 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
242 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
243 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
244 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
245 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
246 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
247 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
248 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
249 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
250 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
251 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
252 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
253 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
254 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
255 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
256 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
257 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
258 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
259 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
260 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
261 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
262 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
263 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
265 Use of these low level functions has been informally discouraged for a long
266 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
267 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
268 L<EVP_PKEY_decrypt(3)>.
272 * X509 certificates signed using SHA1 are no longer allowed at security
274 In TLS/SSL the default security level is 1. It can be set either
275 using the cipher string with @SECLEVEL, or calling
276 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
277 a call to SSL_CTX_use_certificate() will fail if the security level is not
279 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
280 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
285 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
286 modified to use PKEY APIs. These commands are now in maintenance mode
287 and no new features will be added to them.
291 * The command line utility rsautl has been deprecated.
292 Instead use the pkeyutl program.
296 * The command line utilities genrsa and rsa have been modified to use PKEY
297 APIs These commands are now in maintenance mode and no new features will
302 * All of the low level DH functions have been deprecated including:
304 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
305 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
306 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
307 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
308 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
309 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
310 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
311 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
312 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
313 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
314 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
315 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
316 DH_meth_get_generate_params and DH_meth_set_generate_params.
318 Use of these low level functions has been informally discouraged for a long
319 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
320 and L<EVP_PKEY_derive(3)>.
324 * All of the low level DSA functions have been deprecated including:
326 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
327 DSA_get_default_method, DSA_set_method, DSA_get_method,
328 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
329 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
330 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
331 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
332 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
333 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
334 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
335 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
336 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
337 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
338 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
341 Use of these low level functions has been informally discouraged for a long
342 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
343 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
347 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
348 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
349 This means that applications don't have to look at the curve NID and
350 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
351 However, they still can, that EVP_PKEY_set_alias_type() call acts as
352 a no-op when the EVP_PKEY is already of the given type.
354 Parameter and key generation is also reworked to make it possible
355 to generate EVP_PKEY_SM2 parameters and keys without having to go
356 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
357 However, code that does the latter will still work as before.
361 * Deprecated low level ECDH and ECDSA functions. These include:
363 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
364 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
367 Use of these low level functions has been informally discouraged for a long
368 time. Instead applications should use the EVP_PKEY_derive(3),
369 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
373 * Deprecated the EC_KEY_METHOD functions. These include:
375 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
376 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
377 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
378 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
379 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
380 EC_KEY_METHOD_get_verify.
382 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
386 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
387 and EVP_PKEY_decrypt() instead.
388 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
389 and EVP_PKEY_encrypt() instead.
393 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
394 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
395 a new formulation to include all the things it can be used for,
396 as well as words of caution.
400 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
401 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
405 * All of the low level HMAC functions have been deprecated including:
407 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
408 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
411 Use of these low level functions has been informally discouraged for a long
412 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
413 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
414 and L<EVP_MAC_final(3)>.
418 * Over two thousand fixes were made to the documentation, including:
419 - Common options (such as -rand/-writerand, TLS version control, etc)
420 were refactored and point to newly-enhanced descriptions in openssl.pod.
421 - Added style conformance for all options (with help from Richard Levitte),
422 documented all reported missing options, added a CI build to check
423 that all options are documented and that no unimplemented options
425 - Documented some internals, such as all use of environment variables.
426 - Addressed all internal broken L<> references.
430 * All of the low level CMAC functions have been deprecated including:
432 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
433 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
435 Use of these low level functions has been informally discouraged for a long
436 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
437 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
438 and L<EVP_MAC_final(3)>.
442 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
443 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
446 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
447 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
448 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
449 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
450 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
451 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
452 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
453 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
454 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
455 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
457 Use of these low level functions has been informally discouraged
458 for a long time. Applications should use the EVP_DigestInit_ex(3),
459 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
463 * Corrected the documentation of the return values from the `EVP_DigestSign*`
464 set of functions. The documentation mentioned negative values for some
465 errors, but this was never the case, so the mention of negative values
468 Code that followed the documentation and thereby check with something
469 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
473 * All of the low level cipher functions have been deprecated including:
475 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
476 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
477 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
478 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
479 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
480 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
481 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
482 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
483 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
484 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
485 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
486 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
487 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
488 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
489 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
490 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
491 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
492 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
493 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
494 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
495 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
496 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
497 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
498 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
499 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
500 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
501 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
502 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
503 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
505 Use of these low level functions has been informally discouraged for
506 a long time. Applications should use the high level EVP APIs, e.g.
507 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
508 equivalently named decrypt functions instead.
510 *Matt Caswell and Paul Dale*
512 * Removed include/openssl/opensslconf.h.in and replaced it with
513 include/openssl/configuration.h.in, which differs in not including
514 <openssl/macros.h>. A short header include/openssl/opensslconf.h
515 was added to include both.
517 This allows internal hacks where one might need to modify the set
518 of configured macros, for example this if deprecated symbols are
519 still supposed to be available internally:
521 #include <openssl/configuration.h>
523 #undef OPENSSL_NO_DEPRECATED
524 #define OPENSSL_SUPPRESS_DEPRECATED
526 #include <openssl/macros.h>
528 This should not be used by applications that use the exported
529 symbols, as that will lead to linking errors.
533 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
534 used in exponentiation with 512-bit moduli. No EC algorithms are
535 affected. Analysis suggests that attacks against 2-prime RSA1024,
536 3-prime RSA1536, and DSA1024 as a result of this defect would be very
537 difficult to perform and are not believed likely. Attacks against DH512
538 are considered just feasible. However, for an attack the target would
539 have to re-use the DH512 private key, which is not recommended anyway.
540 Also applications directly using the low level API BN_mod_exp may be
541 affected if they use BN_FLG_CONSTTIME.
546 * Most memory-debug features have been deprecated, and the functionality
547 replaced with no-ops.
551 * Added documentation for the STACK API. OpenSSL only defines the STACK
552 functions where they are used.
556 * Introduced a new method type and API, OSSL_SERIALIZER, to
557 represent generic serializers. An implementation is expected to
558 be able to serialize an object associated with a given name (such
559 as an algorithm name for an asymmetric key) into forms given by
560 implementation properties.
562 Serializers are primarily used from inside libcrypto, through
563 calls to functions like EVP_PKEY_print_private(),
564 PEM_write_bio_PrivateKey() and similar.
566 Serializers are specified in such a way that they can be made to
567 directly handle the provider side portion of an object, if this
568 provider side part comes from the same provider as the serializer
569 itself, but can also be made to handle objects in parametrized
570 form (as an OSSL_PARAM array of data). This allows a provider to
571 offer generic serializers as a service for any other provider.
575 * Added a .pragma directive to the syntax of configuration files, to
576 allow varying behavior in a supported and predictable manner.
577 Currently added pragma:
581 This allows dollar signs to be a keyword character unless it's
582 followed by a opening brace or parenthesis. This is useful for
583 platforms where dollar signs are commonly used in names, such as
584 volume names and system directory names on VMS.
588 * Added functionality to create an EVP_PKEY from user data. This
589 is effectively the same as creating a RSA, DH or DSA object and
590 then assigning them to an EVP_PKEY, but directly using algorithm
591 agnostic EVP functions. A benefit is that this should be future
592 proof for public key algorithms to come.
596 * Change the interpretation of the '--api' configuration option to
597 mean that this is a desired API compatibility level with no
598 further meaning. The previous interpretation, that this would
599 also mean to remove all deprecated symbols up to and including
600 the given version, no requires that 'no-deprecated' is also used
601 in the configuration.
603 When building applications, the desired API compatibility level
604 can be set with the OPENSSL_API_COMPAT macro like before. For
605 API compatibility version below 3.0, the old style numerical
606 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
607 For version 3.0 and on, the value is expected to be the decimal
608 value calculated from the major and minor version like this:
610 MAJOR * 10000 + MINOR * 100
614 -DOPENSSL_API_COMPAT=30000 For 3.0
615 -DOPENSSL_API_COMPAT=30200 For 3.2
617 To hide declarations that are deprecated up to and including the
618 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
619 given when building the application as well.
623 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
624 access to certificate and CRL stores via URIs and OSSL_STORE
627 This adds the following functions:
629 - X509_LOOKUP_store()
630 - X509_STORE_load_file()
631 - X509_STORE_load_path()
632 - X509_STORE_load_store()
633 - SSL_add_store_cert_subjects_to_stack()
634 - SSL_CTX_set_default_verify_store()
635 - SSL_CTX_load_verify_file()
636 - SSL_CTX_load_verify_dir()
637 - SSL_CTX_load_verify_store()
641 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
642 The presence of this system service is determined at run-time.
646 * Added functionality to create an EVP_PKEY context based on data
647 for methods from providers. This takes an algorithm name and a
648 property query string and simply stores them, with the intent
649 that any operation that uses this context will use those strings
650 to fetch the needed methods implicitly, thereby making the port
651 of application written for pre-3.0 OpenSSL easier.
655 * The undocumented function NCONF_WIN32() has been deprecated; for
656 conversion details see the HISTORY section of doc/man5/config.pod
660 * Introduced the new functions EVP_DigestSignInit_ex() and
661 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
662 EVP_DigestVerifyUpdate() have been converted to functions. See the man
663 pages for further details.
667 * Over two thousand fixes were made to the documentation, including:
668 adding missing command flags, better style conformance, documentation
671 *Rich Salz, Richard Levitte*
673 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
674 X25519, X448, Ed25519 and Ed448.
678 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
683 * Deprecated the public definition of ERR_STATE as well as the function
684 ERR_get_state(). This is done in preparation of making ERR_STATE an
689 * Added ERR functionality to give callers access to the stored function
690 names that have replaced the older function code based functions.
692 New functions are ERR_get_error_func(), ERR_peek_error_func(),
693 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
694 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
695 and ERR_peek_last_error_all().
697 These functions have become deprecated: ERR_get_error_line_data(),
698 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
699 ERR_func_error_string().
703 * Extended testing to be verbose for failing tests only. The make variables
704 VERBOSE_FAILURE or VF can be used to enable this:
706 $ make VF=1 test # Unix
707 $ mms /macro=(VF=1) test ! OpenVMS
708 $ nmake VF=1 test # Windows
712 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
713 used even when parsing explicit parameters, when loading a serialized key
714 or calling `EC_GROUP_new_from_ecpkparameters()`/
715 `EC_GROUP_new_from_ecparameters()`.
716 This prevents bypass of security hardening and performance gains,
717 especially for curves with specialized EC_METHODs.
718 By default, if a key encoded with explicit parameters is loaded and later
719 serialized, the output is still encoded with explicit parameters, even if
720 internally a "named" EC_GROUP is used for computation.
724 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
725 this change, EC_GROUP_set_generator would accept order and/or cofactor as
726 NULL. After this change, only the cofactor parameter can be NULL. It also
727 does some minimal sanity checks on the passed order.
732 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
733 An attack is simple, if the first CMS_recipientInfo is valid but the
734 second CMS_recipientInfo is chosen ciphertext. If the second
735 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
736 encryption key will be replaced by garbage, and the message cannot be
737 decoded, but if the RSA decryption fails, the correct encryption key is
738 used and the recipient will not notice the attack.
739 As a work around for this potential attack the length of the decrypted
740 key must be equal to the cipher default key length, in case the
741 certifiate is not given and all recipientInfo are tried out.
742 The old behaviour can be re-enabled in the CMS code by setting the
743 CMS_DEBUG_DECRYPT flag.
747 * Early start up entropy quality from the DEVRANDOM seed source has been
748 improved for older Linux systems. The RAND subsystem will wait for
749 /dev/random to be producing output before seeding from /dev/urandom.
750 The seeded state is stored for future library initialisations using
751 a system global shared memory segment. The shared memory identifier
752 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
753 the desired value. The default identifier is 114.
757 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
758 when primes for RSA keys are computed.
759 Since we previously always generated primes == 2 (mod 3) for RSA keys,
760 the 2-prime and 3-prime RSA modules were easy to distinguish, since
761 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
762 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
763 This avoids possible fingerprinting of newly generated RSA modules.
767 * Correct the extended master secret constant on EBCDIC systems. Without this
768 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
769 negotiate EMS will fail. Unfortunately this also means that TLS connections
770 between EBCDIC systems with this fix, and EBCDIC systems without this
771 fix will fail if they negotiate EMS.
775 * Changed the library initialisation so that the config file is now loaded
776 by default. This was already the case for libssl. It now occurs for both
777 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
778 OPENSSL_init_crypto() to suppress automatic loading of a config file.
782 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
783 where the former acts as a replacement for ERR_put_error(), and the
784 latter replaces the combination ERR_put_error()+ERR_add_error_data().
785 ERR_raise_data() adds more flexibility by taking a format string and
786 an arbitrary number of arguments following it, to be processed with
791 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
792 to check if a named provider is loaded and available. When called, it
793 will also activate all fallback providers if such are still present.
797 * Enforce a minimum DH modulus size of 512 bits.
801 * Changed DH parameters to generate the order q subgroup instead of 2q.
802 Previously generated DH parameters are still accepted by DH_check
803 but DH_generate_key works around that by clearing bit 0 of the
804 private key for those. This avoids leaking bit 0 of the private key.
808 * Significantly reduce secure memory usage by the randomness pools.
812 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
817 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
818 algorithms. An implementation of a key exchange algorithm can be obtained
819 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
820 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
821 the older EVP_PKEY_derive_init() function. See the man pages for the new
822 functions for further details.
826 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
830 * Removed the function names from error messages and deprecated the
833 * Removed NextStep support and the macro OPENSSL_UNISTD
837 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
838 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
839 Also removed "export var as function" capability; we do not export
840 variables, only functions.
844 * RC5_32_set_key has been changed to return an int type, with 0 indicating
845 an error and 1 indicating success. In previous versions of OpenSSL this
846 was a void type. If a key was set longer than the maximum possible this
851 * Support SM2 signing and verification schemes with X509 certificate.
855 * Use SHA256 as the default digest for TS query in the ts app.
859 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
860 This checks that the salt length is at least 128 bits, the derived key
861 length is at least 112 bits, and that the iteration count is at least 1000.
862 For backwards compatibility these checks are disabled by default in the
863 default provider, but are enabled by default in the fips provider.
864 To enable or disable these checks use the control
865 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
869 * Default cipher lists/suites are now available via a function, the
870 #defines are deprecated.
874 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
875 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
876 for Windows Store apps easier. Also, the "no-uplink" option has been added.
880 * Join the directories crypto/x509 and crypto/x509v3
884 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
885 This changes the size when using the genpkey app when no size is given. It
886 fixes an omission in earlier changes that changed all RSA, DSA and DH
887 generation apps to use 2048 bits by default.
891 * Added command 'openssl kdf' that uses the EVP_KDF API.
895 * Added command 'openssl mac' that uses the EVP_MAC API.
899 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
900 as default directories. Also added the command 'openssl info'
901 for scripting purposes.
905 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
906 deprecated. These undocumented functions were never integrated into the EVP
907 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
908 Bi-directional IGE mode. These modes were never formally standardised and
909 usage of these functions is believed to be very small. In particular
910 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
911 is ever used. The security implications are believed to be minimal, but
912 this issue was never fixed for backwards compatibility reasons. New code
913 should not use these modes.
917 * Add prediction resistance to the DRBG reseeding process.
921 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
922 mandated by IEEE Std 1619-2018.
926 * Added newline escaping functionality to a filename when using openssl dgst.
927 This output format is to replicate the output format found in the `*sum`
928 checksum programs. This aims to preserve backward compatibility.
930 *Matt Eaton, Richard Levitte, and Paul Dale*
932 * Removed the heartbeat message in DTLS feature, as it has very
933 little usage and doesn't seem to fulfill a valuable purpose.
934 The configuration option is now deprecated.
938 * Changed the output of 'openssl {digestname} < file' to display the
939 digest name in its output.
943 * Added a new generic trace API which provides support for enabling
944 instrumentation through trace output. This feature is mainly intended
945 as an aid for developers and is disabled by default. To utilize it,
946 OpenSSL needs to be configured with the `enable-trace` option.
948 If the tracing API is enabled, the application can activate trace output
949 by registering BIOs as trace channels for a number of tracing and debugging
952 The 'openssl' application has been expanded to enable any of the types
953 available via environment variables defined by the user, and serves as
954 one possible example on how to use this functionality.
956 *Richard Levitte & Matthias St. Pierre*
958 * Added build tests for C++. These are generated files that only do one
959 thing, to include one public OpenSSL head file each. This tests that
960 the public header files can be usefully included in a C++ application.
962 This test isn't enabled by default. It can be enabled with the option
963 'enable-buildtest-c++'.
967 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
971 * Add KMAC to EVP_MAC.
975 * Added property based algorithm implementation selection framework to
980 * Added SCA hardening for modular field inversion in EC_GROUP through
981 a new dedicated field_inv() pointer in EC_METHOD.
982 This also addresses a leakage affecting conversions from projective
983 to affine coordinates.
985 *Billy Bob Brumley, Nicola Tuveri*
987 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
988 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
989 those algorithms that were already supported through the EVP_PKEY API
990 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
991 and scrypt are now wrappers that call EVP_KDF.
995 * Build devcrypto engine as a dynamic engine.
999 * Add keyed BLAKE2 to EVP_MAC.
1003 * Fix a bug in the computation of the endpoint-pair shared secret used
1004 by DTLS over SCTP. This breaks interoperability with older versions
1005 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1006 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1007 interoperability with such broken implementations. However, enabling
1008 this switch breaks interoperability with correct implementations.
1010 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1011 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1015 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1019 * Change the license to the Apache License v2.0.
1023 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1025 - Major releases (indicated by incrementing the MAJOR release number)
1026 may introduce incompatible API/ABI changes.
1027 - Minor releases (indicated by incrementing the MINOR release number)
1028 may introduce new features but retain API/ABI compatibility.
1029 - Patch releases (indicated by incrementing the PATCH number)
1030 are intended for bug fixes and other improvements of existing
1031 features only (like improving performance or adding documentation)
1032 and retain API/ABI compatibility.
1036 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1040 * Remove the 'dist' target and add a tarball building script. The
1041 'dist' target has fallen out of use, and it shouldn't be
1042 necessary to configure just to create a source distribution.
1046 * Recreate the OS390-Unix config target. It no longer relies on a
1047 special script like it did for OpenSSL pre-1.1.0.
1051 * Instead of having the source directories listed in Configure, add
1052 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1057 * Add GMAC to EVP_MAC.
1061 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1065 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1066 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1067 to facilitate the continued use of MACs through raw private keys in
1068 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1072 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1073 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1077 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1078 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1079 are retained for backwards compatibility.
1083 * AES-XTS mode now enforces that its two keys are different to mitigate
1084 the attacked described in "Efficient Instantiations of Tweakable
1085 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1086 Details of this attack can be obtained from:
1087 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1091 * Rename the object files, i.e. give them other names than in previous
1092 versions. Their names now include the name of the final product, as
1093 well as its type mnemonic (bin, lib, shlib).
1097 * Added new option for 'openssl list', '-objects', which will display the
1098 list of built in objects, i.e. OIDs with names.
1102 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1103 improves application performance by removing data copies and providing
1104 applications with zero-copy system calls such as sendfile and splice.
1108 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1109 option is set, openssl cleanses (zeroize) plaintext bytes from
1110 internal buffers after delivering them to the application. Note,
1111 the application is still responsible for cleansing other copies
1112 (e.g.: data received by SSL_read(3)).
1119 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1121 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1123 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1124 while reading in libssl then we would report an error back to the
1125 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1126 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1127 therefore give a hint as to what went wrong.
1131 * Check that ed25519 and ed448 are allowed by the security level. Previously
1132 signature algorithms not using an MD were not being checked that they were
1133 allowed by the security level.
1137 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1138 was not quite right. The behaviour was not consistent between resumption
1139 and normal handshakes, and also not quite consistent with historical
1140 behaviour. The behaviour in various scenarios has been clarified and
1141 it has been updated to make it match historical behaviour as closely as
1146 * *[VMS only]* The header files that the VMS compilers include automatically,
1147 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1148 that the C++ compiler doesn't understand. This is a shortcoming in the
1149 compiler, but can be worked around with `__cplusplus` guards.
1151 C++ applications that use OpenSSL libraries must be compiled using the
1152 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1153 functions. Otherwise, only functions with symbols of less than 31
1154 characters can be used, as the linker will not be able to successfully
1155 resolve symbols with longer names.
1159 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1160 The presence of this system service is determined at run-time.
1164 * Added newline escaping functionality to a filename when using openssl dgst.
1165 This output format is to replicate the output format found in the `*sum`
1166 checksum programs. This aims to preserve backward compatibility.
1168 *Matt Eaton, Richard Levitte, and Paul Dale*
1170 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1175 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1177 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1178 number generator (RNG). This was intended to include protection in the
1179 event of a fork() system call in order to ensure that the parent and child
1180 processes did not share the same RNG state. However this protection was not
1181 being used in the default case.
1183 A partial mitigation for this issue is that the output from a high
1184 precision timer is mixed into the RNG state so the likelihood of a parent
1185 and child process sharing state is significantly reduced.
1187 If an application already calls OPENSSL_init_crypto() explicitly using
1188 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1191 *Matthias St. Pierre*
1193 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1194 used even when parsing explicit parameters, when loading a serialized key
1195 or calling `EC_GROUP_new_from_ecpkparameters()`/
1196 `EC_GROUP_new_from_ecparameters()`.
1197 This prevents bypass of security hardening and performance gains,
1198 especially for curves with specialized EC_METHODs.
1199 By default, if a key encoded with explicit parameters is loaded and later
1200 serialized, the output is still encoded with explicit parameters, even if
1201 internally a "named" EC_GROUP is used for computation.
1205 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1206 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1207 NULL. After this change, only the cofactor parameter can be NULL. It also
1208 does some minimal sanity checks on the passed order.
1213 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1214 An attack is simple, if the first CMS_recipientInfo is valid but the
1215 second CMS_recipientInfo is chosen ciphertext. If the second
1216 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1217 encryption key will be replaced by garbage, and the message cannot be
1218 decoded, but if the RSA decryption fails, the correct encryption key is
1219 used and the recipient will not notice the attack.
1220 As a work around for this potential attack the length of the decrypted
1221 key must be equal to the cipher default key length, in case the
1222 certifiate is not given and all recipientInfo are tried out.
1223 The old behaviour can be re-enabled in the CMS code by setting the
1224 CMS_DEBUG_DECRYPT flag.
1229 * Early start up entropy quality from the DEVRANDOM seed source has been
1230 improved for older Linux systems. The RAND subsystem will wait for
1231 /dev/random to be producing output before seeding from /dev/urandom.
1232 The seeded state is stored for future library initialisations using
1233 a system global shared memory segment. The shared memory identifier
1234 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1235 the desired value. The default identifier is 114.
1239 * Correct the extended master secret constant on EBCDIC systems. Without this
1240 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1241 negotiate EMS will fail. Unfortunately this also means that TLS connections
1242 between EBCDIC systems with this fix, and EBCDIC systems without this
1243 fix will fail if they negotiate EMS.
1247 * Use Windows installation paths in the mingw builds
1249 Mingw isn't a POSIX environment per se, which means that Windows
1250 paths should be used for installation.
1255 * Changed DH_check to accept parameters with order q and 2q subgroups.
1256 With order 2q subgroups the bit 0 of the private key is not secret
1257 but DH_generate_key works around that by clearing bit 0 of the
1258 private key for those. This avoids leaking bit 0 of the private key.
1262 * Significantly reduce secure memory usage by the randomness pools.
1266 * Revert the DEVRANDOM_WAIT feature for Linux systems
1268 The DEVRANDOM_WAIT feature added a select() call to wait for the
1269 /dev/random device to become readable before reading from the
1270 /dev/urandom device.
1272 It turned out that this change had negative side effects on
1273 performance which were not acceptable. After some discussion it
1274 was decided to revert this feature and leave it up to the OS
1275 resp. the platform maintainer to ensure a proper initialization
1276 during early boot time.
1278 *Matthias St. Pierre*
1280 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1282 * Add build tests for C++. These are generated files that only do one
1283 thing, to include one public OpenSSL head file each. This tests that
1284 the public header files can be usefully included in a C++ application.
1286 This test isn't enabled by default. It can be enabled with the option
1287 'enable-buildtest-c++'.
1291 * Enable SHA3 pre-hashing for ECDSA and DSA.
1295 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1296 This changes the size when using the genpkey app when no size is given. It
1297 fixes an omission in earlier changes that changed all RSA, DSA and DH
1298 generation apps to use 2048 bits by default.
1302 * Reorganize the manual pages to consistently have RETURN VALUES,
1303 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1304 util/fix-doc-nits accordingly.
1306 *Paul Yang, Joshua Lock*
1308 * Add the missing accessor EVP_PKEY_get0_engine()
1312 * Have apps like 's_client' and 's_server' output the signature scheme
1313 along with other cipher suite parameters when debugging.
1317 * Make OPENSSL_config() error agnostic again.
1321 * Do the error handling in RSA decryption constant time.
1325 * Prevent over long nonces in ChaCha20-Poly1305.
1327 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1328 for every encryption operation. RFC 7539 specifies that the nonce value
1329 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1330 and front pads the nonce with 0 bytes if it is less than 12
1331 bytes. However it also incorrectly allows a nonce to be set of up to 16
1332 bytes. In this case only the last 12 bytes are significant and any
1333 additional leading bytes are ignored.
1335 It is a requirement of using this cipher that nonce values are
1336 unique. Messages encrypted using a reused nonce value are susceptible to
1337 serious confidentiality and integrity attacks. If an application changes
1338 the default nonce length to be longer than 12 bytes and then makes a
1339 change to the leading bytes of the nonce expecting the new value to be a
1340 new unique nonce then such an application could inadvertently encrypt
1341 messages with a reused nonce.
1343 Additionally the ignored bytes in a long nonce are not covered by the
1344 integrity guarantee of this cipher. Any application that relies on the
1345 integrity of these ignored leading bytes of a long nonce may be further
1346 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1347 is safe because no such use sets such a long nonce value. However user
1348 applications that use this cipher directly and set a non-default nonce
1349 length to be longer than 12 bytes may be vulnerable.
1351 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1357 * Add DEVRANDOM_WAIT feature for Linux systems
1359 On older Linux systems where the getrandom() system call is not available,
1360 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1361 Contrary to getrandom(), the /dev/urandom device will not block during
1362 early boot when the kernel CSPRNG has not been seeded yet.
1364 To mitigate this known weakness, use select() to wait for /dev/random to
1365 become readable before reading from /dev/urandom.
1367 * Ensure that SM2 only uses SM3 as digest algorithm
1371 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1373 * Change the info callback signals for the start and end of a post-handshake
1374 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1375 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1376 confused by this and assume that a TLSv1.2 renegotiation has started. This
1377 can break KeyUpdate handling. Instead we no longer signal the start and end
1378 of a post handshake message exchange (although the messages themselves are
1379 still signalled). This could break some applications that were expecting
1380 the old signals. However without this KeyUpdate is not usable for many
1385 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1387 * Timing vulnerability in DSA signature generation
1389 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1390 timing side channel attack. An attacker could use variations in the signing
1391 algorithm to recover the private key.
1393 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1398 * Timing vulnerability in ECDSA signature generation
1400 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1401 timing side channel attack. An attacker could use variations in the signing
1402 algorithm to recover the private key.
1404 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1409 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1410 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1411 of two gigabytes and the error handling improved.
1413 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1414 categorized as a normal bug, not a security issue, because the DRBG reseeds
1415 automatically and is fully functional even without additional randomness
1416 provided by the application.
1418 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1420 * Add a new ClientHello callback. Provides a callback interface that gives
1421 the application the ability to adjust the nascent SSL object at the
1422 earliest stage of ClientHello processing, immediately after extensions have
1423 been collected but before they have been processed. In particular, this
1424 callback can adjust the supported TLS versions in response to the contents
1429 * Add SM2 base algorithm support.
1433 * s390x assembly pack: add (improved) hardware-support for the following
1434 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1435 aes-cfb/cfb8, aes-ecb.
1439 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1440 parameter is no longer accepted, as it leads to a corrupt table. NULL
1441 pem_str is reserved for alias entries only.
1445 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1446 step for prime curves. The new implementation is based on formulae from
1447 differential addition-and-doubling in homogeneous projective coordinates
1448 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1449 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1450 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1451 to work in projective coordinates.
1453 *Billy Bob Brumley, Nicola Tuveri*
1455 * Change generating and checking of primes so that the error rate of not
1456 being prime depends on the intended use based on the size of the input.
1457 For larger primes this will result in more rounds of Miller-Rabin.
1458 The maximal error rate for primes with more than 1080 bits is lowered
1461 *Kurt Roeckx, Annie Yousar*
1463 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1467 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1468 moving between systems, and to avoid confusion when a Windows build is
1469 done with mingw vs with MSVC. For POSIX installs, there's still a
1470 symlink or copy named 'tsget' to avoid that confusion as well.
1474 * Revert blinding in ECDSA sign and instead make problematic addition
1475 length-invariant. Switch even to fixed-length Montgomery multiplication.
1479 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1480 step for binary curves. The new implementation is based on formulae from
1481 differential addition-and-doubling in mixed Lopez-Dahab projective
1482 coordinates, modified to independently blind the operands.
1484 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1486 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1487 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1488 EC_METHODs to implement their own specialized "ladder step", to take
1489 advantage of more favorable coordinate systems or more efficient
1490 differential addition-and-doubling algorithms.
1492 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1494 * Modified the random device based seed sources to keep the relevant
1495 file descriptors open rather than reopening them on each access.
1496 This allows such sources to operate in a chroot() jail without
1497 the associated device nodes being available. This behaviour can be
1498 controlled using RAND_keep_random_devices_open().
1502 * Numerous side-channel attack mitigations have been applied. This may have
1503 performance impacts for some algorithms for the benefit of improved
1504 security. Specific changes are noted in this change log by their respective
1509 * AIX shared library support overhaul. Switch to AIX "natural" way of
1510 handling shared libraries, which means collecting shared objects of
1511 different versions and bitnesses in one common archive. This allows to
1512 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1513 doesn't affect the way 3rd party applications are linked, only how
1514 multi-version installation is managed.
1518 * Make ec_group_do_inverse_ord() more robust and available to other
1519 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1520 mitigations are applied to the fallback BN_mod_inverse().
1521 When using this function rather than BN_mod_inverse() directly, new
1522 EC cryptosystem implementations are then safer-by-default.
1526 * Add coordinate blinding for EC_POINT and implement projective
1527 coordinate blinding for generic prime curves as a countermeasure to
1528 chosen point SCA attacks.
1530 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1532 * Add blinding to ECDSA and DSA signatures to protect against side channel
1533 attacks discovered by Keegan Ryan (NCC Group).
1537 * Enforce checking in the pkeyutl command line app to ensure that the input
1538 length does not exceed the maximum supported digest length when performing
1539 a sign, verify or verifyrecover operation.
1543 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1544 I/O in combination with something like select() or poll() will hang. This
1545 can be turned off again using SSL_CTX_clear_mode().
1546 Many applications do not properly handle non-application data records, and
1547 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1548 around the problems in those applications, but can also break some.
1549 It's recommended to read the manpages about SSL_read(), SSL_write(),
1550 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1551 SSL_CTX_set_read_ahead() again.
1555 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1556 now allow empty (zero character) pass phrases.
1560 * Apply blinding to binary field modular inversion and remove patent
1561 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1565 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1566 binary and prime elliptic curves.
1570 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1571 constant time fixed point multiplication.
1575 * Revise elliptic curve scalar multiplication with timing attack
1576 defenses: ec_wNAF_mul redirects to a constant time implementation
1577 when computing fixed point and variable point multiplication (which
1578 in OpenSSL are mostly used with secret scalars in keygen, sign,
1579 ECDH derive operations).
1580 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1583 * Updated CONTRIBUTING
1587 * Updated DRBG / RAND to request nonce and additional low entropy
1588 randomness from the system.
1590 *Matthias St. Pierre*
1592 * Updated 'openssl rehash' to use OpenSSL consistent default.
1596 * Moved the load of the ssl_conf module to libcrypto, which helps
1597 loading engines that libssl uses before libssl is initialised.
1601 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1605 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1607 *Ingo Schwarze, Rich Salz*
1609 * Added output of accepting IP address and port for 'openssl s_server'
1613 * Added a new API for TLSv1.3 ciphersuites:
1614 SSL_CTX_set_ciphersuites()
1615 SSL_set_ciphersuites()
1619 * Memory allocation failures consistently add an error to the error
1624 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1625 in libcrypto when run as setuid/setgid.
1629 * Load any config file by default when libssl is used.
1633 * Added new public header file <openssl/rand_drbg.h> and documentation
1634 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1636 *Matthias St. Pierre*
1638 * QNX support removed (cannot find contributors to get their approval
1639 for the license change).
1643 * TLSv1.3 replay protection for early data has been implemented. See the
1644 SSL_read_early_data() man page for further details.
1648 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1649 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1650 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1651 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1652 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1653 configuration has been separated out. See the ciphers man page or the
1654 SSL_CTX_set_ciphersuites() man page for more information.
1658 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1659 in responder mode now supports the new "-multi" option, which
1660 spawns the specified number of child processes to handle OCSP
1661 requests. The "-timeout" option now also limits the OCSP
1662 responder's patience to wait to receive the full client request
1663 on a newly accepted connection. Child processes are respawned
1664 as needed, and the CA index file is automatically reloaded
1665 when changed. This makes it possible to run the "ocsp" responder
1666 as a long-running service, making the OpenSSL CA somewhat more
1667 feature-complete. In this mode, most diagnostic messages logged
1668 after entering the event loop are logged via syslog(3) rather than
1673 * Added support for X448 and Ed448. Heavily based on original work by
1678 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1679 objects loaded. This adds the functions OSSL_STORE_expect() and
1680 OSSL_STORE_find() as well as needed tools to construct searches and
1681 get the search data out of them.
1685 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1686 version of OpenSSL should review their configuration settings to ensure
1687 that they are still appropriate for TLSv1.3. For further information see:
1688 <https://wiki.openssl.org/index.php/TLS1.3>
1692 * Grand redesign of the OpenSSL random generator
1694 The default RAND method now utilizes an AES-CTR DRBG according to
1695 NIST standard SP 800-90Ar1. The new random generator is essentially
1696 a port of the default random generator from the OpenSSL FIPS 2.0
1697 object module. It is a hybrid deterministic random bit generator
1698 using an AES-CTR bit stream and which seeds and reseeds itself
1699 automatically using trusted system entropy sources.
1701 Some of its new features are:
1702 - Support for multiple DRBG instances with seed chaining.
1703 - The default RAND method makes use of a DRBG.
1704 - There is a public and private DRBG instance.
1705 - The DRBG instances are fork-safe.
1706 - Keep all global DRBG instances on the secure heap if it is enabled.
1707 - The public and private DRBG instance are per thread for lock free
1710 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1712 * Changed Configure so it only says what it does and doesn't dump
1713 so much data. Instead, ./configdata.pm should be used as a script
1714 to display all sorts of configuration data.
1718 * Added processing of "make variables" to Configure.
1722 * Added SHA512/224 and SHA512/256 algorithm support.
1726 * The last traces of Netware support, first removed in 1.1.0, have
1731 * Get rid of Makefile.shared, and in the process, make the processing
1732 of certain files (rc.obj, or the .def/.map/.opt files produced from
1733 the ordinal files) more visible and hopefully easier to trace and
1734 debug (or make silent).
1738 * Make it possible to have environment variable assignments as
1739 arguments to config / Configure.
1743 * Add multi-prime RSA (RFC 8017) support.
1747 * Add SM3 implemented according to GB/T 32905-2016
1748 *Jack Lloyd <jack.lloyd@ribose.com>,*
1749 *Ronald Tse <ronald.tse@ribose.com>,*
1750 *Erick Borsboom <erick.borsboom@ribose.com>*
1752 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1753 as documented in RFC6066.
1754 Based on a patch from Tomasz Moń
1756 *Filipe Raimundo da Silva*
1758 * Add SM4 implemented according to GB/T 32907-2016.
1759 *Jack Lloyd <jack.lloyd@ribose.com>,*
1760 *Ronald Tse <ronald.tse@ribose.com>,*
1761 *Erick Borsboom <erick.borsboom@ribose.com>*
1763 * Reimplement -newreq-nodes and ERR_error_string_n; the
1764 original author does not agree with the license change.
1768 * Add ARIA AEAD TLS support.
1772 * Some macro definitions to support VS6 have been removed. Visual
1773 Studio 6 has not worked since 1.1.0
1777 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1778 without clearing the errors.
1782 * Add "atfork" functions. If building on a system that without
1783 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1784 requirements. The RAND facility now uses/requires this.
1792 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1793 not possible to disable entirely. However, it's still possible to
1794 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1797 To disable, configure with 'no-ui-console'. 'no-ui' is still
1798 possible to use as an alias. Check at compile time with the
1799 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1800 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1804 * Add a STORE module, which implements a uniform and URI based reader of
1805 stores that can contain keys, certificates, CRLs and numerous other
1806 objects. The main API is loosely based on a few stdio functions,
1807 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1808 OSSL_STORE_error and OSSL_STORE_close.
1809 The implementation uses backends called "loaders" to implement arbitrary
1810 URI schemes. There is one built in "loader" for the 'file' scheme.
1814 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1815 then adjusted to work on FreeBSD 8.4 as well.
1816 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1817 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1821 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1822 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1823 error code calls like this:
1825 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1827 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1828 that can be encoded in C. For the foreseeable future, this will only
1831 *Richard Levitte and Tim Hudson*
1833 * Removed BSD cryptodev engine.
1837 * Add a build target 'build_all_generated', to build all generated files
1838 and only that. This can be used to prepare everything that requires
1839 things like perl for a system that lacks perl and then move everything
1840 to that system and do the rest of the build there.
1844 * In the UI interface, make it possible to duplicate the user data. This
1845 can be used by engines that need to retain the data for a longer time
1846 than just the call where this user data is passed.
1850 * Ignore the '-named_curve auto' value for compatibility of applications
1853 *Tomas Mraz <tmraz@fedoraproject.org>*
1855 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1856 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1857 alerts across multiple records (some of which could be empty). In practice
1858 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1859 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1860 support this at all. Supporting it adds significant complexity to the
1861 record layer, and its removal is unlikely to cause interoperability
1866 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1867 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1868 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1873 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1874 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1876 *Richard Levitte, Andy Polyakov*
1878 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1883 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1884 platform rather than 'mingw'.
1888 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1889 success if they are asked to add an object which already exists
1890 in the store. This change cascades to other functions which load
1891 certificates and CRLs.
1895 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1896 facilitate stack unwinding even from assembly subroutines.
1900 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1901 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1905 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1906 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1907 which is the minimum version we support.
1911 * Certificate time validation (X509_cmp_time) enforces stricter
1912 compliance with RFC 5280. Fractional seconds and timezone offsets
1913 are no longer allowed.
1917 * Add support for ARIA
1921 * s_client will now send the Server Name Indication (SNI) extension by
1922 default unless the new "-noservername" option is used. The server name is
1923 based on the host provided to the "-connect" option unless overridden by
1924 using "-servername".
1928 * Add support for SipHash
1932 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1933 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1934 prevent issues where no progress is being made and the peer continually
1935 sends unrecognised record types, using up resources processing them.
1939 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1940 using the algorithm defined in
1941 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1945 * Heartbeat support has been removed; the ABI is changed for now.
1947 *Richard Levitte, Rich Salz*
1949 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1953 * The RSA "null" method, which was partially supported to avoid patent
1954 issues, has been replaced to always returns NULL.
1961 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1963 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1964 used even when parsing explicit parameters, when loading a serialized key
1965 or calling `EC_GROUP_new_from_ecpkparameters()`/
1966 `EC_GROUP_new_from_ecparameters()`.
1967 This prevents bypass of security hardening and performance gains,
1968 especially for curves with specialized EC_METHODs.
1969 By default, if a key encoded with explicit parameters is loaded and later
1970 serialized, the output is still encoded with explicit parameters, even if
1971 internally a "named" EC_GROUP is used for computation.
1975 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1976 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1977 NULL. After this change, only the cofactor parameter can be NULL. It also
1978 does some minimal sanity checks on the passed order.
1983 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1984 An attack is simple, if the first CMS_recipientInfo is valid but the
1985 second CMS_recipientInfo is chosen ciphertext. If the second
1986 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1987 encryption key will be replaced by garbage, and the message cannot be
1988 decoded, but if the RSA decryption fails, the correct encryption key is
1989 used and the recipient will not notice the attack.
1990 As a work around for this potential attack the length of the decrypted
1991 key must be equal to the cipher default key length, in case the
1992 certifiate is not given and all recipientInfo are tried out.
1993 The old behaviour can be re-enabled in the CMS code by setting the
1994 CMS_DEBUG_DECRYPT flag.
1999 * Use Windows installation paths in the mingw builds
2001 Mingw isn't a POSIX environment per se, which means that Windows
2002 paths should be used for installation.
2007 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2009 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2010 This changes the size when using the genpkey app when no size is given. It
2011 fixes an omission in earlier changes that changed all RSA, DSA and DH
2012 generation apps to use 2048 bits by default.
2016 * Prevent over long nonces in ChaCha20-Poly1305.
2018 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2019 for every encryption operation. RFC 7539 specifies that the nonce value
2020 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2021 and front pads the nonce with 0 bytes if it is less than 12
2022 bytes. However it also incorrectly allows a nonce to be set of up to 16
2023 bytes. In this case only the last 12 bytes are significant and any
2024 additional leading bytes are ignored.
2026 It is a requirement of using this cipher that nonce values are
2027 unique. Messages encrypted using a reused nonce value are susceptible to
2028 serious confidentiality and integrity attacks. If an application changes
2029 the default nonce length to be longer than 12 bytes and then makes a
2030 change to the leading bytes of the nonce expecting the new value to be a
2031 new unique nonce then such an application could inadvertently encrypt
2032 messages with a reused nonce.
2034 Additionally the ignored bytes in a long nonce are not covered by the
2035 integrity guarantee of this cipher. Any application that relies on the
2036 integrity of these ignored leading bytes of a long nonce may be further
2037 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2038 is safe because no such use sets such a long nonce value. However user
2039 applications that use this cipher directly and set a non-default nonce
2040 length to be longer than 12 bytes may be vulnerable.
2042 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2048 * Added SCA hardening for modular field inversion in EC_GROUP through
2049 a new dedicated field_inv() pointer in EC_METHOD.
2050 This also addresses a leakage affecting conversions from projective
2051 to affine coordinates.
2053 *Billy Bob Brumley, Nicola Tuveri*
2055 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2056 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2060 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2064 * Remove the 'dist' target and add a tarball building script. The
2065 'dist' target has fallen out of use, and it shouldn't be
2066 necessary to configure just to create a source distribution.
2070 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2072 * Timing vulnerability in DSA signature generation
2074 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2075 timing side channel attack. An attacker could use variations in the signing
2076 algorithm to recover the private key.
2078 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2083 * Timing vulnerability in ECDSA signature generation
2085 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2086 timing side channel attack. An attacker could use variations in the signing
2087 algorithm to recover the private key.
2089 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2094 * Add coordinate blinding for EC_POINT and implement projective
2095 coordinate blinding for generic prime curves as a countermeasure to
2096 chosen point SCA attacks.
2098 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2100 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2102 * Client DoS due to large DH parameter
2104 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2105 malicious server can send a very large prime value to the client. This will
2106 cause the client to spend an unreasonably long period of time generating a
2107 key for this prime resulting in a hang until the client has finished. This
2108 could be exploited in a Denial Of Service attack.
2110 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2115 * Cache timing vulnerability in RSA Key Generation
2117 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2118 a cache timing side channel attack. An attacker with sufficient access to
2119 mount cache timing attacks during the RSA key generation process could
2120 recover the private key.
2122 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2123 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2128 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2129 parameter is no longer accepted, as it leads to a corrupt table. NULL
2130 pem_str is reserved for alias entries only.
2134 * Revert blinding in ECDSA sign and instead make problematic addition
2135 length-invariant. Switch even to fixed-length Montgomery multiplication.
2139 * Change generating and checking of primes so that the error rate of not
2140 being prime depends on the intended use based on the size of the input.
2141 For larger primes this will result in more rounds of Miller-Rabin.
2142 The maximal error rate for primes with more than 1080 bits is lowered
2145 *Kurt Roeckx, Annie Yousar*
2147 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2151 * Add blinding to ECDSA and DSA signatures to protect against side channel
2152 attacks discovered by Keegan Ryan (NCC Group).
2156 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2157 now allow empty (zero character) pass phrases.
2161 * Certificate time validation (X509_cmp_time) enforces stricter
2162 compliance with RFC 5280. Fractional seconds and timezone offsets
2163 are no longer allowed.
2167 * Fixed a text canonicalisation bug in CMS
2169 Where a CMS detached signature is used with text content the text goes
2170 through a canonicalisation process first prior to signing or verifying a
2171 signature. This process strips trailing space at the end of lines, converts
2172 line terminators to CRLF and removes additional trailing line terminators
2173 at the end of a file. A bug in the canonicalisation process meant that
2174 some characters, such as form-feed, were incorrectly treated as whitespace
2175 and removed. This is contrary to the specification (RFC5485). This fix
2176 could mean that detached text data signed with an earlier version of
2177 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2178 signed with a fixed OpenSSL may fail to verify with an earlier version of
2179 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2180 and use the "-binary" flag (for the "cms" command line application) or set
2181 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2185 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2187 * Constructed ASN.1 types with a recursive definition could exceed the stack
2189 Constructed ASN.1 types with a recursive definition (such as can be found
2190 in PKCS7) could eventually exceed the stack given malicious input with
2191 excessive recursion. This could result in a Denial Of Service attack. There
2192 are no such structures used within SSL/TLS that come from untrusted sources
2193 so this is considered safe.
2195 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2201 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2203 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2204 effectively reduced to only comparing the least significant bit of each
2205 byte. This allows an attacker to forge messages that would be considered as
2206 authenticated in an amount of tries lower than that guaranteed by the
2207 security claims of the scheme. The module can only be compiled by the
2208 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2210 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2216 * Add a build target 'build_all_generated', to build all generated files
2217 and only that. This can be used to prepare everything that requires
2218 things like perl for a system that lacks perl and then move everything
2219 to that system and do the rest of the build there.
2223 * Backport SSL_OP_NO_RENGOTIATION
2225 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2226 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2227 changes this is no longer possible in 1.1.0. Therefore the new
2228 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2229 1.1.0 to provide equivalent functionality.
2231 Note that if an application built against 1.1.0h headers (or above) is run
2232 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2233 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2237 * Removed the OS390-Unix config target. It relied on a script that doesn't
2242 * rsaz_1024_mul_avx2 overflow bug on x86_64
2244 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2245 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2246 Analysis suggests that attacks against RSA and DSA as a result of this
2247 defect would be very difficult to perform and are not believed likely.
2248 Attacks against DH1024 are considered just feasible, because most of the
2249 work necessary to deduce information about a private key may be performed
2250 offline. The amount of resources required for such an attack would be
2251 significant. However, for an attack on TLS to be meaningful, the server
2252 would have to share the DH1024 private key among multiple clients, which is
2253 no longer an option since CVE-2016-0701.
2255 This only affects processors that support the AVX2 but not ADX extensions
2256 like Intel Haswell (4th generation).
2258 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2259 was originally found via the OSS-Fuzz project.
2264 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2266 * bn_sqrx8x_internal carry bug on x86_64
2268 There is a carry propagating bug in the x86_64 Montgomery squaring
2269 procedure. No EC algorithms are affected. Analysis suggests that attacks
2270 against RSA and DSA as a result of this defect would be very difficult to
2271 perform and are not believed likely. Attacks against DH are considered just
2272 feasible (although very difficult) because most of the work necessary to
2273 deduce information about a private key may be performed offline. The amount
2274 of resources required for such an attack would be very significant and
2275 likely only accessible to a limited number of attackers. An attacker would
2276 additionally need online access to an unpatched system using the target
2277 private key in a scenario with persistent DH parameters and a private
2278 key that is shared between multiple clients.
2280 This only affects processors that support the BMI1, BMI2 and ADX extensions
2281 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2283 This issue was reported to OpenSSL by the OSS-Fuzz project.
2288 * Malformed X.509 IPAddressFamily could cause OOB read
2290 If an X.509 certificate has a malformed IPAddressFamily extension,
2291 OpenSSL could do a one-byte buffer overread. The most likely result
2292 would be an erroneous display of the certificate in text format.
2294 This issue was reported to OpenSSL by the OSS-Fuzz project.
2299 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2301 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2302 platform rather than 'mingw'.
2306 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2307 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2308 which is the minimum version we support.
2312 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2314 * Encrypt-Then-Mac renegotiation crash
2316 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2317 negotiated where it was not in the original handshake (or vice-versa) then
2318 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2319 and servers are affected.
2321 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2326 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2328 * Truncated packet could crash via OOB read
2330 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2331 cipher is being used, then a truncated packet can cause that host to
2332 perform an out-of-bounds read, usually resulting in a crash.
2334 This issue was reported to OpenSSL by Robert Święcki of Google.
2339 * Bad (EC)DHE parameters cause a client crash
2341 If a malicious server supplies bad parameters for a DHE or ECDHE key
2342 exchange then this can result in the client attempting to dereference a
2343 NULL pointer leading to a client crash. This could be exploited in a Denial
2346 This issue was reported to OpenSSL by Guido Vranken.
2351 * BN_mod_exp may produce incorrect results on x86_64
2353 There is a carry propagating bug in the x86_64 Montgomery squaring
2354 procedure. No EC algorithms are affected. Analysis suggests that attacks
2355 against RSA and DSA as a result of this defect would be very difficult to
2356 perform and are not believed likely. Attacks against DH are considered just
2357 feasible (although very difficult) because most of the work necessary to
2358 deduce information about a private key may be performed offline. The amount
2359 of resources required for such an attack would be very significant and
2360 likely only accessible to a limited number of attackers. An attacker would
2361 additionally need online access to an unpatched system using the target
2362 private key in a scenario with persistent DH parameters and a private
2363 key that is shared between multiple clients. For example this can occur by
2364 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2365 similar to CVE-2015-3193 but must be treated as a separate problem.
2367 This issue was reported to OpenSSL by the OSS-Fuzz project.
2372 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2374 * ChaCha20/Poly1305 heap-buffer-overflow
2376 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2377 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2378 crash. This issue is not considered to be exploitable beyond a DoS.
2380 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2385 * CMS Null dereference
2387 Applications parsing invalid CMS structures can crash with a NULL pointer
2388 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2389 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2390 structure callback if an attempt is made to free certain invalid encodings.
2391 Only CHOICE structures using a callback which do not handle NULL value are
2394 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2399 * Montgomery multiplication may produce incorrect results
2401 There is a carry propagating bug in the Broadwell-specific Montgomery
2402 multiplication procedure that handles input lengths divisible by, but
2403 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2404 and DH private keys are impossible. This is because the subroutine in
2405 question is not used in operations with the private key itself and an input
2406 of the attacker's direct choice. Otherwise the bug can manifest itself as
2407 transient authentication and key negotiation failures or reproducible
2408 erroneous outcome of public-key operations with specially crafted input.
2409 Among EC algorithms only Brainpool P-512 curves are affected and one
2410 presumably can attack ECDH key negotiation. Impact was not analyzed in
2411 detail, because pre-requisites for attack are considered unlikely. Namely
2412 multiple clients have to choose the curve in question and the server has to
2413 share the private key among them, neither of which is default behaviour.
2414 Even then only clients that chose the curve will be affected.
2416 This issue was publicly reported as transient failures and was not
2417 initially recognized as a security issue. Thanks to Richard Morgan for
2418 providing reproducible case.
2423 * Removed automatic addition of RPATH in shared libraries and executables,
2424 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2428 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2430 * Fix Use After Free for large message sizes
2432 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2433 message larger than approx 16k is received then the underlying buffer to
2434 store the incoming message is reallocated and moved. Unfortunately a
2435 dangling pointer to the old location is left which results in an attempt to
2436 write to the previously freed location. This is likely to result in a
2437 crash, however it could potentially lead to execution of arbitrary code.
2439 This issue only affects OpenSSL 1.1.0a.
2441 This issue was reported to OpenSSL by Robert Święcki.
2446 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2448 * OCSP Status Request extension unbounded memory growth
2450 A malicious client can send an excessively large OCSP Status Request
2451 extension. If that client continually requests renegotiation, sending a
2452 large OCSP Status Request extension each time, then there will be unbounded
2453 memory growth on the server. This will eventually lead to a Denial Of
2454 Service attack through memory exhaustion. Servers with a default
2455 configuration are vulnerable even if they do not support OCSP. Builds using
2456 the "no-ocsp" build time option are not affected.
2458 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2463 * SSL_peek() hang on empty record
2465 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2466 sends an empty record. This could be exploited by a malicious peer in a
2467 Denial Of Service attack.
2469 This issue was reported to OpenSSL by Alex Gaynor.
2474 * Excessive allocation of memory in tls_get_message_header() and
2475 dtls1_preprocess_fragment()
2477 A (D)TLS message includes 3 bytes for its length in the header for the
2478 message. This would allow for messages up to 16Mb in length. Messages of
2479 this length are excessive and OpenSSL includes a check to ensure that a
2480 peer is sending reasonably sized messages in order to avoid too much memory
2481 being consumed to service a connection. A flaw in the logic of version
2482 1.1.0 means that memory for the message is allocated too early, prior to
2483 the excessive message length check. Due to way memory is allocated in
2484 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2485 to service a connection. This could lead to a Denial of Service through
2486 memory exhaustion. However, the excessive message length check still takes
2487 place, and this would cause the connection to immediately fail. Assuming
2488 that the application calls SSL_free() on the failed connection in a timely
2489 manner then the 21Mb of allocated memory will then be immediately freed
2490 again. Therefore the excessive memory allocation will be transitory in
2491 nature. This then means that there is only a security impact if:
2493 1) The application does not call SSL_free() in a timely manner in the event
2494 that the connection fails
2496 2) The application is working in a constrained environment where there is
2497 very little free memory
2499 3) The attacker initiates multiple connection attempts such that there are
2500 multiple connections in a state where memory has been allocated for the
2501 connection; SSL_free() has not yet been called; and there is insufficient
2502 memory to service the multiple requests.
2504 Except in the instance of (1) above any Denial Of Service is likely to be
2505 transitory because as soon as the connection fails the memory is
2506 subsequently freed again in the SSL_free() call. However there is an
2507 increased risk during this period of application crashes due to the lack of
2508 memory - which would then mean a more serious Denial of Service.
2510 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2511 (CVE-2016-6307 and CVE-2016-6308)
2515 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2516 had to be removed. Primary reason is that vendor assembler can't
2517 assemble our modules with -KPIC flag. As result it, assembly
2518 support, was not even available as option. But its lack means
2519 lack of side-channel resistant code, which is incompatible with
2520 security by todays standards. Fortunately gcc is readily available
2521 prepackaged option, which we firmly point at...
2525 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2527 * Windows command-line tool supports UTF-8 opt-in option for arguments
2528 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2529 (to any value) allows Windows user to access PKCS#12 file generated
2530 with Windows CryptoAPI and protected with non-ASCII password, as well
2531 as files generated under UTF-8 locale on Linux also protected with
2536 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2537 have been disabled by default and removed from DEFAULT, just like RC4.
2538 See the RC4 item below to re-enable both.
2542 * The method for finding the storage location for the Windows RAND seed file
2543 has changed. First we check %RANDFILE%. If that is not set then we check
2544 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2545 all else fails we fall back to C:\.
2549 * The EVP_EncryptUpdate() function has had its return type changed from void
2550 to int. A return of 0 indicates and error while a return of 1 indicates
2555 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2556 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2557 off the constant time implementation for RSA, DSA and DH have been made
2558 no-ops and deprecated.
2562 * Windows RAND implementation was simplified to only get entropy by
2563 calling CryptGenRandom(). Various other RAND-related tickets
2566 *Joseph Wylie Yandle, Rich Salz*
2568 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2569 and `OPENSSL_LH_`, respectively. The old names are available
2570 with API compatibility. They new names are now completely documented.
2574 * Unify TYPE_up_ref(obj) methods signature.
2575 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2576 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2577 int (instead of void) like all others TYPE_up_ref() methods.
2578 So now these methods also check the return value of CRYPTO_atomic_add(),
2579 and the validity of object reference counter.
2581 *fdasilvayy@gmail.com*
2583 * With Windows Visual Studio builds, the .pdb files are installed
2584 alongside the installed libraries and executables. For a static
2585 library installation, ossl_static.pdb is the associate compiler
2586 generated .pdb file to be used when linking programs.
2590 * Remove openssl.spec. Packaging files belong with the packagers.
2594 * Automatic Darwin/OSX configuration has had a refresh, it will now
2595 recognise x86_64 architectures automatically. You can still decide
2596 to build for a different bitness with the environment variable
2597 KERNEL_BITS (can be 32 or 64), for example:
2599 KERNEL_BITS=32 ./config
2603 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2604 256 bit AES and HMAC with SHA256.
2608 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2612 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2616 * To enable users to have their own config files and build file templates,
2617 Configure looks in the directory indicated by the environment variable
2618 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2619 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2620 name and is used as is.
2624 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2625 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2626 X509_CERT_FILE_CTX was removed.
2630 * "shared" builds are now the default. To create only static libraries use
2631 the "no-shared" Configure option.
2635 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2636 All of these option have not worked for some while and are fundamental
2641 * Make various cleanup routines no-ops and mark them as deprecated. Most
2642 global cleanup functions are no longer required because they are handled
2643 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2644 Explicitly de-initing can cause problems (e.g. where a library that uses
2645 OpenSSL de-inits, but an application is still using it). The affected
2646 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2647 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2648 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2649 COMP_zlib_cleanup().
2653 * --strict-warnings no longer enables runtime debugging options
2654 such as REF_DEBUG. Instead, debug options are automatically
2655 enabled with '--debug' builds.
2657 *Andy Polyakov, Emilia Käsper*
2659 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2660 have been moved out of the public header files. New functions for managing
2661 these have been added.
2665 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2666 objects have been moved out of the public header files. New
2667 functions for managing these have been added.
2671 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2672 have been moved out of the public header files. New functions for managing
2673 these have been added.
2677 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2678 moved out of the public header files. New functions for managing these
2683 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2687 * Removed the mk1mf build scripts.
2691 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2692 it is always safe to #include a header now.
2696 * Removed the aged BC-32 config and all its supporting scripts
2700 * Removed support for Ultrix, Netware, and OS/2.
2704 * Add support for HKDF.
2706 *Alessandro Ghedini*
2708 * Add support for blake2b and blake2s
2712 * Added support for "pipelining". Ciphers that have the
2713 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2714 encryptions/decryptions simultaneously. There are currently no built-in
2715 ciphers with this property but the expectation is that engines will be able
2716 to offer it to significantly improve throughput. Support has been extended
2717 into libssl so that multiple records for a single connection can be
2718 processed in one go (for >=TLS 1.1).
2722 * Added the AFALG engine. This is an async capable engine which is able to
2723 offload work to the Linux kernel. In this initial version it only supports
2724 AES128-CBC. The kernel must be version 4.1.0 or greater.
2728 * OpenSSL now uses a new threading API. It is no longer necessary to
2729 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2730 are two supported threading models: pthreads and windows threads. It is
2731 also possible to configure OpenSSL at compile time for "no-threads". The
2732 old threading API should no longer be used. The functions have been
2733 replaced with "no-op" compatibility macros.
2735 *Alessandro Ghedini, Matt Caswell*
2737 * Modify behavior of ALPN to invoke callback after SNI/servername
2738 callback, such that updates to the SSL_CTX affect ALPN.
2742 * Add SSL_CIPHER queries for authentication and key-exchange.
2746 * Changes to the DEFAULT cipherlist:
2747 - Prefer (EC)DHE handshakes over plain RSA.
2748 - Prefer AEAD ciphers over legacy ciphers.
2749 - Prefer ECDSA over RSA when both certificates are available.
2750 - Prefer TLSv1.2 ciphers/PRF.
2751 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2756 * Change the ECC default curve list to be this, in order: x25519,
2757 secp256r1, secp521r1, secp384r1.
2761 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2762 disabled by default. They can be re-enabled using the
2763 enable-weak-ssl-ciphers option to Configure.
2767 * If the server has ALPN configured, but supports no protocols that the
2768 client advertises, send a fatal "no_application_protocol" alert.
2769 This behaviour is SHALL in RFC 7301, though it isn't universally
2770 implemented by other servers.
2774 * Add X25519 support.
2775 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2776 for public and private key encoding using the format documented in
2777 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2778 key generation and key derivation.
2780 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2785 * Deprecate SRP_VBASE_get_by_user.
2786 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2787 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2788 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2789 seed, even if the seed is configured.
2791 Users should use SRP_VBASE_get1_by_user instead. Note that in
2792 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2793 also that even though configuring the SRP seed attempts to hide
2794 invalid usernames by continuing the handshake with fake
2795 credentials, this behaviour is not constant time and no strong
2796 guarantees are made that the handshake is indistinguishable from
2797 that of a valid user.
2801 * Configuration change; it's now possible to build dynamic engines
2802 without having to build shared libraries and vice versa. This
2803 only applies to the engines in engines/, those in crypto/engine/
2804 will always be built into libcrypto (i.e. "static").
2806 Building dynamic engines is enabled by default; to disable, use
2807 the configuration option "disable-dynamic-engine".
2809 The only requirements for building dynamic engines are the
2810 presence of the DSO module and building with position independent
2811 code, so they will also automatically be disabled if configuring
2812 with "disable-dso" or "disable-pic".
2814 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2815 are also taken away from openssl/opensslconf.h, as they are
2820 * Configuration change; if there is a known flag to compile
2821 position independent code, it will always be applied on the
2822 libcrypto and libssl object files, and never on the application
2823 object files. This means other libraries that use routines from
2824 libcrypto / libssl can be made into shared libraries regardless
2825 of how OpenSSL was configured.
2827 If this isn't desirable, the configuration options "disable-pic"
2828 or "no-pic" can be used to disable the use of PIC. This will
2829 also disable building shared libraries and dynamic engines.
2833 * Removed JPAKE code. It was experimental and has no wide use.
2837 * The INSTALL_PREFIX Makefile variable has been renamed to
2838 DESTDIR. That makes for less confusion on what this variable
2839 is for. Also, the configuration option --install_prefix is
2844 * Heartbeat for TLS has been removed and is disabled by default
2845 for DTLS; configure with enable-heartbeats. Code that uses the
2846 old #define's might need to be updated.
2848 *Emilia Käsper, Rich Salz*
2850 * Rename REF_CHECK to REF_DEBUG.
2854 * New "unified" build system
2856 The "unified" build system is aimed to be a common system for all
2857 platforms we support. With it comes new support for VMS.
2859 This system builds supports building in a different directory tree
2860 than the source tree. It produces one Makefile (for unix family
2861 or lookalikes), or one descrip.mms (for VMS).
2863 The source of information to make the Makefile / descrip.mms is
2864 small files called 'build.info', holding the necessary
2865 information for each directory with source to compile, and a
2866 template in Configurations, like unix-Makefile.tmpl or
2869 With this change, the library names were also renamed on Windows
2870 and on VMS. They now have names that are closer to the standard
2871 on Unix, and include the major version number, and in certain
2872 cases, the architecture they are built for. See "Notes on shared
2873 libraries" in INSTALL.
2875 We rely heavily on the perl module Text::Template.
2879 * Added support for auto-initialisation and de-initialisation of the library.
2880 OpenSSL no longer requires explicit init or deinit routines to be called,
2881 except in certain circumstances. See the OPENSSL_init_crypto() and
2882 OPENSSL_init_ssl() man pages for further information.
2886 * The arguments to the DTLSv1_listen function have changed. Specifically the
2887 "peer" argument is now expected to be a BIO_ADDR object.
2889 * Rewrite of BIO networking library. The BIO library lacked consistent
2890 support of IPv6, and adding it required some more extensive
2891 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2892 which hold all types of addresses and chains of address information.
2893 It also introduces a new API, with functions like BIO_socket,
2894 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2895 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2896 have been adapted accordingly.
2900 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2905 * CRIME protection: disable compression by default, even if OpenSSL is
2906 compiled with zlib enabled. Applications can still enable compression
2907 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2908 using the SSL_CONF library to configure compression.
2912 * The signature of the session callback configured with
2913 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2914 was explicitly marked as `const unsigned char*` instead of
2919 * Always DPURIFY. Remove the use of uninitialized memory in the
2920 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2924 * Removed many obsolete configuration items, including
2925 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2926 MD2_CHAR, MD2_INT, MD2_LONG
2928 IDEA_SHORT, IDEA_LONG
2929 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2931 *Rich Salz, with advice from Andy Polyakov*
2933 * Many BN internals have been moved to an internal header file.
2935 *Rich Salz with help from Andy Polyakov*
2937 * Configuration and writing out the results from it has changed.
2938 Files such as Makefile include/openssl/opensslconf.h and are now
2939 produced through general templates, such as Makefile.in and
2940 crypto/opensslconf.h.in and some help from the perl module
2943 Also, the center of configuration information is no longer
2944 Makefile. Instead, Configure produces a perl module in
2945 configdata.pm which holds most of the config data (in the hash
2946 table %config), the target data that comes from the target
2947 configuration in one of the `Configurations/*.conf` files (in
2952 * To clarify their intended purposes, the Configure options
2953 --prefix and --openssldir change their semantics, and become more
2954 straightforward and less interdependent.
2956 --prefix shall be used exclusively to give the location INSTALLTOP
2957 where programs, scripts, libraries, include files and manuals are
2958 going to be installed. The default is now /usr/local.
2960 --openssldir shall be used exclusively to give the default
2961 location OPENSSLDIR where certificates, private keys, CRLs are
2962 managed. This is also where the default openssl.cnf gets
2964 If the directory given with this option is a relative path, the
2965 values of both the --prefix value and the --openssldir value will
2966 be combined to become OPENSSLDIR.
2967 The default for --openssldir is INSTALLTOP/ssl.
2969 Anyone who uses --openssldir to specify where OpenSSL is to be
2970 installed MUST change to use --prefix instead.
2974 * The GOST engine was out of date and therefore it has been removed. An up
2975 to date GOST engine is now being maintained in an external repository.
2976 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2977 support for GOST ciphersuites (these are only activated if a GOST engine
2982 * EGD is no longer supported by default; use enable-egd when
2985 *Ben Kaduk and Rich Salz*
2987 * The distribution now has Makefile.in files, which are used to
2988 create Makefile's when Configure is run. *Configure must be run
2989 before trying to build now.*
2993 * The return value for SSL_CIPHER_description() for error conditions
2998 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3000 Obtaining and performing DNSSEC validation of TLSA records is
3001 the application's responsibility. The application provides
3002 the TLSA records of its choice to OpenSSL, and these are then
3003 used to authenticate the peer.
3005 The TLSA records need not even come from DNS. They can, for
3006 example, be used to implement local end-entity certificate or
3007 trust-anchor "pinning", where the "pin" data takes the form
3008 of TLSA records, which can augment or replace verification
3009 based on the usual WebPKI public certification authorities.
3013 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3014 continues to support deprecated interfaces in default builds.
3015 However, applications are strongly advised to compile their
3016 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3017 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3018 or the 1.1.0 releases.
3020 In environments in which all applications have been ported to
3021 not use any deprecated interfaces OpenSSL's Configure script
3022 should be used with the --api=1.1.0 option to entirely remove
3023 support for the deprecated features from the library and
3024 unconditionally disable them in the installed headers.
3025 Essentially the same effect can be achieved with the "no-deprecated"
3026 argument to Configure, except that this will always restrict
3027 the build to just the latest API, rather than a fixed API
3030 As applications are ported to future revisions of the API,
3031 they should update their compile-time OPENSSL_API_COMPAT define
3032 accordingly, but in most cases should be able to continue to
3033 compile with later releases.
3035 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3036 0x10000000L and 0x00908000L, respectively. However those
3037 versions did not support the OPENSSL_API_COMPAT feature, and
3038 so applications are not typically tested for explicit support
3039 of just the undeprecated features of either release.
3043 * Add support for setting the minimum and maximum supported protocol.
3044 It can bet set via the SSL_set_min_proto_version() and
3045 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3046 MaxProtocol. It's recommended to use the new APIs to disable
3047 protocols instead of disabling individual protocols using
3048 SSL_set_options() or SSL_CONF's Protocol. This change also
3049 removes support for disabling TLS 1.2 in the OpenSSL TLS
3050 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3054 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3058 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3059 and integrates ECDSA and ECDH functionality into EC. Implementations can
3060 now redirect key generation and no longer need to convert to or from
3063 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3064 include the ec.h header file instead.
3068 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3069 ciphers who are no longer supported and drops support the ephemeral RSA key
3070 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3074 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3075 opaque. For HMAC_CTX, the following constructors and destructors
3078 HMAC_CTX *HMAC_CTX_new(void);
3079 void HMAC_CTX_free(HMAC_CTX *ctx);
3081 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3082 destroy such methods has been added. See EVP_MD_meth_new(3) and
3083 EVP_CIPHER_meth_new(3) for documentation.
3086 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3087 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3088 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3089 an already created structure.
3090 2) For consistency with the majority of our object creators and
3091 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3092 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3093 for deprecated builds.
3097 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3098 cryptographic operations to be performed asynchronously as long as an
3099 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3100 further details. Libssl has also had this capability integrated with the
3101 introduction of the new mode SSL_MODE_ASYNC and associated error
3102 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3103 pages. This work was developed in partnership with Intel Corp.
3107 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3108 always enabled now. If you want to disable the support you should
3109 exclude it using the list of supported ciphers. This also means that the
3110 "-no_ecdhe" option has been removed from s_server.
3114 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3115 SSL_{CTX_}set1_curves() which can set a list.
3119 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3120 curve you want to support using SSL_{CTX_}set1_curves().
3124 * State machine rewrite. The state machine code has been significantly
3125 refactored in order to remove much duplication of code and solve issues
3126 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3127 further details). This change does have some associated API changes.
3128 Notably the SSL_state() function has been removed and replaced by
3129 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3130 SSL_set_state() has been removed altogether. The previous handshake states
3131 defined in ssl.h and ssl3.h have also been removed.
3135 * All instances of the string "ssleay" in the public API were replaced
3136 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3137 Some error codes related to internal RSA_eay API's were renamed.
3141 * The demo files in crypto/threads were moved to demo/threads.
3145 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3148 *Matt Caswell, Rich Salz*
3150 * New ASN.1 embed macro.
3152 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3153 structure is not allocated: it is part of the parent. That is instead of
3161 This reduces memory fragmentation and make it impossible to accidentally
3162 set a mandatory field to NULL.
3164 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3165 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3166 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3171 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3175 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3176 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3177 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3178 DES and RC4 ciphersuites.
3182 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3183 This changes the decoding behaviour for some invalid messages,
3184 though the change is mostly in the more lenient direction, and
3185 legacy behaviour is preserved as much as possible.
3189 * Fix no-stdio build.
3190 *David Woodhouse <David.Woodhouse@intel.com> and also*
3191 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3193 * New testing framework
3194 The testing framework has been largely rewritten and is now using
3195 perl and the perl modules Test::Harness and an extended variant of
3196 Test::More called OpenSSL::Test to do its work. All test scripts in
3197 test/ have been rewritten into test recipes, and all direct calls to
3198 executables in test/Makefile have become individual recipes using the
3199 simplified testing OpenSSL::Test::Simple.
3201 For documentation on our testing modules, do:
3203 perldoc test/testlib/OpenSSL/Test/Simple.pm
3204 perldoc test/testlib/OpenSSL/Test.pm
3208 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3209 are used; the latter aborts on memory leaks (usually checked on exit).
3210 Some undocumented "set malloc, etc., hooks" functions were removed
3211 and others were changed. All are now documented.
3215 * In DSA_generate_parameters_ex, if the provided seed is too short,
3218 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3220 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3221 from RFC4279, RFC4785, RFC5487, RFC5489.
3223 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3224 original RSA_PSK patch.
3228 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3229 era flag was never set throughout the codebase (only read). Also removed
3230 SSL3_FLAGS_POP_BUFFER which was only used if
3231 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3235 * Changed the default name options in the "ca", "crl", "req" and "x509"
3236 to be "oneline" instead of "compat".
3240 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3241 not aware of clients that still exhibit this bug, and the workaround
3242 hasn't been working properly for a while.
3246 * The return type of BIO_number_read() and BIO_number_written() as well as
3247 the corresponding num_read and num_write members in the BIO structure has
3248 changed from unsigned long to uint64_t. On platforms where an unsigned
3249 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3254 * Given the pervasive nature of TLS extensions it is inadvisable to run
3255 OpenSSL without support for them. It also means that maintaining
3256 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3257 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3261 * Removed support for the two export grade static DH ciphersuites
3262 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3263 were newly added (along with a number of other static DH ciphersuites) to
3264 1.0.2. However the two export ones have *never* worked since they were
3265 introduced. It seems strange in any case to be adding new export
3266 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3270 * Version negotiation has been rewritten. In particular SSLv23_method(),
3271 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3272 and turned into macros which simply call the new preferred function names
3273 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3274 should use the new names instead. Also as part of this change the ssl23.h
3275 header file has been removed.
3279 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3280 code and the associated standard is no longer considered fit-for-purpose.
3284 * RT2547 was closed. When generating a private key, try to make the
3285 output file readable only by the owner. This behavior change might
3286 be noticeable when interacting with other software.
3288 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3293 * Added HTTP GET support to the ocsp command.
3297 * Changed default digest for the dgst and enc commands from MD5 to
3302 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3306 * Added support for TLS extended master secret from
3307 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3308 initial patch which was a great help during development.
3312 * All libssl internal structures have been removed from the public header
3313 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3314 now redundant). Users should not attempt to access internal structures
3315 directly. Instead they should use the provided API functions.
3319 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3320 Access to deprecated functions can be re-enabled by running config with
3321 "enable-deprecated". In addition applications wishing to use deprecated
3322 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3323 will, by default, disable some transitive includes that previously existed
3324 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3328 * Added support for OCB mode. OpenSSL has been granted a patent license
3329 compatible with the OpenSSL license for use of OCB. Details are available
3330 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3331 for OCB can be removed by calling config with no-ocb.
3335 * SSLv2 support has been removed. It still supports receiving a SSLv2
3336 compatible client hello.
3340 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3341 done while fixing the error code for the key-too-small case.
3343 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3345 * CA.sh has been removed; use CA.pl instead.
3349 * Removed old DES API.
3353 * Remove various unsupported platforms:
3359 Sinix/ReliantUNIX RM400
3364 16-bit platforms such as WIN16
3368 * Clean up OPENSSL_NO_xxx #define's
3369 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3370 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3371 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3372 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3373 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3374 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3375 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3376 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3377 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3378 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3382 * Cleaned up dead code
3383 Remove all but one '#ifdef undef' which is to be looked at.
3387 * Clean up calling of xxx_free routines.
3388 Just like free(), fix most of the xxx_free routines to accept
3389 NULL. Remove the non-null checks from callers. Save much code.
3393 * Add secure heap for storage of private keys (when possible).
3394 Add BIO_s_secmem(), CBIGNUM, etc.
3395 Contributed by Akamai Technologies under our Corporate CLA.
3399 * Experimental support for a new, fast, unbiased prime candidate generator,
3400 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3402 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3404 * New output format NSS in the sess_id command line tool. This allows
3405 exporting the session id and the master key in NSS keylog format.
3407 *Martin Kaiser <martin@kaiser.cx>*
3409 * Harmonize version and its documentation. -f flag is used to display
3412 *mancha <mancha1@zoho.com>*
3414 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3415 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3417 *mancha <mancha1@zoho.com>*
3419 * Fix some double frees. These are not thought to be exploitable.
3421 *mancha <mancha1@zoho.com>*
3423 * A missing bounds check in the handling of the TLS heartbeat extension
3424 can be used to reveal up to 64k of memory to a connected client or
3427 Thanks for Neel Mehta of Google Security for discovering this bug and to
3428 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3429 preparing the fix [CVE-2014-0160][]
3431 *Adam Langley, Bodo Moeller*
3433 * Fix for the attack described in the paper "Recovering OpenSSL
3434 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3435 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3436 <http://eprint.iacr.org/2014/140>
3438 Thanks to Yuval Yarom and Naomi Benger for discovering this
3439 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3441 *Yuval Yarom and Naomi Benger*
3443 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3444 this fixes a limitation in previous versions of OpenSSL.
3448 * Experimental encrypt-then-mac support.
3450 Experimental support for encrypt then mac from
3451 draft-gutmann-tls-encrypt-then-mac-02.txt
3453 To enable it set the appropriate extension number (0x42 for the test
3454 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3456 For non-compliant peers (i.e. just about everything) this should have no
3459 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3463 * Add EVP support for key wrapping algorithms, to avoid problems with
3464 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3465 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3466 algorithms and include tests cases.
3470 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3475 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3476 MGF1 digest and OAEP label.
3480 * Make openssl verify return errors.
3482 *Chris Palmer <palmer@google.com> and Ben Laurie*
3484 * New function ASN1_TIME_diff to calculate the difference between two
3485 ASN1_TIME structures or one structure and the current time.
3489 * Update fips_test_suite to support multiple command line options. New
3490 test to induce all self test errors in sequence and check expected
3495 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3496 sign or verify all in one operation.
3500 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3501 test programs and fips_test_suite. Includes functionality to parse
3502 the minimal script output of fipsalgest.pl directly.
3506 * Add authorisation parameter to FIPS_module_mode_set().
3510 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3514 * Use separate DRBG fields for internal and external flags. New function
3515 FIPS_drbg_health_check() to perform on demand health checking. Add
3516 generation tests to fips_test_suite with reduced health check interval to
3517 demonstrate periodic health checking. Add "nodh" option to
3518 fips_test_suite to skip very slow DH test.
3522 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3527 * More extensive health check for DRBG checking many more failure modes.
3528 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3529 combination: call this in fips_test_suite.
3533 * Add support for canonical generation of DSA parameter 'g'. See
3536 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3537 POST to handle HMAC cases.
3541 * Add functions FIPS_module_version() and FIPS_module_version_text()
3542 to return numerical and string versions of the FIPS module number.
3546 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3547 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3548 outside the validated module in the FIPS capable OpenSSL.
3552 * Minor change to DRBG entropy callback semantics. In some cases
3553 there is no multiple of the block length between min_len and
3554 max_len. Allow the callback to return more than max_len bytes
3555 of entropy but discard any extra: it is the callback's responsibility
3556 to ensure that the extra data discarded does not impact the
3557 requested amount of entropy.
3561 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3562 information in FIPS186-3, SP800-57 and SP800-131A.
3566 * CCM support via EVP. Interface is very similar to GCM case except we
3567 must supply all data in one chunk (i.e. no update, final) and the
3568 message length must be supplied if AAD is used. Add algorithm test
3573 * Initial version of POST overhaul. Add POST callback to allow the status
3574 of POST to be monitored and/or failures induced. Modify fips_test_suite
3575 to use callback. Always run all selftests even if one fails.
3579 * XTS support including algorithm test driver in the fips_gcmtest program.
3580 Note: this does increase the maximum key length from 32 to 64 bytes but
3581 there should be no binary compatibility issues as existing applications
3582 will never use XTS mode.
3586 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3587 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3588 performs algorithm blocking for unapproved PRNG types. Also do not
3589 set PRNG type in FIPS_mode_set(): leave this to the application.
3590 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3591 the standard OpenSSL PRNG: set additional data to a date time vector.
3595 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3596 This shouldn't present any incompatibility problems because applications
3597 shouldn't be using these directly and any that are will need to rethink
3598 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3602 * Extensive self tests and health checking required by SP800-90 DRBG.
3603 Remove strength parameter from FIPS_drbg_instantiate and always
3604 instantiate at maximum supported strength.
3608 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3612 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3616 * New function DH_compute_key_padded() to compute a DH key and pad with
3617 leading zeroes if needed: this complies with SP800-56A et al.
3621 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3622 anything, incomplete, subject to change and largely untested at present.
3626 * Modify fipscanisteronly build option to only build the necessary object
3627 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3631 * Add experimental option FIPSSYMS to give all symbols in
3632 fipscanister.o and FIPS or fips prefix. This will avoid
3633 conflicts with future versions of OpenSSL. Add perl script
3634 util/fipsas.pl to preprocess assembly language source files
3635 and rename any affected symbols.
3639 * Add selftest checks and algorithm block of non-fips algorithms in
3640 FIPS mode. Remove DES2 from selftests.
3644 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3645 return internal method without any ENGINE dependencies. Add new
3646 tiny fips sign and verify functions.
3650 * New build option no-ec2m to disable characteristic 2 code.
3654 * New build option "fipscanisteronly". This only builds fipscanister.o
3655 and (currently) associated fips utilities. Uses the file Makefile.fips
3656 instead of Makefile.org as the prototype.
3660 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3661 Update fips_gcmtest to use IV generator.
3665 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3666 setting output buffer to NULL. The `*Final` function must be
3667 called although it will not retrieve any additional data. The tag
3668 can be set or retrieved with a ctrl. The IV length is by default 12
3669 bytes (96 bits) but can be set to an alternative value. If the IV
3670 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3675 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3676 underlying do_cipher function handles all cipher semantics itself
3677 including padding and finalisation. This is useful if (for example)
3678 an ENGINE cipher handles block padding itself. The behaviour of
3679 do_cipher is subtly changed if this flag is set: the return value
3680 is the number of characters written to the output buffer (zero is
3681 no longer an error code) or a negative error code. Also if the
3682 input buffer is NULL and length 0 finalisation should be performed.
3686 * If a candidate issuer certificate is already part of the constructed
3687 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3691 * Improve forward-security support: add functions
3693 void SSL_CTX_set_not_resumable_session_callback(
3694 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3695 void SSL_set_not_resumable_session_callback(
3696 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3698 for use by SSL/TLS servers; the callback function will be called whenever a
3699 new session is created, and gets to decide whether the session may be
3700 cached to make it resumable (return 0) or not (return 1). (As by the
3701 SSL/TLS protocol specifications, the session_id sent by the server will be
3702 empty to indicate that the session is not resumable; also, the server will
3703 not generate RFC 4507 (RFC 5077) session tickets.)
3705 A simple reasonable callback implementation is to return is_forward_secure.
3706 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3707 by the SSL/TLS server library, indicating whether it can provide forward
3710 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3712 * New -verify_name option in command line utilities to set verification
3717 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3718 Add CMAC pkey methods.
3722 * Experimental renegotiation in s_server -www mode. If the client
3723 browses /reneg connection is renegotiated. If /renegcert it is
3724 renegotiated requesting a certificate.
3728 * Add an "external" session cache for debugging purposes to s_server. This
3729 should help trace issues which normally are only apparent in deployed
3730 multi-process servers.
3734 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3735 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3736 BIO_set_cipher() and some obscure PEM functions were changed so they
3737 can now return an error. The RAND changes required a change to the
3738 RAND_METHOD structure.
3742 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3743 a gcc attribute to warn if the result of a function is ignored. This
3744 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3745 whose return value is often ignored.
3749 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3750 These allow SCTs (signed certificate timestamps) to be requested and
3751 validated when establishing a connection.
3753 *Rob Percival <robpercival@google.com>*
3758 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3760 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3761 used even when parsing explicit parameters, when loading a serialized key
3762 or calling `EC_GROUP_new_from_ecpkparameters()`/
3763 `EC_GROUP_new_from_ecparameters()`.
3764 This prevents bypass of security hardening and performance gains,
3765 especially for curves with specialized EC_METHODs.
3766 By default, if a key encoded with explicit parameters is loaded and later
3767 serialized, the output is still encoded with explicit parameters, even if
3768 internally a "named" EC_GROUP is used for computation.
3772 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3773 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3774 NULL. After this change, only the cofactor parameter can be NULL. It also
3775 does some minimal sanity checks on the passed order.
3780 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3781 An attack is simple, if the first CMS_recipientInfo is valid but the
3782 second CMS_recipientInfo is chosen ciphertext. If the second
3783 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3784 encryption key will be replaced by garbage, and the message cannot be
3785 decoded, but if the RSA decryption fails, the correct encryption key is
3786 used and the recipient will not notice the attack.
3787 As a work around for this potential attack the length of the decrypted
3788 key must be equal to the cipher default key length, in case the
3789 certifiate is not given and all recipientInfo are tried out.
3790 The old behaviour can be re-enabled in the CMS code by setting the
3791 CMS_DEBUG_DECRYPT flag.
3796 * Document issue with installation paths in diverse Windows builds
3798 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3799 binaries and run-time config file.
3804 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3806 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3807 This changes the size when using the genpkey app when no size is given. It
3808 fixes an omission in earlier changes that changed all RSA, DSA and DH
3809 generation apps to use 2048 bits by default.
3813 * Add FIPS support for Android Arm 64-bit
3815 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3816 Module in Version 2.0.10. For some reason, the corresponding target
3817 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3818 built with FIPS support on Android Arm 64-bit. This omission has been
3821 *Matthias St. Pierre*
3823 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3825 * 0-byte record padding oracle
3827 If an application encounters a fatal protocol error and then calls
3828 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3829 then OpenSSL can respond differently to the calling application if a 0 byte
3830 record is received with invalid padding compared to if a 0 byte record is
3831 received with an invalid MAC. If the application then behaves differently
3832 based on that in a way that is detectable to the remote peer, then this
3833 amounts to a padding oracle that could be used to decrypt data.
3835 In order for this to be exploitable "non-stitched" ciphersuites must be in
3836 use. Stitched ciphersuites are optimised implementations of certain
3837 commonly used ciphersuites. Also the application must call SSL_shutdown()
3838 twice even if a protocol error has occurred (applications should not do
3839 this but some do anyway).
3841 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3842 Aviram, with additional investigation by Steven Collison and Andrew
3843 Hourselt. It was reported to OpenSSL on 10th December 2018.
3848 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3852 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3854 * Microarchitecture timing vulnerability in ECC scalar multiplication
3856 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3857 shown to be vulnerable to a microarchitecture timing side channel attack.
3858 An attacker with sufficient access to mount local timing attacks during
3859 ECDSA signature generation could recover the private key.
3861 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3862 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3868 * Timing vulnerability in DSA signature generation
3870 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3871 timing side channel attack. An attacker could use variations in the signing
3872 algorithm to recover the private key.
3874 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3879 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3880 Module, accidentally introduced while backporting security fixes from the
3881 development branch and hindering the use of ECC in FIPS mode.
3885 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3887 * Client DoS due to large DH parameter
3889 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3890 malicious server can send a very large prime value to the client. This will
3891 cause the client to spend an unreasonably long period of time generating a
3892 key for this prime resulting in a hang until the client has finished. This
3893 could be exploited in a Denial Of Service attack.
3895 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3900 * Cache timing vulnerability in RSA Key Generation
3902 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3903 a cache timing side channel attack. An attacker with sufficient access to
3904 mount cache timing attacks during the RSA key generation process could
3905 recover the private key.
3907 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3908 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3913 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3914 parameter is no longer accepted, as it leads to a corrupt table. NULL
3915 pem_str is reserved for alias entries only.
3919 * Revert blinding in ECDSA sign and instead make problematic addition
3920 length-invariant. Switch even to fixed-length Montgomery multiplication.
3924 * Change generating and checking of primes so that the error rate of not
3925 being prime depends on the intended use based on the size of the input.
3926 For larger primes this will result in more rounds of Miller-Rabin.
3927 The maximal error rate for primes with more than 1080 bits is lowered
3930 *Kurt Roeckx, Annie Yousar*
3932 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3936 * Add blinding to ECDSA and DSA signatures to protect against side channel
3937 attacks discovered by Keegan Ryan (NCC Group).
3941 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3942 now allow empty (zero character) pass phrases.
3946 * Certificate time validation (X509_cmp_time) enforces stricter
3947 compliance with RFC 5280. Fractional seconds and timezone offsets
3948 are no longer allowed.
3952 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3954 * Constructed ASN.1 types with a recursive definition could exceed the stack
3956 Constructed ASN.1 types with a recursive definition (such as can be found
3957 in PKCS7) could eventually exceed the stack given malicious input with
3958 excessive recursion. This could result in a Denial Of Service attack. There
3959 are no such structures used within SSL/TLS that come from untrusted sources
3960 so this is considered safe.
3962 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3968 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3970 * Read/write after SSL object in error state
3972 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3973 mechanism. The intent was that if a fatal error occurred during a handshake
3974 then OpenSSL would move into the error state and would immediately fail if
3975 you attempted to continue the handshake. This works as designed for the
3976 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3977 SSL_connect()), however due to a bug it does not work correctly if
3978 SSL_read() or SSL_write() is called directly. In that scenario, if the
3979 handshake fails then a fatal error will be returned in the initial function
3980 call. If SSL_read()/SSL_write() is subsequently called by the application
3981 for the same SSL object then it will succeed and the data is passed without
3982 being decrypted/encrypted directly from the SSL/TLS record layer.
3984 In order to exploit this issue an application bug would have to be present
3985 that resulted in a call to SSL_read()/SSL_write() being issued after having
3986 already received a fatal error.
3988 This issue was reported to OpenSSL by David Benjamin (Google).
3993 * rsaz_1024_mul_avx2 overflow bug on x86_64
3995 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3996 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3997 Analysis suggests that attacks against RSA and DSA as a result of this
3998 defect would be very difficult to perform and are not believed likely.
3999 Attacks against DH1024 are considered just feasible, because most of the
4000 work necessary to deduce information about a private key may be performed
4001 offline. The amount of resources required for such an attack would be
4002 significant. However, for an attack on TLS to be meaningful, the server
4003 would have to share the DH1024 private key among multiple clients, which is
4004 no longer an option since CVE-2016-0701.
4006 This only affects processors that support the AVX2 but not ADX extensions
4007 like Intel Haswell (4th generation).
4009 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4010 was originally found via the OSS-Fuzz project.
4015 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4017 * bn_sqrx8x_internal carry bug on x86_64
4019 There is a carry propagating bug in the x86_64 Montgomery squaring
4020 procedure. No EC algorithms are affected. Analysis suggests that attacks
4021 against RSA and DSA as a result of this defect would be very difficult to
4022 perform and are not believed likely. Attacks against DH are considered just
4023 feasible (although very difficult) because most of the work necessary to
4024 deduce information about a private key may be performed offline. The amount
4025 of resources required for such an attack would be very significant and
4026 likely only accessible to a limited number of attackers. An attacker would
4027 additionally need online access to an unpatched system using the target
4028 private key in a scenario with persistent DH parameters and a private
4029 key that is shared between multiple clients.
4031 This only affects processors that support the BMI1, BMI2 and ADX extensions
4032 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4034 This issue was reported to OpenSSL by the OSS-Fuzz project.
4039 * Malformed X.509 IPAddressFamily could cause OOB read
4041 If an X.509 certificate has a malformed IPAddressFamily extension,
4042 OpenSSL could do a one-byte buffer overread. The most likely result
4043 would be an erroneous display of the certificate in text format.
4045 This issue was reported to OpenSSL by the OSS-Fuzz project.
4050 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4052 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4053 platform rather than 'mingw'.
4057 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4059 * Truncated packet could crash via OOB read
4061 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4062 cipher is being used, then a truncated packet can cause that host to
4063 perform an out-of-bounds read, usually resulting in a crash.
4065 This issue was reported to OpenSSL by Robert Święcki of Google.
4070 * BN_mod_exp may produce incorrect results on x86_64
4072 There is a carry propagating bug in the x86_64 Montgomery squaring
4073 procedure. No EC algorithms are affected. Analysis suggests that attacks
4074 against RSA and DSA as a result of this defect would be very difficult to
4075 perform and are not believed likely. Attacks against DH are considered just
4076 feasible (although very difficult) because most of the work necessary to
4077 deduce information about a private key may be performed offline. The amount
4078 of resources required for such an attack would be very significant and
4079 likely only accessible to a limited number of attackers. An attacker would
4080 additionally need online access to an unpatched system using the target
4081 private key in a scenario with persistent DH parameters and a private
4082 key that is shared between multiple clients. For example this can occur by
4083 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4084 similar to CVE-2015-3193 but must be treated as a separate problem.
4086 This issue was reported to OpenSSL by the OSS-Fuzz project.
4091 * Montgomery multiplication may produce incorrect results
4093 There is a carry propagating bug in the Broadwell-specific Montgomery
4094 multiplication procedure that handles input lengths divisible by, but
4095 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4096 and DH private keys are impossible. This is because the subroutine in
4097 question is not used in operations with the private key itself and an input
4098 of the attacker's direct choice. Otherwise the bug can manifest itself as
4099 transient authentication and key negotiation failures or reproducible
4100 erroneous outcome of public-key operations with specially crafted input.
4101 Among EC algorithms only Brainpool P-512 curves are affected and one
4102 presumably can attack ECDH key negotiation. Impact was not analyzed in
4103 detail, because pre-requisites for attack are considered unlikely. Namely
4104 multiple clients have to choose the curve in question and the server has to
4105 share the private key among them, neither of which is default behaviour.
4106 Even then only clients that chose the curve will be affected.
4108 This issue was publicly reported as transient failures and was not
4109 initially recognized as a security issue. Thanks to Richard Morgan for
4110 providing reproducible case.
4115 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4116 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4117 prevent issues where no progress is being made and the peer continually
4118 sends unrecognised record types, using up resources processing them.
4122 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4124 * Missing CRL sanity check
4126 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4127 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4128 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4130 This issue only affects the OpenSSL 1.0.2i
4135 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4137 * OCSP Status Request extension unbounded memory growth
4139 A malicious client can send an excessively large OCSP Status Request
4140 extension. If that client continually requests renegotiation, sending a
4141 large OCSP Status Request extension each time, then there will be unbounded
4142 memory growth on the server. This will eventually lead to a Denial Of
4143 Service attack through memory exhaustion. Servers with a default
4144 configuration are vulnerable even if they do not support OCSP. Builds using
4145 the "no-ocsp" build time option are not affected.
4147 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4152 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4155 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4161 * OOB write in MDC2_Update()
4163 An overflow can occur in MDC2_Update() either if called directly or
4164 through the EVP_DigestUpdate() function using MDC2. If an attacker
4165 is able to supply very large amounts of input data after a previous
4166 call to EVP_EncryptUpdate() with a partial block then a length check
4167 can overflow resulting in a heap corruption.
4169 The amount of data needed is comparable to SIZE_MAX which is impractical
4172 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4177 * Malformed SHA512 ticket DoS
4179 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4180 DoS attack where a malformed ticket will result in an OOB read which will
4183 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4184 a custom server callback and ticket lookup mechanism.
4186 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4191 * OOB write in BN_bn2dec()
4193 The function BN_bn2dec() does not check the return value of BN_div_word().
4194 This can cause an OOB write if an application uses this function with an
4195 overly large BIGNUM. This could be a problem if an overly large certificate
4196 or CRL is printed out from an untrusted source. TLS is not affected because
4197 record limits will reject an oversized certificate before it is parsed.
4199 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4204 * OOB read in TS_OBJ_print_bio()
4206 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4207 the total length the OID text representation would use and not the amount
4208 of data written. This will result in OOB reads when large OIDs are
4211 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4216 * Pointer arithmetic undefined behaviour
4218 Avoid some undefined pointer arithmetic
4220 A common idiom in the codebase is to check limits in the following manner:
4223 Where "p" points to some malloc'd data of SIZE bytes and
4226 "len" here could be from some externally supplied data (e.g. from a TLS
4229 The rules of C pointer arithmetic are such that "p + len" is only well
4230 defined where len <= SIZE. Therefore the above idiom is actually
4231 undefined behaviour.
4233 For example this could cause problems if some malloc implementation
4234 provides an address for "p" such that "p + len" actually overflows for
4235 values of len that are too big and therefore p + len < limit.
4237 This issue was reported to OpenSSL by Guido Vranken
4242 * Constant time flag not preserved in DSA signing
4244 Operations in the DSA signing algorithm should run in constant time in
4245 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4246 implementation means that a non-constant time codepath is followed for
4247 certain operations. This has been demonstrated through a cache-timing
4248 attack to be sufficient for an attacker to recover the private DSA key.
4250 This issue was reported by César Pereida (Aalto University), Billy Brumley
4251 (Tampere University of Technology), and Yuval Yarom (The University of
4252 Adelaide and NICTA).
4257 * DTLS buffered message DoS
4259 In a DTLS connection where handshake messages are delivered out-of-order
4260 those messages that OpenSSL is not yet ready to process will be buffered
4261 for later use. Under certain circumstances, a flaw in the logic means that
4262 those messages do not get removed from the buffer even though the handshake
4263 has been completed. An attacker could force up to approx. 15 messages to
4264 remain in the buffer when they are no longer required. These messages will
4265 be cleared when the DTLS connection is closed. The default maximum size for
4266 a message is 100k. Therefore the attacker could force an additional 1500k
4267 to be consumed per connection. By opening many simulataneous connections an
4268 attacker could cause a DoS attack through memory exhaustion.
4270 This issue was reported to OpenSSL by Quan Luo.
4275 * DTLS replay protection DoS
4277 A flaw in the DTLS replay attack protection mechanism means that records
4278 that arrive for future epochs update the replay protection "window" before
4279 the MAC for the record has been validated. This could be exploited by an
4280 attacker by sending a record for the next epoch (which does not have to
4281 decrypt or have a valid MAC), with a very large sequence number. This means
4282 that all subsequent legitimate packets are dropped causing a denial of
4283 service for a specific DTLS connection.
4285 This issue was reported to OpenSSL by the OCAP audit team.
4290 * Certificate message OOB reads
4292 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4293 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4294 theoretical DoS risk but this has not been observed in practice on common
4297 The messages affected are client certificate, client certificate request
4298 and server certificate. As a result the attack can only be performed
4299 against a client or a server which enables client authentication.
4301 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4306 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4308 * Prevent padding oracle in AES-NI CBC MAC check
4310 A MITM attacker can use a padding oracle attack to decrypt traffic
4311 when the connection uses an AES CBC cipher and the server support
4314 This issue was introduced as part of the fix for Lucky 13 padding
4315 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4316 constant time by making sure that always the same bytes are read and
4317 compared against either the MAC or padding bytes. But it no longer
4318 checked that there was enough data to have both the MAC and padding
4321 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4326 * Fix EVP_EncodeUpdate overflow
4328 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4329 Base64 encoding of binary data. If an attacker is able to supply very large
4330 amounts of input data then a length check can overflow resulting in a heap
4333 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4334 the `PEM_write_bio*` family of functions. These are mainly used within the
4335 OpenSSL command line applications, so any application which processes data
4336 from an untrusted source and outputs it as a PEM file should be considered
4337 vulnerable to this issue. User applications that call these APIs directly
4338 with large amounts of untrusted data may also be vulnerable.
4340 This issue was reported by Guido Vranken.
4345 * Fix EVP_EncryptUpdate overflow
4347 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4348 is able to supply very large amounts of input data after a previous call to
4349 EVP_EncryptUpdate() with a partial block then a length check can overflow
4350 resulting in a heap corruption. Following an analysis of all OpenSSL
4351 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4352 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4353 the first called function after an EVP_EncryptInit(), and therefore that
4354 specific call must be safe. The second form is where the length passed to
4355 EVP_EncryptUpdate() can be seen from the code to be some small value and
4356 therefore there is no possibility of an overflow. Since all instances are
4357 one of these two forms, it is believed that there can be no overflows in
4358 internal code due to this problem. It should be noted that
4359 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4360 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4361 of these calls have also been analysed too and it is believed there are no
4362 instances in internal usage where an overflow could occur.
4364 This issue was reported by Guido Vranken.
4369 * Prevent ASN.1 BIO excessive memory allocation
4371 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4372 a short invalid encoding can cause allocation of large amounts of memory
4373 potentially consuming excessive resources or exhausting memory.
4375 Any application parsing untrusted data through d2i BIO functions is
4376 affected. The memory based functions such as d2i_X509() are *not* affected.
4377 Since the memory based functions are used by the TLS library, TLS
4378 applications are not affected.
4380 This issue was reported by Brian Carpenter.
4387 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4388 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4389 in arbitrary stack data being returned in the buffer.
4391 This issue was reported by Guido Vranken.
4396 * Modify behavior of ALPN to invoke callback after SNI/servername
4397 callback, such that updates to the SSL_CTX affect ALPN.
4401 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4406 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4407 methods are enabled and ssl2 is disabled the methods return NULL.
4411 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4413 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4414 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4415 provide any "EXPORT" or "LOW" strength ciphers.
4419 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4420 is by default disabled at build-time. Builds that are not configured with
4421 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4422 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4423 will need to explicitly call either of:
4425 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4427 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4429 as appropriate. Even if either of those is used, or the application
4430 explicitly uses the version-specific SSLv2_method() or its client and
4431 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4432 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4433 ciphers, and SSLv2 56-bit DES are no longer available.
4438 * Fix a double-free in DSA code
4440 A double free bug was discovered when OpenSSL parses malformed DSA private
4441 keys and could lead to a DoS attack or memory corruption for applications
4442 that receive DSA private keys from untrusted sources. This scenario is
4445 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4451 * Disable SRP fake user seed to address a server memory leak.
4453 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4455 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4456 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4457 was changed to ignore the "fake user" SRP seed, even if the seed
4460 Users should use SRP_VBASE_get1_by_user instead. Note that in
4461 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4462 also that even though configuring the SRP seed attempts to hide
4463 invalid usernames by continuing the handshake with fake
4464 credentials, this behaviour is not constant time and no strong
4465 guarantees are made that the handshake is indistinguishable from
4466 that of a valid user.
4471 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4473 In the BN_hex2bn function the number of hex digits is calculated using an
4474 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4475 large values of `i` this can result in `bn_expand` not allocating any
4476 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4477 field as NULL leading to a subsequent NULL ptr deref. For very large values
4478 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4479 In this case memory is allocated to the internal BIGNUM data field, but it
4480 is insufficiently sized leading to heap corruption. A similar issue exists
4481 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4482 is ever called by user applications with very large untrusted hex/dec data.
4483 This is anticipated to be a rare occurrence.
4485 All OpenSSL internal usage of these functions use data that is not expected
4486 to be untrusted, e.g. config file data or application command line
4487 arguments. If user developed applications generate config file data based
4488 on untrusted data then it is possible that this could also lead to security
4489 consequences. This is also anticipated to be rare.
4491 This issue was reported to OpenSSL by Guido Vranken.
4496 * Fix memory issues in `BIO_*printf` functions
4498 The internal `fmtstr` function used in processing a "%s" format string in
4499 the `BIO_*printf` functions could overflow while calculating the length of a
4500 string and cause an OOB read when printing very long strings.
4502 Additionally the internal `doapr_outch` function can attempt to write to an
4503 OOB memory location (at an offset from the NULL pointer) in the event of a
4504 memory allocation failure. In 1.0.2 and below this could be caused where
4505 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4506 could be in processing a very long "%s" format string. Memory leaks can
4509 The first issue may mask the second issue dependent on compiler behaviour.
4510 These problems could enable attacks where large amounts of untrusted data
4511 is passed to the `BIO_*printf` functions. If applications use these functions
4512 in this way then they could be vulnerable. OpenSSL itself uses these
4513 functions when printing out human-readable dumps of ASN.1 data. Therefore
4514 applications that print this data could be vulnerable if the data is from
4515 untrusted sources. OpenSSL command line applications could also be
4516 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4517 as command line arguments.
4519 Libssl is not considered directly vulnerable. Additionally certificates etc
4520 received via remote connections via libssl are also unlikely to be able to
4521 trigger these issues because of message size limits enforced within libssl.
4523 This issue was reported to OpenSSL Guido Vranken.
4528 * Side channel attack on modular exponentiation
4530 A side-channel attack was found which makes use of cache-bank conflicts on
4531 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4532 of RSA keys. The ability to exploit this issue is limited as it relies on
4533 an attacker who has control of code in a thread running on the same
4534 hyper-threaded core as the victim thread which is performing decryptions.
4536 This issue was reported to OpenSSL by Yuval Yarom, The University of
4537 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4538 Nadia Heninger, University of Pennsylvania with more information at
4539 <http://cachebleed.info>.
4544 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4545 if no keysize is specified with default_bits. This fixes an
4546 omission in an earlier change that changed all RSA/DSA key generation
4547 apps to use 2048 bits by default.
4551 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4553 * DH small subgroups
4555 Historically OpenSSL only ever generated DH parameters based on "safe"
4556 primes. More recently (in version 1.0.2) support was provided for
4557 generating X9.42 style parameter files such as those required for RFC 5114
4558 support. The primes used in such files may not be "safe". Where an
4559 application is using DH configured with parameters based on primes that are
4560 not "safe" then an attacker could use this fact to find a peer's private
4561 DH exponent. This attack requires that the attacker complete multiple
4562 handshakes in which the peer uses the same private DH exponent. For example
4563 this could be used to discover a TLS server's private DH exponent if it's
4564 reusing the private DH exponent or it's using a static DH ciphersuite.
4566 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4567 TLS. It is not on by default. If the option is not set then the server
4568 reuses the same private DH exponent for the life of the server process and
4569 would be vulnerable to this attack. It is believed that many popular
4570 applications do set this option and would therefore not be at risk.
4572 The fix for this issue adds an additional check where a "q" parameter is
4573 available (as is the case in X9.42 based parameters). This detects the
4574 only known attack, and is the only possible defense for static DH
4575 ciphersuites. This could have some performance impact.
4577 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4578 default and cannot be disabled. This could have some performance impact.
4580 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4585 * SSLv2 doesn't block disabled ciphers
4587 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4588 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4589 been disabled, provided that the SSLv2 protocol was not also disabled via
4592 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4593 and Sebastian Schinzel.
4598 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4600 * BN_mod_exp may produce incorrect results on x86_64
4602 There is a carry propagating bug in the x86_64 Montgomery squaring
4603 procedure. No EC algorithms are affected. Analysis suggests that attacks
4604 against RSA and DSA as a result of this defect would be very difficult to
4605 perform and are not believed likely. Attacks against DH are considered just
4606 feasible (although very difficult) because most of the work necessary to
4607 deduce information about a private key may be performed offline. The amount
4608 of resources required for such an attack would be very significant and
4609 likely only accessible to a limited number of attackers. An attacker would
4610 additionally need online access to an unpatched system using the target
4611 private key in a scenario with persistent DH parameters and a private
4612 key that is shared between multiple clients. For example this can occur by
4613 default in OpenSSL DHE based SSL/TLS ciphersuites.
4615 This issue was reported to OpenSSL by Hanno Böck.
4620 * Certificate verify crash with missing PSS parameter
4622 The signature verification routines will crash with a NULL pointer
4623 dereference if presented with an ASN.1 signature using the RSA PSS
4624 algorithm and absent mask generation function parameter. Since these
4625 routines are used to verify certificate signature algorithms this can be
4626 used to crash any certificate verification operation and exploited in a
4627 DoS attack. Any application which performs certificate verification is
4628 vulnerable including OpenSSL clients and servers which enable client
4631 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4636 * X509_ATTRIBUTE memory leak
4638 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4639 memory. This structure is used by the PKCS#7 and CMS routines so any
4640 application which reads PKCS#7 or CMS data from untrusted sources is
4641 affected. SSL/TLS is not affected.
4643 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4649 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4650 This changes the decoding behaviour for some invalid messages,
4651 though the change is mostly in the more lenient direction, and
4652 legacy behaviour is preserved as much as possible.
4656 * In DSA_generate_parameters_ex, if the provided seed is too short,
4659 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4661 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4663 * Alternate chains certificate forgery
4665 During certificate verification, OpenSSL will attempt to find an
4666 alternative certificate chain if the first attempt to build such a chain
4667 fails. An error in the implementation of this logic can mean that an
4668 attacker could cause certain checks on untrusted certificates to be
4669 bypassed, such as the CA flag, enabling them to use a valid leaf
4670 certificate to act as a CA and "issue" an invalid certificate.
4672 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4677 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4679 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4680 incompatibility in the handling of HMAC. The previous ABI has now been
4685 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4687 * Malformed ECParameters causes infinite loop
4689 When processing an ECParameters structure OpenSSL enters an infinite loop
4690 if the curve specified is over a specially malformed binary polynomial
4693 This can be used to perform denial of service against any
4694 system which processes public keys, certificate requests or
4695 certificates. This includes TLS clients and TLS servers with
4696 client authentication enabled.
4698 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4703 * Exploitable out-of-bounds read in X509_cmp_time
4705 X509_cmp_time does not properly check the length of the ASN1_TIME
4706 string and can read a few bytes out of bounds. In addition,
4707 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4710 An attacker can use this to craft malformed certificates and CRLs of
4711 various sizes and potentially cause a segmentation fault, resulting in
4712 a DoS on applications that verify certificates or CRLs. TLS clients
4713 that verify CRLs are affected. TLS clients and servers with client
4714 authentication enabled may be affected if they use custom verification
4717 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4718 independently by Hanno Böck.
4723 * PKCS7 crash with missing EnvelopedContent
4725 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4726 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4727 with missing content and trigger a NULL pointer dereference on parsing.
4729 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4730 structures from untrusted sources are affected. OpenSSL clients and
4731 servers are not affected.
4733 This issue was reported to OpenSSL by Michal Zalewski (Google).
4738 * CMS verify infinite loop with unknown hash function
4740 When verifying a signedData message the CMS code can enter an infinite loop
4741 if presented with an unknown hash function OID. This can be used to perform
4742 denial of service against any system which verifies signedData messages using
4744 This issue was reported to OpenSSL by Johannes Bauer.
4749 * Race condition handling NewSessionTicket
4751 If a NewSessionTicket is received by a multi-threaded client when attempting to
4752 reuse a previous ticket then a race condition can occur potentially leading to
4753 a double free of the ticket data.
4758 * Only support 256-bit or stronger elliptic curves with the
4759 'ecdh_auto' setting (server) or by default (client). Of supported
4760 curves, prefer P-256 (both).
4764 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4766 * ClientHello sigalgs DoS fix
4768 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4769 invalid signature algorithms extension a NULL pointer dereference will
4770 occur. This can be exploited in a DoS attack against the server.
4772 This issue was was reported to OpenSSL by David Ramos of Stanford
4776 *Stephen Henson and Matt Caswell*
4778 * Multiblock corrupted pointer fix
4780 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4781 feature only applies on 64 bit x86 architecture platforms that support AES
4782 NI instructions. A defect in the implementation of "multiblock" can cause
4783 OpenSSL's internal write buffer to become incorrectly set to NULL when
4784 using non-blocking IO. Typically, when the user application is using a
4785 socket BIO for writing, this will only result in a failed connection.
4786 However if some other BIO is used then it is likely that a segmentation
4787 fault will be triggered, thus enabling a potential DoS attack.
4789 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4794 * Segmentation fault in DTLSv1_listen fix
4796 The DTLSv1_listen function is intended to be stateless and processes the
4797 initial ClientHello from many peers. It is common for user code to loop
4798 over the call to DTLSv1_listen until a valid ClientHello is received with
4799 an associated cookie. A defect in the implementation of DTLSv1_listen means
4800 that state is preserved in the SSL object from one invocation to the next
4801 that can lead to a segmentation fault. Errors processing the initial
4802 ClientHello can trigger this scenario. An example of such an error could be
4803 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4806 This issue was reported to OpenSSL by Per Allansson.
4811 * Segmentation fault in ASN1_TYPE_cmp fix
4813 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4814 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4815 certificate signature algorithm consistency this can be used to crash any
4816 certificate verification operation and exploited in a DoS attack. Any
4817 application which performs certificate verification is vulnerable including
4818 OpenSSL clients and servers which enable client authentication.
4823 * Segmentation fault for invalid PSS parameters fix
4825 The signature verification routines will crash with a NULL pointer
4826 dereference if presented with an ASN.1 signature using the RSA PSS
4827 algorithm and invalid parameters. Since these routines are used to verify
4828 certificate signature algorithms this can be used to crash any
4829 certificate verification operation and exploited in a DoS attack. Any
4830 application which performs certificate verification is vulnerable including
4831 OpenSSL clients and servers which enable client authentication.
4833 This issue was was reported to OpenSSL by Brian Carpenter.
4838 * ASN.1 structure reuse memory corruption fix
4840 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4841 memory corruption via an invalid write. Such reuse is and has been
4842 strongly discouraged and is believed to be rare.
4844 Applications that parse structures containing CHOICE or ANY DEFINED BY
4845 components may be affected. Certificate parsing (d2i_X509 and related
4846 functions) are however not affected. OpenSSL clients and servers are
4852 * PKCS7 NULL pointer dereferences fix
4854 The PKCS#7 parsing code does not handle missing outer ContentInfo
4855 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4856 missing content and trigger a NULL pointer dereference on parsing.
4858 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4859 otherwise parse PKCS#7 structures from untrusted sources are
4860 affected. OpenSSL clients and servers are not affected.
4862 This issue was reported to OpenSSL by Michal Zalewski (Google).
4867 * DoS via reachable assert in SSLv2 servers fix
4869 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4870 servers that both support SSLv2 and enable export cipher suites by sending
4871 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4873 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4874 (OpenSSL development team).
4879 * Empty CKE with client auth and DHE fix
4881 If client auth is used then a server can seg fault in the event of a DHE
4882 ciphersuite being selected and a zero length ClientKeyExchange message
4883 being sent by the client. This could be exploited in a DoS attack.
4888 * Handshake with unseeded PRNG fix
4890 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4891 with an unseeded PRNG. The conditions are:
4892 - The client is on a platform where the PRNG has not been seeded
4893 automatically, and the user has not seeded manually
4894 - A protocol specific client method version has been used (i.e. not
4895 SSL_client_methodv23)
4896 - A ciphersuite is used that does not require additional random data from
4897 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4899 If the handshake succeeds then the client random that has been used will
4900 have been generated from a PRNG with insufficient entropy and therefore the
4901 output may be predictable.
4903 For example using the following command with an unseeded openssl will
4904 succeed on an unpatched platform:
4906 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4911 * Use After Free following d2i_ECPrivatekey error fix
4913 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4914 could cause a use after free condition. This, in turn, could cause a double
4915 free in several private key parsing functions (such as d2i_PrivateKey
4916 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4917 for applications that receive EC private keys from untrusted
4918 sources. This scenario is considered rare.
4920 This issue was discovered by the BoringSSL project and fixed in their
4926 * X509_to_X509_REQ NULL pointer deref fix
4928 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4929 the certificate key is invalid. This function is rarely used in practice.
4931 This issue was discovered by Brian Carpenter.
4936 * Removed the export ciphers from the DEFAULT ciphers
4940 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4942 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4943 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4944 So far those who have to target multiple platforms would compromise
4945 and argue that binary targeting say ARMv5 would still execute on
4946 ARMv8. "Universal" build resolves this compromise by providing
4947 near-optimal performance even on newer platforms.
4951 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4952 (other platforms pending).
4954 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4956 * Add support for the SignedCertificateTimestampList certificate and
4957 OCSP response extensions from RFC6962.
4961 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4962 for corner cases. (Certain input points at infinity could lead to
4963 bogus results, with non-infinity inputs mapped to infinity too.)
4967 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4968 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4969 common cases are optimized and there still is room for further
4970 improvements. Vector Permutation AES for Altivec is also added.
4974 * Add support for little-endian ppc64 Linux target.
4976 *Marcelo Cerri (IBM)*
4978 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4979 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4980 are optimized and there still is room for further improvements.
4981 Both 32- and 64-bit modes are supported.
4983 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4985 * Improved ARMv7 NEON support.
4989 * Support for SPARC Architecture 2011 crypto extensions, first
4990 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4991 SHA256/512, MD5, GHASH and modular exponentiation.
4993 *Andy Polyakov, David Miller*
4995 * Accelerated modular exponentiation for Intel processors, a.k.a.
4998 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5000 * Support for new and upcoming Intel processors, including AVX2,
5001 BMI and SHA ISA extensions. This includes additional "stitched"
5002 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5005 This work was sponsored by Intel Corp.
5009 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5010 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5011 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5015 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5016 this fixes a limitation in previous versions of OpenSSL.
5020 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5021 MGF1 digest and OAEP label.
5025 * Add EVP support for key wrapping algorithms, to avoid problems with
5026 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5027 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5028 algorithms and include tests cases.
5032 * Add functions to allocate and set the fields of an ECDSA_METHOD
5035 *Douglas E. Engert, Steve Henson*
5037 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5038 difference in days and seconds between two tm or ASN1_TIME structures.
5042 * Add -rev test option to s_server to just reverse order of characters
5043 received by client and send back to server. Also prints an abbreviated
5044 summary of the connection parameters.
5048 * New option -brief for s_client and s_server to print out a brief summary
5049 of connection parameters.
5053 * Add callbacks for arbitrary TLS extensions.
5055 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5057 * New option -crl_download in several openssl utilities to download CRLs
5058 from CRLDP extension in certificates.
5062 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5066 * New function X509_CRL_diff to generate a delta CRL from the difference
5067 of two full CRLs. Add support to "crl" utility.
5071 * New functions to set lookup_crls function and to retrieve
5072 X509_STORE from X509_STORE_CTX.
5076 * Print out deprecated issuer and subject unique ID fields in
5081 * Extend OCSP I/O functions so they can be used for simple general purpose
5082 HTTP as well as OCSP. New wrapper function which can be used to download
5083 CRLs using the OCSP API.
5087 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5091 * `SSL_CONF*` functions. These provide a common framework for application
5092 configuration using configuration files or command lines.
5096 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5097 message callback and prints the results. Needs compile time option
5098 "enable-ssl-trace". New options to s_client and s_server to enable
5103 * New ctrl and macro to retrieve supported points extensions.
5104 Print out extension in s_server and s_client.
5108 * New functions to retrieve certificate signature and signature
5113 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5118 * New Suite B modes for TLS code. These use and enforce the requirements
5119 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5120 only use Suite B curves. The Suite B modes can be set by using the
5121 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5125 * New chain verification flags for Suite B levels of security. Check
5126 algorithms are acceptable when flags are set in X509_verify_cert.
5130 * Make tls1_check_chain return a set of flags indicating checks passed
5131 by a certificate chain. Add additional tests to handle client
5132 certificates: checks for matching certificate type and issuer name
5137 * If an attempt is made to use a signature algorithm not in the peer
5138 preference list abort the handshake. If client has no suitable
5139 signature algorithms in response to a certificate request do not
5140 use the certificate.
5144 * If server EC tmp key is not in client preference list abort handshake.
5148 * Add support for certificate stores in CERT structure. This makes it
5149 possible to have different stores per SSL structure or one store in
5150 the parent SSL_CTX. Include distinct stores for certificate chain
5151 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5152 to build and store a certificate chain in CERT structure: returning
5153 an error if the chain cannot be built: this will allow applications
5154 to test if a chain is correctly configured.
5156 Note: if the CERT based stores are not set then the parent SSL_CTX
5157 store is used to retain compatibility with existing behaviour.
5161 * New function ssl_set_client_disabled to set a ciphersuite disabled
5162 mask based on the current session, check mask when sending client
5163 hello and checking the requested ciphersuite.
5167 * New ctrls to retrieve and set certificate types in a certificate
5168 request message. Print out received values in s_client. If certificate
5169 types is not set with custom values set sensible values based on
5170 supported signature algorithms.
5174 * Support for distinct client and server supported signature algorithms.
5178 * Add certificate callback. If set this is called whenever a certificate
5179 is required by client or server. An application can decide which
5180 certificate chain to present based on arbitrary criteria: for example
5181 supported signature algorithms. Add very simple example to s_server.
5182 This fixes many of the problems and restrictions of the existing client
5183 certificate callback: for example you can now clear an existing
5184 certificate and specify the whole chain.
5188 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5189 the certificate can be used for (if anything). Set valid_flags field
5190 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5191 to have similar checks in it.
5193 Add new "cert_flags" field to CERT structure and include a "strict mode".
5194 This enforces some TLS certificate requirements (such as only permitting
5195 certificate signature algorithms contained in the supported algorithms
5196 extension) which some implementations ignore: this option should be used
5197 with caution as it could cause interoperability issues.
5201 * Update and tidy signature algorithm extension processing. Work out
5202 shared signature algorithms based on preferences and peer algorithms
5203 and print them out in s_client and s_server. Abort handshake if no
5204 shared signature algorithms.
5208 * Add new functions to allow customised supported signature algorithms
5209 for SSL and SSL_CTX structures. Add options to s_client and s_server
5214 * New function SSL_certs_clear() to delete all references to certificates
5215 from an SSL structure. Before this once a certificate had been added
5216 it couldn't be removed.
5220 * Integrate hostname, email address and IP address checking with certificate
5221 verification. New verify options supporting checking in openssl utility.
5225 * Fixes and wildcard matching support to hostname and email checking
5226 functions. Add manual page.
5228 *Florian Weimer (Red Hat Product Security Team)*
5230 * New functions to check a hostname email or IP address against a
5231 certificate. Add options x509 utility to print results of checks against
5236 * Fix OCSP checking.
5238 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5240 * Initial experimental support for explicitly trusted non-root CAs.
5241 OpenSSL still tries to build a complete chain to a root but if an
5242 intermediate CA has a trust setting included that is used. The first
5243 setting is used: whether to trust (e.g., -addtrust option to the x509
5248 * Add -trusted_first option which attempts to find certificates in the
5249 trusted store even if an untrusted chain is also supplied.
5253 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5254 platform support for Linux and Android.
5258 * Support for linux-x32, ILP32 environment in x86_64 framework.
5262 * Experimental multi-implementation support for FIPS capable OpenSSL.
5263 When in FIPS mode the approved implementations are used as normal,
5264 when not in FIPS mode the internal unapproved versions are used instead.
5265 This means that the FIPS capable OpenSSL isn't forced to use the
5266 (often lower performance) FIPS implementations outside FIPS mode.
5270 * Transparently support X9.42 DH parameters when calling
5271 PEM_read_bio_DHparameters. This means existing applications can handle
5272 the new parameter format automatically.
5276 * Initial experimental support for X9.42 DH parameter format: mainly
5277 to support use of 'q' parameter for RFC5114 parameters.
5281 * Add DH parameters from RFC5114 including test data to dhtest.
5285 * Support for automatic EC temporary key parameter selection. If enabled
5286 the most preferred EC parameters are automatically used instead of
5287 hardcoded fixed parameters. Now a server just has to call:
5288 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5289 support ECDH and use the most appropriate parameters.
5293 * Enhance and tidy EC curve and point format TLS extension code. Use
5294 static structures instead of allocation if default values are used.
5295 New ctrls to set curves we wish to support and to retrieve shared curves.
5296 Print out shared curves in s_server. New options to s_server and s_client
5297 to set list of supported curves.
5301 * New ctrls to retrieve supported signature algorithms and
5302 supported curve values as an array of NIDs. Extend openssl utility
5303 to print out received values.
5307 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5308 between NIDs and the more common NIST names such as "P-256". Enhance
5309 ecparam utility and ECC method to recognise the NIST names for curves.
5313 * Enhance SSL/TLS certificate chain handling to support different
5314 chains for each certificate instead of one chain in the parent SSL_CTX.
5318 * Support for fixed DH ciphersuite client authentication: where both
5319 server and client use DH certificates with common parameters.
5323 * Support for fixed DH ciphersuites: those requiring DH server
5328 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5330 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5331 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5332 X509_CINF_get_signature were reverted post internal team review.
5337 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5339 * OCSP Status Request extension unbounded memory growth
5341 A malicious client can send an excessively large OCSP Status Request
5342 extension. If that client continually requests renegotiation, sending a
5343 large OCSP Status Request extension each time, then there will be unbounded
5344 memory growth on the server. This will eventually lead to a Denial Of
5345 Service attack through memory exhaustion. Servers with a default
5346 configuration are vulnerable even if they do not support OCSP. Builds using
5347 the "no-ocsp" build time option are not affected.
5349 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5354 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5357 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5363 * OOB write in MDC2_Update()
5365 An overflow can occur in MDC2_Update() either if called directly or
5366 through the EVP_DigestUpdate() function using MDC2. If an attacker
5367 is able to supply very large amounts of input data after a previous
5368 call to EVP_EncryptUpdate() with a partial block then a length check
5369 can overflow resulting in a heap corruption.
5371 The amount of data needed is comparable to SIZE_MAX which is impractical
5374 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5379 * Malformed SHA512 ticket DoS
5381 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5382 DoS attack where a malformed ticket will result in an OOB read which will
5385 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5386 a custom server callback and ticket lookup mechanism.
5388 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5393 * OOB write in BN_bn2dec()
5395 The function BN_bn2dec() does not check the return value of BN_div_word().
5396 This can cause an OOB write if an application uses this function with an
5397 overly large BIGNUM. This could be a problem if an overly large certificate
5398 or CRL is printed out from an untrusted source. TLS is not affected because
5399 record limits will reject an oversized certificate before it is parsed.
5401 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5406 * OOB read in TS_OBJ_print_bio()
5408 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5409 the total length the OID text representation would use and not the amount
5410 of data written. This will result in OOB reads when large OIDs are
5413 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5418 * Pointer arithmetic undefined behaviour
5420 Avoid some undefined pointer arithmetic
5422 A common idiom in the codebase is to check limits in the following manner:
5425 Where "p" points to some malloc'd data of SIZE bytes and
5428 "len" here could be from some externally supplied data (e.g. from a TLS
5431 The rules of C pointer arithmetic are such that "p + len" is only well
5432 defined where len <= SIZE. Therefore the above idiom is actually
5433 undefined behaviour.
5435 For example this could cause problems if some malloc implementation
5436 provides an address for "p" such that "p + len" actually overflows for
5437 values of len that are too big and therefore p + len < limit.
5439 This issue was reported to OpenSSL by Guido Vranken
5444 * Constant time flag not preserved in DSA signing
5446 Operations in the DSA signing algorithm should run in constant time in
5447 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5448 implementation means that a non-constant time codepath is followed for
5449 certain operations. This has been demonstrated through a cache-timing
5450 attack to be sufficient for an attacker to recover the private DSA key.
5452 This issue was reported by César Pereida (Aalto University), Billy Brumley
5453 (Tampere University of Technology), and Yuval Yarom (The University of
5454 Adelaide and NICTA).
5459 * DTLS buffered message DoS
5461 In a DTLS connection where handshake messages are delivered out-of-order
5462 those messages that OpenSSL is not yet ready to process will be buffered
5463 for later use. Under certain circumstances, a flaw in the logic means that
5464 those messages do not get removed from the buffer even though the handshake
5465 has been completed. An attacker could force up to approx. 15 messages to
5466 remain in the buffer when they are no longer required. These messages will
5467 be cleared when the DTLS connection is closed. The default maximum size for
5468 a message is 100k. Therefore the attacker could force an additional 1500k
5469 to be consumed per connection. By opening many simulataneous connections an
5470 attacker could cause a DoS attack through memory exhaustion.
5472 This issue was reported to OpenSSL by Quan Luo.
5477 * DTLS replay protection DoS
5479 A flaw in the DTLS replay attack protection mechanism means that records
5480 that arrive for future epochs update the replay protection "window" before
5481 the MAC for the record has been validated. This could be exploited by an
5482 attacker by sending a record for the next epoch (which does not have to
5483 decrypt or have a valid MAC), with a very large sequence number. This means
5484 that all subsequent legitimate packets are dropped causing a denial of
5485 service for a specific DTLS connection.
5487 This issue was reported to OpenSSL by the OCAP audit team.
5492 * Certificate message OOB reads
5494 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5495 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5496 theoretical DoS risk but this has not been observed in practice on common
5499 The messages affected are client certificate, client certificate request
5500 and server certificate. As a result the attack can only be performed
5501 against a client or a server which enables client authentication.
5503 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5508 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5510 * Prevent padding oracle in AES-NI CBC MAC check
5512 A MITM attacker can use a padding oracle attack to decrypt traffic
5513 when the connection uses an AES CBC cipher and the server support
5516 This issue was introduced as part of the fix for Lucky 13 padding
5517 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5518 constant time by making sure that always the same bytes are read and
5519 compared against either the MAC or padding bytes. But it no longer
5520 checked that there was enough data to have both the MAC and padding
5523 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5528 * Fix EVP_EncodeUpdate overflow
5530 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5531 Base64 encoding of binary data. If an attacker is able to supply very large
5532 amounts of input data then a length check can overflow resulting in a heap
5535 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5536 the `PEM_write_bio*` family of functions. These are mainly used within the
5537 OpenSSL command line applications, so any application which processes data
5538 from an untrusted source and outputs it as a PEM file should be considered
5539 vulnerable to this issue. User applications that call these APIs directly
5540 with large amounts of untrusted data may also be vulnerable.
5542 This issue was reported by Guido Vranken.
5547 * Fix EVP_EncryptUpdate overflow
5549 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5550 is able to supply very large amounts of input data after a previous call to
5551 EVP_EncryptUpdate() with a partial block then a length check can overflow
5552 resulting in a heap corruption. Following an analysis of all OpenSSL
5553 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5554 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5555 the first called function after an EVP_EncryptInit(), and therefore that
5556 specific call must be safe. The second form is where the length passed to
5557 EVP_EncryptUpdate() can be seen from the code to be some small value and
5558 therefore there is no possibility of an overflow. Since all instances are
5559 one of these two forms, it is believed that there can be no overflows in
5560 internal code due to this problem. It should be noted that
5561 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5562 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5563 of these calls have also been analysed too and it is believed there are no
5564 instances in internal usage where an overflow could occur.
5566 This issue was reported by Guido Vranken.
5571 * Prevent ASN.1 BIO excessive memory allocation
5573 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5574 a short invalid encoding can casuse allocation of large amounts of memory
5575 potentially consuming excessive resources or exhausting memory.
5577 Any application parsing untrusted data through d2i BIO functions is
5578 affected. The memory based functions such as d2i_X509() are *not* affected.
5579 Since the memory based functions are used by the TLS library, TLS
5580 applications are not affected.
5582 This issue was reported by Brian Carpenter.
5589 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5590 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5591 in arbitrary stack data being returned in the buffer.
5593 This issue was reported by Guido Vranken.
5598 * Modify behavior of ALPN to invoke callback after SNI/servername
5599 callback, such that updates to the SSL_CTX affect ALPN.
5603 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5608 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5609 methods are enabled and ssl2 is disabled the methods return NULL.
5613 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5615 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5616 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5617 provide any "EXPORT" or "LOW" strength ciphers.
5621 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5622 is by default disabled at build-time. Builds that are not configured with
5623 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5624 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5625 will need to explicitly call either of:
5627 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5629 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5631 as appropriate. Even if either of those is used, or the application
5632 explicitly uses the version-specific SSLv2_method() or its client and
5633 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5634 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5635 ciphers, and SSLv2 56-bit DES are no longer available.
5640 * Fix a double-free in DSA code
5642 A double free bug was discovered when OpenSSL parses malformed DSA private
5643 keys and could lead to a DoS attack or memory corruption for applications
5644 that receive DSA private keys from untrusted sources. This scenario is
5647 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5653 * Disable SRP fake user seed to address a server memory leak.
5655 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5657 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5658 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5659 was changed to ignore the "fake user" SRP seed, even if the seed
5662 Users should use SRP_VBASE_get1_by_user instead. Note that in
5663 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5664 also that even though configuring the SRP seed attempts to hide
5665 invalid usernames by continuing the handshake with fake
5666 credentials, this behaviour is not constant time and no strong
5667 guarantees are made that the handshake is indistinguishable from
5668 that of a valid user.
5673 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5675 In the BN_hex2bn function the number of hex digits is calculated using an
5676 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5677 large values of `i` this can result in `bn_expand` not allocating any
5678 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5679 field as NULL leading to a subsequent NULL ptr deref. For very large values
5680 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5681 In this case memory is allocated to the internal BIGNUM data field, but it
5682 is insufficiently sized leading to heap corruption. A similar issue exists
5683 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5684 is ever called by user applications with very large untrusted hex/dec data.
5685 This is anticipated to be a rare occurrence.
5687 All OpenSSL internal usage of these functions use data that is not expected
5688 to be untrusted, e.g. config file data or application command line
5689 arguments. If user developed applications generate config file data based
5690 on untrusted data then it is possible that this could also lead to security
5691 consequences. This is also anticipated to be rare.
5693 This issue was reported to OpenSSL by Guido Vranken.
5698 * Fix memory issues in `BIO_*printf` functions
5700 The internal `fmtstr` function used in processing a "%s" format string in
5701 the `BIO_*printf` functions could overflow while calculating the length of a
5702 string and cause an OOB read when printing very long strings.
5704 Additionally the internal `doapr_outch` function can attempt to write to an
5705 OOB memory location (at an offset from the NULL pointer) in the event of a
5706 memory allocation failure. In 1.0.2 and below this could be caused where
5707 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5708 could be in processing a very long "%s" format string. Memory leaks can
5711 The first issue may mask the second issue dependent on compiler behaviour.
5712 These problems could enable attacks where large amounts of untrusted data
5713 is passed to the `BIO_*printf` functions. If applications use these functions
5714 in this way then they could be vulnerable. OpenSSL itself uses these
5715 functions when printing out human-readable dumps of ASN.1 data. Therefore
5716 applications that print this data could be vulnerable if the data is from
5717 untrusted sources. OpenSSL command line applications could also be
5718 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5719 as command line arguments.
5721 Libssl is not considered directly vulnerable. Additionally certificates etc
5722 received via remote connections via libssl are also unlikely to be able to
5723 trigger these issues because of message size limits enforced within libssl.
5725 This issue was reported to OpenSSL Guido Vranken.
5730 * Side channel attack on modular exponentiation
5732 A side-channel attack was found which makes use of cache-bank conflicts on
5733 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5734 of RSA keys. The ability to exploit this issue is limited as it relies on
5735 an attacker who has control of code in a thread running on the same
5736 hyper-threaded core as the victim thread which is performing decryptions.
5738 This issue was reported to OpenSSL by Yuval Yarom, The University of
5739 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5740 Nadia Heninger, University of Pennsylvania with more information at
5741 <http://cachebleed.info>.
5746 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5747 if no keysize is specified with default_bits. This fixes an
5748 omission in an earlier change that changed all RSA/DSA key generation
5749 apps to use 2048 bits by default.
5753 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5755 * Protection for DH small subgroup attacks
5757 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5758 switched on by default and cannot be disabled. This could have some
5763 * SSLv2 doesn't block disabled ciphers
5765 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5766 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5767 been disabled, provided that the SSLv2 protocol was not also disabled via
5770 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5771 and Sebastian Schinzel.
5776 * Reject DH handshakes with parameters shorter than 1024 bits.
5780 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5782 * Certificate verify crash with missing PSS parameter
5784 The signature verification routines will crash with a NULL pointer
5785 dereference if presented with an ASN.1 signature using the RSA PSS
5786 algorithm and absent mask generation function parameter. Since these
5787 routines are used to verify certificate signature algorithms this can be
5788 used to crash any certificate verification operation and exploited in a
5789 DoS attack. Any application which performs certificate verification is
5790 vulnerable including OpenSSL clients and servers which enable client
5793 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5798 * X509_ATTRIBUTE memory leak
5800 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5801 memory. This structure is used by the PKCS#7 and CMS routines so any
5802 application which reads PKCS#7 or CMS data from untrusted sources is
5803 affected. SSL/TLS is not affected.
5805 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5811 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5812 This changes the decoding behaviour for some invalid messages,
5813 though the change is mostly in the more lenient direction, and
5814 legacy behaviour is preserved as much as possible.
5818 * In DSA_generate_parameters_ex, if the provided seed is too short,
5819 use a random seed, as already documented.
5821 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5823 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5825 * Alternate chains certificate forgery
5827 During certificate verfification, OpenSSL will attempt to find an
5828 alternative certificate chain if the first attempt to build such a chain
5829 fails. An error in the implementation of this logic can mean that an
5830 attacker could cause certain checks on untrusted certificates to be
5831 bypassed, such as the CA flag, enabling them to use a valid leaf
5832 certificate to act as a CA and "issue" an invalid certificate.
5834 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5840 * Race condition handling PSK identify hint
5842 If PSK identity hints are received by a multi-threaded client then
5843 the values are wrongly updated in the parent SSL_CTX structure. This can
5844 result in a race condition potentially leading to a double free of the
5850 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5852 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5853 incompatibility in the handling of HMAC. The previous ABI has now been
5856 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5858 * Malformed ECParameters causes infinite loop
5860 When processing an ECParameters structure OpenSSL enters an infinite loop
5861 if the curve specified is over a specially malformed binary polynomial
5864 This can be used to perform denial of service against any
5865 system which processes public keys, certificate requests or
5866 certificates. This includes TLS clients and TLS servers with
5867 client authentication enabled.
5869 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5874 * Exploitable out-of-bounds read in X509_cmp_time
5876 X509_cmp_time does not properly check the length of the ASN1_TIME
5877 string and can read a few bytes out of bounds. In addition,
5878 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5881 An attacker can use this to craft malformed certificates and CRLs of
5882 various sizes and potentially cause a segmentation fault, resulting in
5883 a DoS on applications that verify certificates or CRLs. TLS clients
5884 that verify CRLs are affected. TLS clients and servers with client
5885 authentication enabled may be affected if they use custom verification
5888 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5889 independently by Hanno Böck.
5894 * PKCS7 crash with missing EnvelopedContent
5896 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5897 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5898 with missing content and trigger a NULL pointer dereference on parsing.
5900 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5901 structures from untrusted sources are affected. OpenSSL clients and
5902 servers are not affected.
5904 This issue was reported to OpenSSL by Michal Zalewski (Google).
5909 * CMS verify infinite loop with unknown hash function
5911 When verifying a signedData message the CMS code can enter an infinite loop
5912 if presented with an unknown hash function OID. This can be used to perform
5913 denial of service against any system which verifies signedData messages using
5915 This issue was reported to OpenSSL by Johannes Bauer.
5920 * Race condition handling NewSessionTicket
5922 If a NewSessionTicket is received by a multi-threaded client when attempting to
5923 reuse a previous ticket then a race condition can occur potentially leading to
5924 a double free of the ticket data.
5929 * Reject DH handshakes with parameters shorter than 768 bits.
5931 *Kurt Roeckx and Emilia Kasper*
5933 * dhparam: generate 2048-bit parameters by default.
5935 *Kurt Roeckx and Emilia Kasper*
5937 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5939 * Segmentation fault in ASN1_TYPE_cmp fix
5941 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5942 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5943 certificate signature algorithm consistency this can be used to crash any
5944 certificate verification operation and exploited in a DoS attack. Any
5945 application which performs certificate verification is vulnerable including
5946 OpenSSL clients and servers which enable client authentication.
5951 * ASN.1 structure reuse memory corruption fix
5953 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5954 memory corruption via an invalid write. Such reuse is and has been
5955 strongly discouraged and is believed to be rare.
5957 Applications that parse structures containing CHOICE or ANY DEFINED BY
5958 components may be affected. Certificate parsing (d2i_X509 and related
5959 functions) are however not affected. OpenSSL clients and servers are
5965 * PKCS7 NULL pointer dereferences fix
5967 The PKCS#7 parsing code does not handle missing outer ContentInfo
5968 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5969 missing content and trigger a NULL pointer dereference on parsing.
5971 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5972 otherwise parse PKCS#7 structures from untrusted sources are
5973 affected. OpenSSL clients and servers are not affected.
5975 This issue was reported to OpenSSL by Michal Zalewski (Google).
5980 * DoS via reachable assert in SSLv2 servers fix
5982 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5983 servers that both support SSLv2 and enable export cipher suites by sending
5984 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5986 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5987 (OpenSSL development team).
5992 * Use After Free following d2i_ECPrivatekey error fix
5994 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5995 could cause a use after free condition. This, in turn, could cause a double
5996 free in several private key parsing functions (such as d2i_PrivateKey
5997 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5998 for applications that receive EC private keys from untrusted
5999 sources. This scenario is considered rare.
6001 This issue was discovered by the BoringSSL project and fixed in their
6007 * X509_to_X509_REQ NULL pointer deref fix
6009 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6010 the certificate key is invalid. This function is rarely used in practice.
6012 This issue was discovered by Brian Carpenter.
6017 * Removed the export ciphers from the DEFAULT ciphers
6021 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6023 * Build fixes for the Windows and OpenVMS platforms
6025 *Matt Caswell and Richard Levitte*
6027 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6029 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6030 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6031 dereference. This could lead to a Denial Of Service attack. Thanks to
6032 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6037 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6038 dtls1_buffer_record function under certain conditions. In particular this
6039 could occur if an attacker sent repeated DTLS records with the same
6040 sequence number but for the next epoch. The memory leak could be exploited
6041 by an attacker in a Denial of Service attack through memory exhaustion.
6042 Thanks to Chris Mueller for reporting this issue.
6047 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6048 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6049 method would be set to NULL which could later result in a NULL pointer
6050 dereference. Thanks to Frank Schmirler for reporting this issue.
6055 * Abort handshake if server key exchange message is omitted for ephemeral
6058 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6059 reporting this issue.
6064 * Remove non-export ephemeral RSA code on client and server. This code
6065 violated the TLS standard by allowing the use of temporary RSA keys in
6066 non-export ciphersuites and could be used by a server to effectively
6067 downgrade the RSA key length used to a value smaller than the server
6068 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6069 INRIA or reporting this issue.
6074 * Fixed issue where DH client certificates are accepted without verification.
6075 An OpenSSL server will accept a DH certificate for client authentication
6076 without the certificate verify message. This effectively allows a client to
6077 authenticate without the use of a private key. This only affects servers
6078 which trust a client certificate authority which issues certificates
6079 containing DH keys: these are extremely rare and hardly ever encountered.
6080 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6086 * Ensure that the session ID context of an SSL is updated when its
6087 SSL_CTX is updated via SSL_set_SSL_CTX.
6089 The session ID context is typically set from the parent SSL_CTX,
6090 and can vary with the CTX.
6094 * Fix various certificate fingerprint issues.
6096 By using non-DER or invalid encodings outside the signed portion of a
6097 certificate the fingerprint can be changed without breaking the signature.
6098 Although no details of the signed portion of the certificate can be changed
6099 this can cause problems with some applications: e.g. those using the
6100 certificate fingerprint for blacklists.
6102 1. Reject signatures with non zero unused bits.
6104 If the BIT STRING containing the signature has non zero unused bits reject
6105 the signature. All current signature algorithms require zero unused bits.
6107 2. Check certificate algorithm consistency.
6109 Check the AlgorithmIdentifier inside TBS matches the one in the
6110 certificate signature. NB: this will result in signature failure
6111 errors for some broken certificates.
6113 Thanks to Konrad Kraszewski from Google for reporting this issue.
6115 3. Check DSA/ECDSA signatures use DER.
6117 Re-encode DSA/ECDSA signatures and compare with the original received
6118 signature. Return an error if there is a mismatch.
6120 This will reject various cases including garbage after signature
6121 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6122 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6123 (negative or with leading zeroes).
6125 Further analysis was conducted and fixes were developed by Stephen Henson
6126 of the OpenSSL core team.
6132 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6133 results on some platforms, including x86_64. This bug occurs at random
6134 with a very low probability, and is not known to be exploitable in any
6135 way, though its exact impact is difficult to determine. Thanks to Pieter
6136 Wuille (Blockstream) who reported this issue and also suggested an initial
6137 fix. Further analysis was conducted by the OpenSSL development team and
6138 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6139 the OpenSSL core team.
6144 * Do not resume sessions on the server if the negotiated protocol
6145 version does not match the session's version. Resuming with a different
6146 version, while not strictly forbidden by the RFC, is of questionable
6147 sanity and breaks all known clients.
6149 *David Benjamin, Emilia Käsper*
6151 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6152 early CCS messages during renegotiation. (Note that because
6153 renegotiation is encrypted, this early CCS was not exploitable.)
6157 * Tighten client-side session ticket handling during renegotiation:
6158 ensure that the client only accepts a session ticket if the server sends
6159 the extension anew in the ServerHello. Previously, a TLS client would
6160 reuse the old extension state and thus accept a session ticket if one was
6161 announced in the initial ServerHello.
6163 Similarly, ensure that the client requires a session ticket if one
6164 was advertised in the ServerHello. Previously, a TLS client would
6165 ignore a missing NewSessionTicket message.
6169 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6173 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6174 sends a carefully crafted handshake message, to cause OpenSSL to fail
6175 to free up to 64k of memory causing a memory leak. This could be
6176 exploited in a Denial Of Service attack. This issue affects OpenSSL
6177 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6178 whether SRTP is used or configured. Implementations of OpenSSL that
6179 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6181 The fix was developed by the OpenSSL team.
6186 * Session Ticket Memory Leak.
6188 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6189 integrity of that ticket is first verified. In the event of a session
6190 ticket integrity check failing, OpenSSL will fail to free memory
6191 causing a memory leak. By sending a large number of invalid session
6192 tickets an attacker could exploit this issue in a Denial Of Service
6198 * Build option no-ssl3 is incomplete.
6200 When OpenSSL is configured with "no-ssl3" as a build option, servers
6201 could accept and complete a SSL 3.0 handshake, and clients could be
6202 configured to send them.
6205 *Akamai and the OpenSSL team*
6207 * Add support for TLS_FALLBACK_SCSV.
6208 Client applications doing fallback retries should call
6209 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6212 *Adam Langley, Bodo Moeller*
6214 * Add additional DigestInfo checks.
6216 Re-encode DigestInto in DER and check against the original when
6217 verifying RSA signature: this will reject any improperly encoded
6218 DigestInfo structures.
6220 Note: this is a precautionary measure and no attacks are currently known.
6224 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6226 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6227 SRP code can be overrun an internal buffer. Add sanity check that
6228 g, A, B < N to SRP code.
6230 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6231 Group for discovering this issue.
6236 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6237 TLS 1.0 instead of higher protocol versions when the ClientHello message
6238 is badly fragmented. This allows a man-in-the-middle attacker to force a
6239 downgrade to TLS 1.0 even if both the server and the client support a
6240 higher protocol version, by modifying the client's TLS records.
6242 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6243 researching this issue.
6248 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6249 to a denial of service attack. A malicious server can crash the client
6250 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6251 ciphersuite and sending carefully crafted handshake messages.
6253 Thanks to Felix Gröbert (Google) for discovering and researching this
6259 * By sending carefully crafted DTLS packets an attacker could cause openssl
6260 to leak memory. This can be exploited through a Denial of Service attack.
6261 Thanks to Adam Langley for discovering and researching this issue.
6266 * An attacker can force openssl to consume large amounts of memory whilst
6267 processing DTLS handshake messages. This can be exploited through a
6268 Denial of Service attack.
6269 Thanks to Adam Langley for discovering and researching this issue.
6274 * An attacker can force an error condition which causes openssl to crash
6275 whilst processing DTLS packets due to memory being freed twice. This
6276 can be exploited through a Denial of Service attack.
6277 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6283 * If a multithreaded client connects to a malicious server using a resumed
6284 session and the server sends an ec point format extension it could write
6285 up to 255 bytes to freed memory.
6287 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6293 * A malicious server can crash an OpenSSL client with a null pointer
6294 dereference (read) by specifying an SRP ciphersuite even though it was not
6295 properly negotiated with the client. This can be exploited through a
6296 Denial of Service attack.
6298 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6299 discovering and researching this issue.
6304 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6305 X509_name_oneline, X509_name_print_ex et al. to leak some information
6306 from the stack. Applications may be affected if they echo pretty printing
6307 output to the attacker.
6309 Thanks to Ivan Fratric (Google) for discovering this issue.
6312 *Emilia Käsper, and Steve Henson*
6314 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6315 for corner cases. (Certain input points at infinity could lead to
6316 bogus results, with non-infinity inputs mapped to infinity too.)
6320 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6322 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6323 handshake can force the use of weak keying material in OpenSSL
6324 SSL/TLS clients and servers.
6326 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6327 researching this issue. [CVE-2014-0224][]
6329 *KIKUCHI Masashi, Steve Henson*
6331 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6332 OpenSSL DTLS client the code can be made to recurse eventually crashing
6335 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6338 *Imre Rad, Steve Henson*
6340 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6341 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6342 client or server. This is potentially exploitable to run arbitrary
6343 code on a vulnerable client or server.
6345 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6347 *Jüri Aedla, Steve Henson*
6349 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6350 are subject to a denial of service attack.
6352 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6353 this issue. [CVE-2014-3470][]
6355 *Felix Gröbert, Ivan Fratric, Steve Henson*
6357 * Harmonize version and its documentation. -f flag is used to display
6360 *mancha <mancha1@zoho.com>*
6362 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6363 in i2d_ECPrivateKey.
6365 *mancha <mancha1@zoho.com>*
6367 * Fix some double frees. These are not thought to be exploitable.
6369 *mancha <mancha1@zoho.com>*
6371 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6373 * A missing bounds check in the handling of the TLS heartbeat extension
6374 can be used to reveal up to 64k of memory to a connected client or
6377 Thanks for Neel Mehta of Google Security for discovering this bug and to
6378 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6379 preparing the fix [CVE-2014-0160][]
6381 *Adam Langley, Bodo Moeller*
6383 * Fix for the attack described in the paper "Recovering OpenSSL
6384 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6385 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6386 <http://eprint.iacr.org/2014/140>
6388 Thanks to Yuval Yarom and Naomi Benger for discovering this
6389 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6391 *Yuval Yarom and Naomi Benger*
6393 * TLS pad extension: draft-agl-tls-padding-03
6395 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6396 TLS client Hello record length value would otherwise be > 255 and
6397 less that 512 pad with a dummy extension containing zeroes so it
6398 is at least 512 bytes long.
6400 *Adam Langley, Steve Henson*
6402 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6404 * Fix for TLS record tampering bug. A carefully crafted invalid
6405 handshake could crash OpenSSL with a NULL pointer exception.
6406 Thanks to Anton Johansson for reporting this issues.
6409 * Keep original DTLS digest and encryption contexts in retransmission
6410 structures so we can use the previous session parameters if they need
6411 to be resent. [CVE-2013-6450][]
6415 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6416 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6417 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6418 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6419 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6420 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6422 *Rob Stradling, Adam Langley*
6424 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6426 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6427 supporting platforms or when small records were transferred.
6429 *Andy Polyakov, Steve Henson*
6431 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6433 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6435 This addresses the flaw in CBC record processing discovered by
6436 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6437 at: <http://www.isg.rhul.ac.uk/tls/>
6439 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6440 Security Group at Royal Holloway, University of London
6441 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6442 Emilia Käsper for the initial patch.
6445 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6447 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6448 ciphersuites which can be exploited in a denial of service attack.
6449 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6450 and detecting this bug and to Wolfgang Ettlinger
6451 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6456 * Return an error when checking OCSP signatures when key is NULL.
6457 This fixes a DoS attack. [CVE-2013-0166][]
6461 * Make openssl verify return errors.
6463 *Chris Palmer <palmer@google.com> and Ben Laurie*
6465 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6466 the right response is stapled. Also change SSL_get_certificate()
6467 so it returns the certificate actually sent.
6468 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6470 *Rob Stradling <rob.stradling@comodo.com>*
6472 * Fix possible deadlock when decoding public keys.
6476 * Don't use TLS 1.0 record version number in initial client hello
6481 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6483 * Sanity check record length before skipping explicit IV in TLS
6484 1.2, 1.1 and DTLS to fix DoS attack.
6486 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6487 fuzzing as a service testing platform.
6492 * Initialise tkeylen properly when encrypting CMS messages.
6493 Thanks to Solar Designer of Openwall for reporting this issue.
6497 * In FIPS mode don't try to use composite ciphers as they are not
6502 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6504 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6505 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6506 mean any application compiled against OpenSSL 1.0.0 headers setting
6507 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6508 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6509 0x10000000L Any application which was previously compiled against
6510 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6511 will need to be recompiled as a result. Letting be results in
6512 inability to disable specifically TLS 1.1 and in client context,
6513 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6517 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6518 disable just protocol X, but all protocols above X *if* there are
6519 protocols *below* X still enabled. In more practical terms it means
6520 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6521 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6522 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6527 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6529 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6530 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6531 in CRYPTO_realloc_clean.
6533 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6534 issue and to Adam Langley <agl@chromium.org> for fixing it.
6537 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6539 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6543 * Workarounds for some broken servers that "hang" if a client hello
6544 record length exceeds 255 bytes.
6546 1. Do not use record version number > TLS 1.0 in initial client
6547 hello: some (but not all) hanging servers will now work.
6548 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6549 the number of ciphers sent in the client hello. This should be
6550 set to an even number, such as 50, for example by passing:
6551 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6552 Most broken servers should now work.
6553 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6554 TLS 1.2 client support entirely.
6558 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6562 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6564 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6565 STRING form instead of a DigestInfo.
6569 * The format used for MDC2 RSA signatures is inconsistent between EVP
6570 and the RSA_sign/RSA_verify functions. This was made more apparent when
6571 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6572 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6573 the correct format in RSA_verify so both forms transparently work.
6577 * Some servers which support TLS 1.0 can choke if we initially indicate
6578 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6579 encrypted premaster secret. As a workaround use the maximum permitted
6580 client version in client hello, this should keep such servers happy
6581 and still work with previous versions of OpenSSL.
6585 * Add support for TLS/DTLS heartbeats.
6587 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6589 * Add support for SCTP.
6591 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6593 * Improved PRNG seeding for VOS.
6595 *Paul Green <Paul.Green@stratus.com>*
6597 * Extensive assembler packs updates, most notably:
6599 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6600 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6601 - x86_64: bit-sliced AES implementation;
6602 - ARM: NEON support, contemporary platforms optimizations;
6603 - s390x: z196 support;
6604 - `*`: GHASH and GF(2^m) multiplication implementations;
6608 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6609 (removal of unnecessary code)
6611 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6613 * Add TLS key material exporter from RFC 5705.
6617 * Add DTLS-SRTP negotiation from RFC 5764.
6621 * Add Next Protocol Negotiation,
6622 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6623 disabled with a no-npn flag to config or Configure. Code donated
6626 *Adam Langley <agl@google.com> and Ben Laurie*
6628 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6629 NIST-P256, NIST-P521, with constant-time single point multiplication on
6630 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6631 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6632 Code made available under Apache License version 2.0.
6634 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6635 line to include this in your build of OpenSSL, and run "make depend" (or
6636 "make update"). This enables the following EC_METHODs:
6638 EC_GFp_nistp224_method()
6639 EC_GFp_nistp256_method()
6640 EC_GFp_nistp521_method()
6642 EC_GROUP_new_by_curve_name() will automatically use these (while
6643 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6646 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6648 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6649 all platforms. Move ssize_t definition from e_os.h to the public
6650 header file e_os2.h as it now appears in public header file cms.h
6654 * New -sigopt option to the ca, req and x509 utilities. Additional
6655 signature parameters can be passed using this option and in
6660 * Add RSA PSS signing function. This will generate and set the
6661 appropriate AlgorithmIdentifiers for PSS based on those in the
6662 corresponding EVP_MD_CTX structure. No application support yet.
6666 * Support for companion algorithm specific ASN1 signing routines.
6667 New function ASN1_item_sign_ctx() signs a pre-initialised
6668 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6669 the appropriate parameters.
6673 * Add new algorithm specific ASN1 verification initialisation function
6674 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6675 handling will be the same no matter what EVP_PKEY_METHOD is used.
6676 Add a PSS handler to support verification of PSS signatures: checked
6677 against a number of sample certificates.
6681 * Add signature printing for PSS. Add PSS OIDs.
6683 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6685 * Add algorithm specific signature printing. An individual ASN1 method
6686 can now print out signatures instead of the standard hex dump.
6688 More complex signatures (e.g. PSS) can print out more meaningful
6689 information. Include DSA version that prints out the signature
6694 * Password based recipient info support for CMS library: implementing
6699 * Split password based encryption into PBES2 and PBKDF2 functions. This
6700 neatly separates the code into cipher and PBE sections and is required
6701 for some algorithms that split PBES2 into separate pieces (such as
6702 password based CMS).
6706 * Session-handling fixes:
6707 - Fix handling of connections that are resuming with a session ID,
6708 but also support Session Tickets.
6709 - Fix a bug that suppressed issuing of a new ticket if the client
6710 presented a ticket with an expired session.
6711 - Try to set the ticket lifetime hint to something reasonable.
6712 - Make tickets shorter by excluding irrelevant information.
6713 - On the client side, don't ignore renewed tickets.
6715 *Adam Langley, Bodo Moeller (Google)*
6717 * Fix PSK session representation.
6721 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6723 This work was sponsored by Intel.
6727 * Add GCM support to TLS library. Some custom code is needed to split
6728 the IV between the fixed (from PRF) and explicit (from TLS record)
6729 portions. This adds all GCM ciphersuites supported by RFC5288 and
6730 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6731 add a special AESGCM string for GCM only.
6735 * Expand range of ctrls for AES GCM. Permit setting invocation
6736 field on decrypt and retrieval of invocation field only on encrypt.
6740 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6741 As required by RFC5289 these ciphersuites cannot be used if for
6742 versions of TLS earlier than 1.2.
6746 * For FIPS capable OpenSSL interpret a NULL default public key method
6747 as unset and return the appropriate default but do *not* set the default.
6748 This means we can return the appropriate method in applications that
6749 switch between FIPS and non-FIPS modes.
6753 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6754 ENGINE is used then we cannot handle that in the FIPS module so we
6755 keep original code iff non-FIPS operations are allowed.
6759 * Add -attime option to openssl utilities.
6761 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6763 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6767 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6768 FIPS EC methods unconditionally for now.
6772 * New build option no-ec2m to disable characteristic 2 code.
6776 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6777 all cases can be covered as some introduce binary incompatibilities.
6781 * Redirect RSA operations to FIPS module including keygen,
6782 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6786 * Add similar low level API blocking to ciphers.
6790 * Low level digest APIs are not approved in FIPS mode: any attempt
6791 to use these will cause a fatal error. Applications that *really* want
6792 to use them can use the `private_*` version instead.
6796 * Redirect cipher operations to FIPS module for FIPS builds.
6800 * Redirect digest operations to FIPS module for FIPS builds.
6804 * Update build system to add "fips" flag which will link in fipscanister.o
6805 for static and shared library builds embedding a signature if needed.
6809 * Output TLS supported curves in preference order instead of numerical
6810 order. This is currently hardcoded for the highest order curves first.
6811 This should be configurable so applications can judge speed vs strength.
6815 * Add TLS v1.2 server support for client authentication.
6819 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6824 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6825 FIPS modules versions.
6829 * Add TLS v1.2 client side support for client authentication. Keep cache
6830 of handshake records longer as we don't know the hash algorithm to use
6831 until after the certificate request message is received.
6835 * Initial TLS v1.2 client support. Add a default signature algorithms
6836 extension including all the algorithms we support. Parse new signature
6837 format in client key exchange. Relax some ECC signing restrictions for
6838 TLS v1.2 as indicated in RFC5246.
6842 * Add server support for TLS v1.2 signature algorithms extension. Switch
6843 to new signature format when needed using client digest preference.
6844 All server ciphersuites should now work correctly in TLS v1.2. No client
6845 support yet and no support for client certificates.
6849 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6850 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6851 ciphersuites. At present only RSA key exchange ciphersuites work with
6852 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6853 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6854 and version checking.
6858 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6859 with this defined it will not be affected by any changes to ssl internal
6860 structures. Add several utility functions to allow openssl application
6861 to work with OPENSSL_NO_SSL_INTERN defined.
6865 * A long standing patch to add support for SRP from EdelWeb (Peter
6866 Sylvester and Christophe Renou) was integrated.
6867 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6868 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6871 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6875 * Permit abbreviated handshakes when renegotiating using the function
6876 SSL_renegotiate_abbreviated().
6878 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6880 * Add call to ENGINE_register_all_complete() to
6881 ENGINE_load_builtin_engines(), so some implementations get used
6882 automatically instead of needing explicit application support.
6886 * Add support for TLS key exporter as described in RFC5705.
6888 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6890 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6891 a few changes are required:
6893 Add SSL_OP_NO_TLSv1_1 flag.
6894 Add TLSv1_1 methods.
6895 Update version checking logic to handle version 1.1.
6896 Add explicit IV handling (ported from DTLS code).
6897 Add command line options to s_client/s_server.
6904 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6906 * X509_ATTRIBUTE memory leak
6908 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6909 memory. This structure is used by the PKCS#7 and CMS routines so any
6910 application which reads PKCS#7 or CMS data from untrusted sources is
6911 affected. SSL/TLS is not affected.
6913 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6919 * Race condition handling PSK identify hint
6921 If PSK identity hints are received by a multi-threaded client then
6922 the values are wrongly updated in the parent SSL_CTX structure. This can
6923 result in a race condition potentially leading to a double free of the
6929 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6931 * Malformed ECParameters causes infinite loop
6933 When processing an ECParameters structure OpenSSL enters an infinite loop
6934 if the curve specified is over a specially malformed binary polynomial
6937 This can be used to perform denial of service against any
6938 system which processes public keys, certificate requests or
6939 certificates. This includes TLS clients and TLS servers with
6940 client authentication enabled.
6942 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6947 * Exploitable out-of-bounds read in X509_cmp_time
6949 X509_cmp_time does not properly check the length of the ASN1_TIME
6950 string and can read a few bytes out of bounds. In addition,
6951 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6954 An attacker can use this to craft malformed certificates and CRLs of
6955 various sizes and potentially cause a segmentation fault, resulting in
6956 a DoS on applications that verify certificates or CRLs. TLS clients
6957 that verify CRLs are affected. TLS clients and servers with client
6958 authentication enabled may be affected if they use custom verification
6961 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6962 independently by Hanno Böck.
6967 * PKCS7 crash with missing EnvelopedContent
6969 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6970 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6971 with missing content and trigger a NULL pointer dereference on parsing.
6973 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6974 structures from untrusted sources are affected. OpenSSL clients and
6975 servers are not affected.
6977 This issue was reported to OpenSSL by Michal Zalewski (Google).
6982 * CMS verify infinite loop with unknown hash function
6984 When verifying a signedData message the CMS code can enter an infinite loop
6985 if presented with an unknown hash function OID. This can be used to perform
6986 denial of service against any system which verifies signedData messages using
6988 This issue was reported to OpenSSL by Johannes Bauer.
6993 * Race condition handling NewSessionTicket
6995 If a NewSessionTicket is received by a multi-threaded client when attempting to
6996 reuse a previous ticket then a race condition can occur potentially leading to
6997 a double free of the ticket data.
7002 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7004 * Segmentation fault in ASN1_TYPE_cmp fix
7006 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7007 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7008 certificate signature algorithm consistency this can be used to crash any
7009 certificate verification operation and exploited in a DoS attack. Any
7010 application which performs certificate verification is vulnerable including
7011 OpenSSL clients and servers which enable client authentication.
7016 * ASN.1 structure reuse memory corruption fix
7018 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7019 memory corruption via an invalid write. Such reuse is and has been
7020 strongly discouraged and is believed to be rare.
7022 Applications that parse structures containing CHOICE or ANY DEFINED BY
7023 components may be affected. Certificate parsing (d2i_X509 and related
7024 functions) are however not affected. OpenSSL clients and servers are
7030 * PKCS7 NULL pointer dereferences fix
7032 The PKCS#7 parsing code does not handle missing outer ContentInfo
7033 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7034 missing content and trigger a NULL pointer dereference on parsing.
7036 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7037 otherwise parse PKCS#7 structures from untrusted sources are
7038 affected. OpenSSL clients and servers are not affected.
7040 This issue was reported to OpenSSL by Michal Zalewski (Google).
7045 * DoS via reachable assert in SSLv2 servers fix
7047 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7048 servers that both support SSLv2 and enable export cipher suites by sending
7049 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7051 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7052 (OpenSSL development team).
7057 * Use After Free following d2i_ECPrivatekey error fix
7059 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7060 could cause a use after free condition. This, in turn, could cause a double
7061 free in several private key parsing functions (such as d2i_PrivateKey
7062 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7063 for applications that receive EC private keys from untrusted
7064 sources. This scenario is considered rare.
7066 This issue was discovered by the BoringSSL project and fixed in their
7072 * X509_to_X509_REQ NULL pointer deref fix
7074 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7075 the certificate key is invalid. This function is rarely used in practice.
7077 This issue was discovered by Brian Carpenter.
7082 * Removed the export ciphers from the DEFAULT ciphers
7086 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7088 * Build fixes for the Windows and OpenVMS platforms
7090 *Matt Caswell and Richard Levitte*
7092 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7094 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7095 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7096 dereference. This could lead to a Denial Of Service attack. Thanks to
7097 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7102 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7103 dtls1_buffer_record function under certain conditions. In particular this
7104 could occur if an attacker sent repeated DTLS records with the same
7105 sequence number but for the next epoch. The memory leak could be exploited
7106 by an attacker in a Denial of Service attack through memory exhaustion.
7107 Thanks to Chris Mueller for reporting this issue.
7112 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7113 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7114 method would be set to NULL which could later result in a NULL pointer
7115 dereference. Thanks to Frank Schmirler for reporting this issue.
7120 * Abort handshake if server key exchange message is omitted for ephemeral
7123 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7124 reporting this issue.
7129 * Remove non-export ephemeral RSA code on client and server. This code
7130 violated the TLS standard by allowing the use of temporary RSA keys in
7131 non-export ciphersuites and could be used by a server to effectively
7132 downgrade the RSA key length used to a value smaller than the server
7133 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7134 INRIA or reporting this issue.
7139 * Fixed issue where DH client certificates are accepted without verification.
7140 An OpenSSL server will accept a DH certificate for client authentication
7141 without the certificate verify message. This effectively allows a client to
7142 authenticate without the use of a private key. This only affects servers
7143 which trust a client certificate authority which issues certificates
7144 containing DH keys: these are extremely rare and hardly ever encountered.
7145 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7151 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7152 results on some platforms, including x86_64. This bug occurs at random
7153 with a very low probability, and is not known to be exploitable in any
7154 way, though its exact impact is difficult to determine. Thanks to Pieter
7155 Wuille (Blockstream) who reported this issue and also suggested an initial
7156 fix. Further analysis was conducted by the OpenSSL development team and
7157 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7158 the OpenSSL core team.
7163 * Fix various certificate fingerprint issues.
7165 By using non-DER or invalid encodings outside the signed portion of a
7166 certificate the fingerprint can be changed without breaking the signature.
7167 Although no details of the signed portion of the certificate can be changed
7168 this can cause problems with some applications: e.g. those using the
7169 certificate fingerprint for blacklists.
7171 1. Reject signatures with non zero unused bits.
7173 If the BIT STRING containing the signature has non zero unused bits reject
7174 the signature. All current signature algorithms require zero unused bits.
7176 2. Check certificate algorithm consistency.
7178 Check the AlgorithmIdentifier inside TBS matches the one in the
7179 certificate signature. NB: this will result in signature failure
7180 errors for some broken certificates.
7182 Thanks to Konrad Kraszewski from Google for reporting this issue.
7184 3. Check DSA/ECDSA signatures use DER.
7186 Reencode DSA/ECDSA signatures and compare with the original received
7187 signature. Return an error if there is a mismatch.
7189 This will reject various cases including garbage after signature
7190 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7191 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7192 (negative or with leading zeroes).
7194 Further analysis was conducted and fixes were developed by Stephen Henson
7195 of the OpenSSL core team.
7201 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7203 * Session Ticket Memory Leak.
7205 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7206 integrity of that ticket is first verified. In the event of a session
7207 ticket integrity check failing, OpenSSL will fail to free memory
7208 causing a memory leak. By sending a large number of invalid session
7209 tickets an attacker could exploit this issue in a Denial Of Service
7215 * Build option no-ssl3 is incomplete.
7217 When OpenSSL is configured with "no-ssl3" as a build option, servers
7218 could accept and complete a SSL 3.0 handshake, and clients could be
7219 configured to send them.
7222 *Akamai and the OpenSSL team*
7224 * Add support for TLS_FALLBACK_SCSV.
7225 Client applications doing fallback retries should call
7226 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7229 *Adam Langley, Bodo Moeller*
7231 * Add additional DigestInfo checks.
7233 Reencode DigestInto in DER and check against the original when
7234 verifying RSA signature: this will reject any improperly encoded
7235 DigestInfo structures.
7237 Note: this is a precautionary measure and no attacks are currently known.
7241 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7243 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7244 to a denial of service attack. A malicious server can crash the client
7245 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7246 ciphersuite and sending carefully crafted handshake messages.
7248 Thanks to Felix Gröbert (Google) for discovering and researching this
7254 * By sending carefully crafted DTLS packets an attacker could cause openssl
7255 to leak memory. This can be exploited through a Denial of Service attack.
7256 Thanks to Adam Langley for discovering and researching this issue.
7261 * An attacker can force openssl to consume large amounts of memory whilst
7262 processing DTLS handshake messages. This can be exploited through a
7263 Denial of Service attack.
7264 Thanks to Adam Langley for discovering and researching this issue.
7269 * An attacker can force an error condition which causes openssl to crash
7270 whilst processing DTLS packets due to memory being freed twice. This
7271 can be exploited through a Denial of Service attack.
7272 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7278 * If a multithreaded client connects to a malicious server using a resumed
7279 session and the server sends an ec point format extension it could write
7280 up to 255 bytes to freed memory.
7282 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7288 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7289 X509_name_oneline, X509_name_print_ex et al. to leak some information
7290 from the stack. Applications may be affected if they echo pretty printing
7291 output to the attacker.
7293 Thanks to Ivan Fratric (Google) for discovering this issue.
7296 *Emilia Käsper, and Steve Henson*
7298 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7299 for corner cases. (Certain input points at infinity could lead to
7300 bogus results, with non-infinity inputs mapped to infinity too.)
7304 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7306 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7307 handshake can force the use of weak keying material in OpenSSL
7308 SSL/TLS clients and servers.
7310 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7311 researching this issue. [CVE-2014-0224][]
7313 *KIKUCHI Masashi, Steve Henson*
7315 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7316 OpenSSL DTLS client the code can be made to recurse eventually crashing
7319 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7322 *Imre Rad, Steve Henson*
7324 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7325 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7326 client or server. This is potentially exploitable to run arbitrary
7327 code on a vulnerable client or server.
7329 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7331 *Jüri Aedla, Steve Henson*
7333 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7334 are subject to a denial of service attack.
7336 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7337 this issue. [CVE-2014-3470][]
7339 *Felix Gröbert, Ivan Fratric, Steve Henson*
7341 * Harmonize version and its documentation. -f flag is used to display
7344 *mancha <mancha1@zoho.com>*
7346 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7347 in i2d_ECPrivateKey.
7349 *mancha <mancha1@zoho.com>*
7351 * Fix some double frees. These are not thought to be exploitable.
7353 *mancha <mancha1@zoho.com>*
7355 * Fix for the attack described in the paper "Recovering OpenSSL
7356 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7357 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7358 <http://eprint.iacr.org/2014/140>
7360 Thanks to Yuval Yarom and Naomi Benger for discovering this
7361 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7363 *Yuval Yarom and Naomi Benger*
7365 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7367 * Keep original DTLS digest and encryption contexts in retransmission
7368 structures so we can use the previous session parameters if they need
7369 to be resent. [CVE-2013-6450][]
7373 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7374 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7375 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7376 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7377 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7378 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7380 *Rob Stradling, Adam Langley*
7382 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7384 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7386 This addresses the flaw in CBC record processing discovered by
7387 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7388 at: <http://www.isg.rhul.ac.uk/tls/>
7390 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7391 Security Group at Royal Holloway, University of London
7392 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7393 Emilia Käsper for the initial patch.
7396 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7398 * Return an error when checking OCSP signatures when key is NULL.
7399 This fixes a DoS attack. [CVE-2013-0166][]
7403 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7404 the right response is stapled. Also change SSL_get_certificate()
7405 so it returns the certificate actually sent.
7406 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7407 (This is a backport)
7409 *Rob Stradling <rob.stradling@comodo.com>*
7411 * Fix possible deadlock when decoding public keys.
7415 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7417 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7420 * Sanity check record length before skipping explicit IV in DTLS
7423 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7424 fuzzing as a service testing platform.
7429 * Initialise tkeylen properly when encrypting CMS messages.
7430 Thanks to Solar Designer of Openwall for reporting this issue.
7434 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7436 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7437 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7438 in CRYPTO_realloc_clean.
7440 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7441 issue and to Adam Langley <agl@chromium.org> for fixing it.
7444 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7446 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7448 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7449 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7450 content decryption and always return the same error. Note: this attack
7451 needs on average 2^20 messages so it only affects automated senders. The
7452 old behaviour can be re-enabled in the CMS code by setting the
7453 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7454 an MMA defence is not necessary.
7455 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7456 this issue. [CVE-2012-0884][]
7460 * Fix CVE-2011-4619: make sure we really are receiving a
7461 client hello before rejecting multiple SGC restarts. Thanks to
7462 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7466 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7468 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7469 Thanks to Antonio Martin, Enterprise Secure Access Research and
7470 Development, Cisco Systems, Inc. for discovering this bug and
7471 preparing a fix. [CVE-2012-0050][]
7475 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7477 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7478 of the Vaudenay padding oracle attack on CBC mode encryption
7479 which enables an efficient plaintext recovery attack against
7480 the OpenSSL implementation of DTLS. Their attack exploits timing
7481 differences arising during decryption processing. A research
7482 paper describing this attack can be found at:
7483 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7484 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7485 Security Group at Royal Holloway, University of London
7486 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7487 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7488 for preparing the fix. [CVE-2011-4108][]
7490 *Robin Seggelmann, Michael Tuexen*
7492 * Clear bytes used for block padding of SSL 3.0 records.
7495 *Adam Langley (Google)*
7497 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7498 Kadianakis <desnacked@gmail.com> for discovering this issue and
7499 Adam Langley for preparing the fix. [CVE-2011-4619][]
7501 *Adam Langley (Google)*
7503 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7505 *Andrey Kulikov <amdeich@gmail.com>*
7507 * Prevent malformed RFC3779 data triggering an assertion failure.
7508 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7509 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7511 *Rob Austein <sra@hactrn.net>*
7513 * Improved PRNG seeding for VOS.
7515 *Paul Green <Paul.Green@stratus.com>*
7517 * Fix ssl_ciph.c set-up race.
7519 *Adam Langley (Google)*
7521 * Fix spurious failures in ecdsatest.c.
7523 *Emilia Käsper (Google)*
7525 * Fix the BIO_f_buffer() implementation (which was mixing different
7526 interpretations of the `..._len` fields).
7528 *Adam Langley (Google)*
7530 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7531 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7532 threads won't reuse the same blinding coefficients.
7534 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7535 lock to call BN_BLINDING_invert_ex, and avoids one use of
7536 BN_BLINDING_update for each BN_BLINDING structure (previously,
7537 the last update always remained unused).
7539 *Emilia Käsper (Google)*
7541 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7543 *Bob Buckholz (Google)*
7545 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7547 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7548 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7550 *Kaspar Brand <ossl@velox.ch>*
7552 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7553 for multi-threaded use of ECDH. [CVE-2011-3210][]
7555 *Adam Langley (Google)*
7557 * Fix x509_name_ex_d2i memory leak on bad inputs.
7561 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7562 signature public key algorithm by using OID xref utilities instead.
7563 Before this you could only use some ECC ciphersuites with SHA1 only.
7567 * Add protection against ECDSA timing attacks as mentioned in the paper
7568 by Billy Bob Brumley and Nicola Tuveri, see:
7569 <http://eprint.iacr.org/2011/232.pdf>
7571 *Billy Bob Brumley and Nicola Tuveri*
7573 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7575 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7577 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7579 * Fix bug in string printing code: if *any* escaping is enabled we must
7580 escape the escape character (backslash) or the resulting string is
7585 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7587 * Disable code workaround for ancient and obsolete Netscape browsers
7588 and servers: an attacker can use it in a ciphersuite downgrade attack.
7589 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7593 * Fixed J-PAKE implementation error, originally discovered by
7594 Sebastien Martini, further info and confirmation from Stefan
7595 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7599 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7601 * Fix extension code to avoid race conditions which can result in a buffer
7602 overrun vulnerability: resumed sessions must not be modified as they can
7603 be shared by multiple threads. CVE-2010-3864
7607 * Fix WIN32 build system to correctly link an ENGINE directory into
7612 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7614 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7617 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7619 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7621 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7622 context. The operation can be customised via the ctrl mechanism in
7623 case ENGINEs want to include additional functionality.
7627 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7631 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7632 output hashes compatible with older versions of OpenSSL.
7634 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7636 * Fix compression algorithm handling: if resuming a session use the
7637 compression algorithm of the resumed session instead of determining
7638 it from client hello again. Don't allow server to change algorithm.
7642 * Add load_crls() function to apps tidying load_certs() too. Add option
7643 to verify utility to allow additional CRLs to be included.
7647 * Update OCSP request code to permit adding custom headers to the request:
7648 some responders need this.
7652 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7655 *Julia Lawall <julia@diku.dk>*
7657 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7658 needlessly dereferenced structures, used obsolete functions and
7659 didn't handle all updated verify codes correctly.
7663 * Disable MD2 in the default configuration.
7667 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7668 indicate the initial BIO being pushed or popped. This makes it possible
7669 to determine whether the BIO is the one explicitly called or as a result
7670 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7671 it handles reference counts correctly and doesn't zero out the I/O bio
7672 when it is not being explicitly popped. WARNING: applications which
7673 included workarounds for the old buggy behaviour will need to be modified
7674 or they could free up already freed BIOs.
7678 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7679 renaming to all platforms (within the 0.9.8 branch, this was
7680 done conditionally on Netware platforms to avoid a name clash).
7682 *Guenter <lists@gknw.net>*
7684 * Add ECDHE and PSK support to DTLS.
7686 *Michael Tuexen <tuexen@fh-muenster.de>*
7688 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7693 * Add "missing" function EVP_MD_flags() (without this the only way to
7694 retrieve a digest flags is by accessing the structure directly. Update
7695 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7696 or cipher is registered as in the "from" argument. Print out all
7697 registered digests in the dgst usage message instead of manually
7698 attempting to work them out.
7702 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7703 this allows the use of compression and extensions. Change default cipher
7704 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7705 by default unless an application cipher string requests it.
7709 * Alter match criteria in PKCS12_parse(). It used to try to use local
7710 key ids to find matching certificates and keys but some PKCS#12 files
7711 don't follow the (somewhat unwritten) rules and this strategy fails.
7712 Now just gather all certificates together and the first private key
7713 then look for the first certificate that matches the key.
7717 * Support use of registered digest and cipher names for dgst and cipher
7718 commands instead of having to add each one as a special case. So now
7725 openssl dgst -sha256 foo
7727 and this works for ENGINE based algorithms too.
7731 * Update Gost ENGINE to support parameter files.
7733 *Victor B. Wagner <vitus@cryptocom.ru>*
7735 * Support GeneralizedTime in ca utility.
7737 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7739 * Enhance the hash format used for certificate directory links. The new
7740 form uses the canonical encoding (meaning equivalent names will work
7741 even if they aren't identical) and uses SHA1 instead of MD5. This form
7742 is incompatible with the older format and as a result c_rehash should
7743 be used to rebuild symbolic links.
7747 * Make PKCS#8 the default write format for private keys, replacing the
7748 traditional format. This form is standardised, more secure and doesn't
7749 include an implicit MD5 dependency.
7753 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7754 committed to OpenSSL should pass this lot as a minimum.
7758 * Add session ticket override functionality for use by EAP-FAST.
7760 *Jouni Malinen <j@w1.fi>*
7762 * Modify HMAC functions to return a value. Since these can be implemented
7763 in an ENGINE errors can occur.
7767 * Type-checked OBJ_bsearch_ex.
7771 * Type-checked OBJ_bsearch. Also some constification necessitated
7772 by type-checking. Still to come: TXT_DB, bsearch(?),
7773 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7778 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7779 seconds to a tm structure directly, instead of going through OS
7780 specific date routines. This avoids any issues with OS routines such
7781 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7782 and X509_time_adj_ex() to cover the extended range. The existing
7783 X509_time_adj() is still usable and will no longer have any date issues.
7787 * Delta CRL support. New use deltas option which will attempt to locate
7788 and search any appropriate delta CRLs available.
7790 This work was sponsored by Google.
7794 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7795 code and add additional score elements. Validate alternate CRL paths
7796 as part of the CRL checking and indicate a new error "CRL path validation
7797 error" in this case. Applications wanting additional details can use
7798 the verify callback and check the new "parent" field. If this is not
7799 NULL CRL path validation is taking place. Existing applications won't
7800 see this because it requires extended CRL support which is off by
7803 This work was sponsored by Google.
7807 * Support for freshest CRL extension.
7809 This work was sponsored by Google.
7813 * Initial indirect CRL support. Currently only supported in the CRLs
7814 passed directly and not via lookup. Process certificate issuer
7815 CRL entry extension and lookup CRL entries by bother issuer name
7816 and serial number. Check and process CRL issuer entry in IDP extension.
7818 This work was sponsored by Google.
7822 * Add support for distinct certificate and CRL paths. The CRL issuer
7823 certificate is validated separately in this case. Only enabled if
7824 an extended CRL support flag is set: this flag will enable additional
7825 CRL functionality in future.
7827 This work was sponsored by Google.
7831 * Add support for policy mappings extension.
7833 This work was sponsored by Google.
7837 * Fixes to pathlength constraint, self issued certificate handling,
7838 policy processing to align with RFC3280 and PKITS tests.
7840 This work was sponsored by Google.
7844 * Support for name constraints certificate extension. DN, email, DNS
7845 and URI types are currently supported.
7847 This work was sponsored by Google.
7851 * To cater for systems that provide a pointer-based thread ID rather
7852 than numeric, deprecate the current numeric thread ID mechanism and
7853 replace it with a structure and associated callback type. This
7854 mechanism allows a numeric "hash" to be extracted from a thread ID in
7855 either case, and on platforms where pointers are larger than 'long',
7856 mixing is done to help ensure the numeric 'hash' is usable even if it
7857 can't be guaranteed unique. The default mechanism is to use "&errno"
7858 as a pointer-based thread ID to distinguish between threads.
7860 Applications that want to provide their own thread IDs should now use
7861 CRYPTO_THREADID_set_callback() to register a callback that will call
7862 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7864 Note that ERR_remove_state() is now deprecated, because it is tied
7865 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7866 to free the current thread's error state should be replaced by
7867 ERR_remove_thread_state(NULL).
7869 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7870 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7871 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7872 application was previously providing a numeric thread callback that
7873 was inappropriate for distinguishing threads, then uniqueness might
7874 have been obtained with &errno that happened immediately in the
7875 intermediate development versions of OpenSSL; this is no longer the
7876 case, the numeric thread callback will now override the automatic use
7879 *Geoff Thorpe, with help from Bodo Moeller*
7881 * Initial support for different CRL issuing certificates. This covers a
7882 simple case where the self issued certificates in the chain exist and
7883 the real CRL issuer is higher in the existing chain.
7885 This work was sponsored by Google.
7889 * Removed effectively defunct crypto/store from the build.
7893 * Revamp of STACK to provide stronger type-checking. Still to come:
7894 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7895 ASN1_STRING, CONF_VALUE.
7899 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7900 RAM on SSL connections. This option can save about 34k per idle SSL.
7904 * Revamp of LHASH to provide stronger type-checking. Still to come:
7905 STACK, TXT_DB, bsearch, qsort.
7909 * Initial support for Cryptographic Message Syntax (aka CMS) based
7910 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7911 support for data, signedData, compressedData, digestedData and
7912 encryptedData, envelopedData types included. Scripts to check against
7913 RFC4134 examples draft and interop and consistency checks of many
7914 content types and variants.
7918 * Add options to enc utility to support use of zlib compression BIO.
7922 * Extend mk1mf to support importing of options and assembly language
7923 files from Configure script, currently only included in VC-WIN32.
7924 The assembly language rules can now optionally generate the source
7925 files from the associated perl scripts.
7929 * Implement remaining functionality needed to support GOST ciphersuites.
7930 Interop testing has been performed using CryptoPro implementations.
7932 *Victor B. Wagner <vitus@cryptocom.ru>*
7934 * s390x assembler pack.
7938 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7943 * Implement Opaque PRF Input TLS extension as specified in
7944 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7945 official specification yet and no extension type assignment by
7946 IANA exists, this extension (for now) will have to be explicitly
7947 enabled when building OpenSSL by providing the extension number
7948 to use. For example, specify an option
7950 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7952 to the "config" or "Configure" script to enable the extension,
7953 assuming extension number 0x9527 (which is a completely arbitrary
7954 and unofficial assignment based on the MD5 hash of the Internet
7955 Draft). Note that by doing so, you potentially lose
7956 interoperability with other TLS implementations since these might
7957 be using the same extension number for other purposes.
7959 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7960 opaque PRF input value to use in the handshake. This will create
7961 an internal copy of the length-'len' string at 'src', and will
7962 return non-zero for success.
7964 To get more control and flexibility, provide a callback function
7967 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7968 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7972 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7975 Callback function 'cb' will be called in handshakes, and is
7976 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7977 Argument 'arg' is for application purposes (the value as given to
7978 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7979 be provided to the callback function). The callback function
7980 has to return non-zero to report success: usually 1 to use opaque
7981 PRF input just if possible, or 2 to enforce use of the opaque PRF
7982 input. In the latter case, the library will abort the handshake
7983 if opaque PRF input is not successfully negotiated.
7985 Arguments 'peerinput' and 'len' given to the callback function
7986 will always be NULL and 0 in the case of a client. A server will
7987 see the client's opaque PRF input through these variables if
7988 available (NULL and 0 otherwise). Note that if the server
7989 provides an opaque PRF input, the length must be the same as the
7990 length of the client's opaque PRF input.
7992 Note that the callback function will only be called when creating
7993 a new session (session resumption can resume whatever was
7994 previously negotiated), and will not be called in SSL 2.0
7995 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7996 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7997 for applications that need to enforce opaque PRF input.
8001 * Update ssl code to support digests other than SHA1+MD5 for handshake
8004 *Victor B. Wagner <vitus@cryptocom.ru>*
8006 * Add RFC4507 support to OpenSSL. This includes the corrections in
8007 RFC4507bis. The encrypted ticket format is an encrypted encoded
8008 SSL_SESSION structure, that way new session features are automatically
8011 If a client application caches session in an SSL_SESSION structure
8012 support is transparent because tickets are now stored in the encoded
8015 The SSL_CTX structure automatically generates keys for ticket
8016 protection in servers so again support should be possible
8017 with no application modification.
8019 If a client or server wishes to disable RFC4507 support then the option
8020 SSL_OP_NO_TICKET can be set.
8022 Add a TLS extension debugging callback to allow the contents of any client
8023 or server extensions to be examined.
8025 This work was sponsored by Google.
8029 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8030 OpenSSL should now compile cleanly on gcc 4.2
8032 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8034 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8035 support including streaming MAC support: this is required for GOST
8036 ciphersuite support.
8038 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8040 * Add option -stream to use PKCS#7 streaming in smime utility. New
8041 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8042 to output in BER and PEM format.
8046 * Experimental support for use of HMAC via EVP_PKEY interface. This
8047 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8048 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8049 ENGINE support for HMAC keys which are unextractable. New -mac and
8050 -macopt options to dgst utility.
8054 * New option -sigopt to dgst utility. Update dgst to use
8055 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8056 alternative signing parameters such as X9.31 or PSS in the dgst
8061 * Change ssl_cipher_apply_rule(), the internal function that does
8062 the work each time a ciphersuite string requests enabling
8063 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8064 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8065 the order of disabled ciphersuites such that those ciphersuites
8066 that most recently went from enabled to disabled not only stay
8067 in order with respect to each other, but also have higher priority
8068 than other disabled ciphersuites the next time ciphersuites are
8071 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8072 the same ciphersuites as with "HIGH" alone, but in a specific
8073 order where the PSK ciphersuites come first (since they are the
8074 most recently disabled ciphersuites when "HIGH" is parsed).
8076 Also, change ssl_create_cipher_list() (using this new
8077 functionality) such that between otherwise identical
8078 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8083 * Change ssl_create_cipher_list() so that it automatically
8084 arranges the ciphersuites in reasonable order before starting
8085 to process the rule string. Thus, the definition for "DEFAULT"
8086 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8087 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8088 This makes it much easier to arrive at a reasonable default order
8089 in applications for which anonymous ciphers are OK (meaning
8090 that you can't actually use DEFAULT).
8092 *Bodo Moeller; suggested by Victor Duchovni*
8094 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8095 processing) into multiple integers instead of setting
8096 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8097 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8098 (These masks as well as the individual bit definitions are hidden
8099 away into the non-exported interface ssl/ssl_locl.h, so this
8100 change to the definition of the SSL_CIPHER structure shouldn't
8101 affect applications.) This give us more bits for each of these
8102 categories, so there is no longer a need to coagulate AES128 and
8103 AES256 into a single algorithm bit, and to coagulate Camellia128
8104 and Camellia256 into a single algorithm bit, which has led to all
8107 Thus, among other things, the kludge introduced in 0.9.7m and
8108 0.9.8e for masking out AES256 independently of AES128 or masking
8109 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8111 With the change, we also introduce new ciphersuite aliases that
8112 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8117 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8118 Use the leftmost N bytes of the signature input if the input is
8119 larger than the prime q (with N being the size in bytes of q).
8123 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8124 it yet and it is largely untested.
8128 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8132 * Initial incomplete changes to avoid need for function casts in OpenSSL
8133 some compilers (gcc 4.2 and later) reject their use. Safestack is
8134 reimplemented. Update ASN1 to avoid use of legacy functions.
8138 * Win32/64 targets are linked with Winsock2.
8142 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8143 to external functions. This can be used to increase CRL handling
8144 efficiency especially when CRLs are very large by (for example) storing
8145 the CRL revoked certificates in a database.
8149 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8150 new CRLs added to a directory can be used. New command line option
8151 -verify_return_error to s_client and s_server. This causes real errors
8152 to be returned by the verify callback instead of carrying on no matter
8153 what. This reflects the way a "real world" verify callback would behave.
8157 * GOST engine, supporting several GOST algorithms and public key formats.
8158 Kindly donated by Cryptocom.
8162 * Partial support for Issuing Distribution Point CRL extension. CRLs
8163 partitioned by DP are handled but no indirect CRL or reason partitioning
8164 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8165 selected via a scoring technique which handles IDP and AKID in CRLs.
8169 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8170 will ultimately be used for all verify operations: this will remove the
8171 X509_STORE dependency on certificate verification and allow alternative
8172 lookup methods. X509_STORE based implementations of these two callbacks.
8176 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8177 Modify get_crl() to find a valid (unexpired) CRL if possible.
8181 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8182 this would be called X509_CRL_cmp() but that name is already used by
8183 a function that just compares CRL issuer names. Cache several CRL
8184 extensions in X509_CRL structure and cache CRLDP in X509.
8188 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8189 this maps equivalent X509_NAME structures into a consistent structure.
8190 Name comparison can then be performed rapidly using memcmp().
8194 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8199 * Allow digests to supply their own micalg string for S/MIME type using
8200 the ctrl EVP_MD_CTRL_MICALG.
8204 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8205 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8206 ctrl. It can then customise the structure before and/or after signing
8211 * New function OBJ_add_sigid() to allow application defined signature OIDs
8212 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8213 to free up any added signature OIDs.
8217 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8218 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8219 digest and cipher tables. New options added to openssl utility:
8220 list-message-digest-algorithms and list-cipher-algorithms.
8224 * Change the array representation of binary polynomials: the list
8225 of degrees of non-zero coefficients is now terminated with -1.
8226 Previously it was terminated with 0, which was also part of the
8227 value; thus, the array representation was not applicable to
8228 polynomials where t^0 has coefficient zero. This change makes
8229 the array representation useful in a more general context.
8233 * Various modifications and fixes to SSL/TLS cipher string
8234 handling. For ECC, the code now distinguishes between fixed ECDH
8235 with RSA certificates on the one hand and with ECDSA certificates
8236 on the other hand, since these are separate ciphersuites. The
8237 unused code for Fortezza ciphersuites has been removed.
8239 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8240 (not "ECDHE"). For consistency with the code for DH
8241 certificates, use of ECDH certificates is now considered ECDH
8242 authentication, not RSA or ECDSA authentication (the latter is
8243 merely the CA's signing algorithm and not actively used in the
8246 The temporary ciphersuite alias "ECCdraft" is no longer
8247 available, and ECC ciphersuites are no longer excluded from "ALL"
8248 and "DEFAULT". The following aliases now exist for RFC 4492
8249 ciphersuites, most of these by analogy with the DH case:
8251 kECDHr - ECDH cert, signed with RSA
8252 kECDHe - ECDH cert, signed with ECDSA
8253 kECDH - ECDH cert (signed with either RSA or ECDSA)
8254 kEECDH - ephemeral ECDH
8255 ECDH - ECDH cert or ephemeral ECDH
8261 AECDH - anonymous ECDH
8262 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8266 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8267 Use correct micalg parameters depending on digest(s) in signed message.
8271 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8272 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8276 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8277 an engine to register a method. Add ENGINE lookups for methods and
8278 functional reference processing.
8282 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8283 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8288 * New -resign option to smime utility. This adds one or more signers
8289 to an existing PKCS#7 signedData structure. Also -md option to use an
8290 alternative message digest algorithm for signing.
8294 * Tidy up PKCS#7 routines and add new functions to make it easier to
8295 create PKCS7 structures containing multiple signers. Update smime
8296 application to support multiple signers.
8300 * New -macalg option to pkcs12 utility to allow setting of an alternative
8305 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8306 Reorganize PBE internals to lookup from a static table using NIDs,
8307 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8308 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8309 PRF which will be automatically used with PBES2.
8313 * Replace the algorithm specific calls to generate keys in "req" with the
8318 * Update PKCS#7 enveloped data routines to use new API. This is now
8319 supported by any public key method supporting the encrypt operation. A
8320 ctrl is added to allow the public key algorithm to examine or modify
8321 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8326 * Add a ctrl to asn1 method to allow a public key algorithm to express
8327 a default digest type to use. In most cases this will be SHA1 but some
8328 algorithms (such as GOST) need to specify an alternative digest. The
8329 return value indicates how strong the preference is 1 means optional and
8330 2 is mandatory (that is it is the only supported type). Modify
8331 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8332 use the default md. Update openssl utilities to use the default digest
8333 type for signing if it is not explicitly indicated.
8337 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8338 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8339 signing method from the key type. This effectively removes the link
8340 between digests and public key types.
8344 * Add an OID cross reference table and utility functions. Its purpose is to
8345 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8346 rsaEncryption. This will allow some of the algorithm specific hackery
8347 needed to use the correct OID to be removed.
8351 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8352 structures for PKCS7_sign(). They are now set up by the relevant public
8357 * Add provisional EC pkey method with support for ECDSA and ECDH.
8361 * Add support for key derivation (agreement) in the API, DH method and
8366 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8367 public and private key formats. As a side effect these add additional
8368 command line functionality not previously available: DSA signatures can be
8369 generated and verified using pkeyutl and DH key support and generation in
8376 *Oliver Tappe <zooey@hirschkaefer.de>*
8378 * New make target "install_html_docs" installs HTML renditions of the
8381 *Oliver Tappe <zooey@hirschkaefer.de>*
8383 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8384 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8385 support key and parameter generation and add initial key generation
8386 functionality for RSA.
8390 * Add functions for main EVP_PKEY_method operations. The undocumented
8391 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8392 `EVP_PKEY_{encrypt,decrypt}_old`.
8396 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8397 key API, doesn't do much yet.
8401 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8402 public key algorithms. New option to openssl utility:
8403 "list-public-key-algorithms" to print out info.
8407 * Implement the Supported Elliptic Curves Extension for
8408 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8412 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8413 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8417 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8418 utilities such as rsa, dsa, dsaparam etc except they process any key
8423 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8424 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8425 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8430 * Initial support for pluggable public key ASN1.
8431 De-spaghettify the public key ASN1 handling. Move public and private
8432 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8433 algorithm specific handling to a single module within the relevant
8434 algorithm directory. Add functions to allow (near) opaque processing
8435 of public and private key structures.
8439 * Implement the Supported Point Formats Extension for
8440 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8444 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8445 for the psk identity [hint] and the psk callback functions to the
8446 SSL_SESSION, SSL and SSL_CTX structure.
8449 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8453 SSL_CTX_use_psk_identity_hint
8454 SSL_get_psk_identity_hint
8455 SSL_get_psk_identity
8456 SSL_use_psk_identity_hint
8458 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8460 * Add RFC 3161 compliant time stamp request creation, response generation
8461 and response verification functionality.
8463 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8465 * Add initial support for TLS extensions, specifically for the server_name
8466 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8467 have new members for a host name. The SSL data structure has an
8468 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8469 stored in that context to allow for session resumption, even after the
8470 SSL has been switched to a new SSL_CTX in reaction to a client's
8471 server_name extension.
8473 New functions (subject to change):
8475 SSL_get_servername()
8476 SSL_get_servername_type()
8479 New CTRL codes and macros (subject to change):
8481 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8482 - SSL_CTX_set_tlsext_servername_callback()
8483 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8484 - SSL_CTX_set_tlsext_servername_arg()
8485 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8487 openssl s_client has a new '-servername ...' option.
8489 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8490 '-key2 ...', '-servername_fatal' (subject to change). This allows
8491 testing the HostName extension for a specific single host name ('-cert'
8492 and '-key' remain fallbacks for handshakes without HostName
8493 negotiation). If the unrecognized_name alert has to be sent, this by
8494 default is a warning; it becomes fatal with the '-servername_fatal'
8497 *Peter Sylvester, Remy Allais, Christophe Renou*
8499 * Whirlpool hash implementation is added.
8503 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8504 bn(64,32). Because of instruction set limitations it doesn't have
8505 any negative impact on performance. This was done mostly in order
8506 to make it possible to share assembler modules, such as bn_mul_mont
8507 implementations, between 32- and 64-bit builds without hassle.
8511 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8512 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8517 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8518 dedicated Montgomery multiplication procedure, is introduced.
8519 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8520 "64-bit" performance on certain 32-bit targets.
8524 * New option SSL_OP_NO_COMP to disable use of compression selectively
8525 in SSL structures. New SSL ctrl to set maximum send fragment size.
8526 Save memory by setting the I/O buffer sizes dynamically instead of
8527 using the maximum available value.
8531 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8532 in addition to the text details.
8536 * Very, very preliminary EXPERIMENTAL support for printing of general
8537 ASN1 structures. This currently produces rather ugly output and doesn't
8538 handle several customised structures at all.
8542 * Integrated support for PVK file format and some related formats such
8543 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8544 these in the 'rsa' and 'dsa' utilities.
8548 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8552 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8553 place for the (very old) "NETSCAPE" format certificates which are now
8554 handled using new ASN1 code equivalents.
8558 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8559 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8560 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8564 * Modify CRL distribution points extension code to print out previously
8565 unsupported fields. Enhance extension setting code to allow setting of
8570 * Add print and set support for Issuing Distribution Point CRL extension.
8574 * Change 'Configure' script to enable Camellia by default.
8581 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8583 * When rejecting SSL/TLS records due to an incorrect version number, never
8584 update s->server with a new major version number. As of
8585 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8586 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8587 the previous behavior could result in a read attempt at NULL when
8588 receiving specific incorrect SSL/TLS records once record payload
8589 protection is active. [CVE-2010-0740][]
8591 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8593 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8594 could be crashed if the relevant tables were not present (e.g. chrooted).
8596 *Tomas Hoger <thoger@redhat.com>*
8598 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8600 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8602 *Martin Olsson, Neel Mehta*
8604 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8605 accommodate for stack sorting, always a write lock!).
8609 * On some versions of WIN32 Heap32Next is very slow. This can cause
8610 excessive delays in the RAND_poll(): over a minute. As a workaround
8611 include a time check in the inner Heap32Next loop too.
8615 * The code that handled flushing of data in SSL/TLS originally used the
8616 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8617 the problem outlined in PR#1949. The fix suggested there however can
8618 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8619 of Apache). So instead simplify the code to flush unconditionally.
8620 This should be fine since flushing with no data to flush is a no op.
8624 * Handle TLS versions 2.0 and later properly and correctly use the
8625 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8626 off ancient servers have a habit of sticking around for a while...
8630 * Modify compression code so it frees up structures without using the
8631 ex_data callbacks. This works around a problem where some applications
8632 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8633 restarting) then use compression (e.g. SSL with compression) later.
8634 This results in significant per-connection memory leaks and
8635 has caused some security issues including CVE-2008-1678 and
8640 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8641 change when encrypting or decrypting.
8645 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8646 connect and renegotiate with servers which do not support RI.
8647 Until RI is more widely deployed this option is enabled by default.
8651 * Add "missing" ssl ctrls to clear options and mode.
8655 * If client attempts to renegotiate and doesn't support RI respond with
8656 a no_renegotiation alert as required by RFC5746. Some renegotiating
8657 TLS clients will continue a connection gracefully when they receive
8658 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8659 waiting for a server hello which it will never receive. Now we treat a
8660 received no_renegotiation alert as a fatal error. This is because
8661 applications requesting a renegotiation might well expect it to succeed
8662 and would have no code in place to handle the server denying it so the
8663 only safe thing to do is to terminate the connection.
8667 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8668 peer supports secure renegotiation and 0 otherwise. Print out peer
8669 renegotiation support in s_client/s_server.
8673 * Replace the highly broken and deprecated SPKAC certification method with
8674 the updated NID creation version. This should correctly handle UTF8.
8678 * Implement RFC5746. Re-enable renegotiation but require the extension
8679 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8680 turns out to be a bad idea. It has been replaced by
8681 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8682 SSL_CTX_set_options(). This is really not recommended unless you
8683 know what you are doing.
8685 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8687 * Fixes to stateless session resumption handling. Use initial_ctx when
8688 issuing and attempting to decrypt tickets in case it has changed during
8689 servername handling. Use a non-zero length session ID when attempting
8690 stateless session resumption: this makes it possible to determine if
8691 a resumption has occurred immediately after receiving server hello
8692 (several places in OpenSSL subtly assume this) instead of later in
8697 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8698 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8699 fixes for a few places where the return code is not checked
8702 *Julia Lawall <julia@diku.dk>*
8704 * Add --strict-warnings option to Configure script to include devteam
8705 warnings in other configurations.
8709 * Add support for --libdir option and LIBDIR variable in makefiles. This
8710 makes it possible to install openssl libraries in locations which
8711 have names other than "lib", for example "/usr/lib64" which some
8714 *Steve Henson, based on patch from Jeremy Utley*
8716 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8717 X690 8.9.12 and can produce some misleading textual output of OIDs.
8719 *Steve Henson, reported by Dan Kaminsky*
8721 * Delete MD2 from algorithm tables. This follows the recommendation in
8722 several standards that it is not used in new applications due to
8723 several cryptographic weaknesses. For binary compatibility reasons
8724 the MD2 API is still compiled in by default.
8728 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8733 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8734 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8737 *Guenter <lists@gknw.net>*
8739 * Fix the server certificate chain building code to use X509_verify_cert(),
8740 it used to have an ad-hoc builder which was unable to cope with anything
8741 other than a simple chain.
8743 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8745 * Don't check self signed certificate signatures in X509_verify_cert()
8746 by default (a flag can override this): it just wastes time without
8747 adding any security. As a useful side effect self signed root CAs
8748 with non-FIPS digests are now usable in FIPS mode.
8752 * In dtls1_process_out_of_seq_message() the check if the current message
8753 is already buffered was missing. For every new message was memory
8754 allocated, allowing an attacker to perform an denial of service attack
8755 with sending out of seq handshake messages until there is no memory
8756 left. Additionally every future message was buffered, even if the
8757 sequence number made no sense and would be part of another handshake.
8758 So only messages with sequence numbers less than 10 in advance will be
8759 buffered. [CVE-2009-1378][]
8761 *Robin Seggelmann, discovered by Daniel Mentz*
8763 * Records are buffered if they arrive with a future epoch to be
8764 processed after finishing the corresponding handshake. There is
8765 currently no limitation to this buffer allowing an attacker to perform
8766 a DOS attack with sending records with future epochs until there is no
8767 memory left. This patch adds the pqueue_size() function to determine
8768 the size of a buffer and limits the record buffer to 100 entries.
8771 *Robin Seggelmann, discovered by Daniel Mentz*
8773 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8774 parent structure is freed. [CVE-2009-1379][]
8778 * Handle non-blocking I/O properly in SSL_shutdown() call.
8780 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8782 * Add `2.5.4.*` OIDs
8784 *Ilya O. <vrghost@gmail.com>*
8786 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8788 * Disable renegotiation completely - this fixes a severe security
8789 problem [CVE-2009-3555][] at the cost of breaking all
8790 renegotiation. Renegotiation can be re-enabled by setting
8791 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8792 run-time. This is really not recommended unless you know what
8797 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8799 * Don't set val to NULL when freeing up structures, it is freed up by
8800 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8801 zeroing past the valid field. [CVE-2009-0789][]
8803 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8805 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8806 checked correctly. This would allow some invalid signed attributes to
8807 appear to verify correctly. [CVE-2009-0591][]
8809 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8811 * Reject UniversalString and BMPString types with invalid lengths. This
8812 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8813 a legal length. [CVE-2009-0590][]
8817 * Set S/MIME signing as the default purpose rather than setting it
8818 unconditionally. This allows applications to override it at the store
8823 * Permit restricted recursion of ASN1 strings. This is needed in practice
8824 to handle some structures.
8828 * Improve efficiency of mem_gets: don't search whole buffer each time
8831 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8833 * New -hex option for openssl rand.
8837 * Print out UTF8String and NumericString when parsing ASN1.
8841 * Support NumericString type for name components.
8845 * Allow CC in the environment to override the automatically chosen
8846 compiler. Note that nothing is done to ensure flags work with the
8851 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8853 * Properly check EVP_VerifyFinal() and similar return values
8856 *Ben Laurie, Bodo Moeller, Google Security Team*
8858 * Enable TLS extensions by default.
8862 * Allow the CHIL engine to be loaded, whether the application is
8863 multithreaded or not. (This does not release the developer from the
8864 obligation to set up the dynamic locking callbacks.)
8866 *Sander Temme <sander@temme.net>*
8868 * Use correct exit code if there is an error in dgst command.
8870 *Steve Henson; problem pointed out by Roland Dirlewanger*
8872 * Tweak Configure so that you need to say "experimental-jpake" to enable
8873 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8877 * Add experimental JPAKE support, including demo authentication in
8878 s_client and s_server.
8882 * Set the comparison function in v3_addr_canonize().
8884 *Rob Austein <sra@hactrn.net>*
8886 * Add support for XMPP STARTTLS in s_client.
8888 *Philip Paeps <philip@freebsd.org>*
8890 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8891 to ensure that even with this option, only ciphersuites in the
8892 server's preference list will be accepted. (Note that the option
8893 applies only when resuming a session, so the earlier behavior was
8894 just about the algorithm choice for symmetric cryptography.)
8898 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8900 * Fix NULL pointer dereference if a DTLS server received
8901 ChangeCipherSpec as first record [CVE-2009-1386][].
8905 * Fix a state transition in s3_srvr.c and d1_srvr.c
8906 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8910 * The fix in 0.9.8c that supposedly got rid of unsafe
8911 double-checked locking was incomplete for RSA blinding,
8912 addressing just one layer of what turns out to have been
8913 doubly unsafe triple-checked locking.
8915 So now fix this for real by retiring the MONT_HELPER macro
8916 in crypto/rsa/rsa_eay.c.
8918 *Bodo Moeller; problem pointed out by Marius Schilder*
8920 * Various precautionary measures:
8922 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8924 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8925 (NB: This would require knowledge of the secret session ticket key
8926 to exploit, in which case you'd be SOL either way.)
8928 - Change bn_nist.c so that it will properly handle input BIGNUMs
8929 outside the expected range.
8931 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8934 *Neel Mehta, Bodo Moeller*
8936 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8937 the load fails. Useful for distros.
8939 *Ben Laurie and the FreeBSD team*
8941 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8945 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8949 * Expand ENGINE to support engine supplied SSL client certificate functions.
8951 This work was sponsored by Logica.
8955 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8956 keystores. Support for SSL/TLS client authentication too.
8957 Not compiled unless enable-capieng specified to Configure.
8959 This work was sponsored by Logica.
8963 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8964 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8965 attribute creation routines such as certificate requests and PKCS#12
8970 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8972 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8973 handshake which could lead to a client crash as found using the
8974 Codenomicon TLS test suite [CVE-2008-1672][]
8976 *Steve Henson, Mark Cox*
8978 * Fix double free in TLS server name extensions which could lead to
8979 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8983 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8985 Clear the error queue to ensure that error entries left from
8986 older function calls do not interfere with the correct operation.
8988 *Lutz Jaenicke, Erik de Castro Lopo*
8990 * Remove root CA certificates of commercial CAs:
8992 The OpenSSL project does not recommend any specific CA and does not
8993 have any policy with respect to including or excluding any CA.
8994 Therefore it does not make any sense to ship an arbitrary selection
8995 of root CA certificates with the OpenSSL software.
8999 * RSA OAEP patches to fix two separate invalid memory reads.
9000 The first one involves inputs when 'lzero' is greater than
9001 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9002 before the beginning of from). The second one involves inputs where
9003 the 'db' section contains nothing but zeroes (there is a one-byte
9004 invalid read after the end of 'db').
9006 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9008 * Partial backport from 0.9.9-dev:
9010 Introduce bn_mul_mont (dedicated Montgomery multiplication
9011 procedure) as a candidate for BIGNUM assembler implementation.
9012 While 0.9.9-dev uses assembler for various architectures, only
9013 x86_64 is available by default here in the 0.9.8 branch, and
9014 32-bit x86 is available through a compile-time setting.
9016 To try the 32-bit x86 assembler implementation, use Configure
9017 option "enable-montasm" (which exists only for this backport).
9019 As "enable-montasm" for 32-bit x86 disclaims code stability
9020 anyway, in this constellation we activate additional code
9021 backported from 0.9.9-dev for further performance improvements,
9022 namely BN_from_montgomery_word. (To enable this otherwise,
9023 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9025 *Andy Polyakov (backport partially by Bodo Moeller)*
9027 * Add TLS session ticket callback. This allows an application to set
9028 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9029 values. This is useful for key rollover for example where several key
9030 sets may exist with different names.
9034 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9035 This was broken until now in 0.9.8 releases, such that the only way
9036 a registered ENGINE could be used (assuming it initialises
9037 successfully on the host) was to explicitly set it as the default
9038 for the relevant algorithms. This is in contradiction with 0.9.7
9039 behaviour and the documentation. With this fix, when an ENGINE is
9040 registered into a given algorithm's table of implementations, the
9041 'uptodate' flag is reset so that auto-discovery will be used next
9042 time a new context for that algorithm attempts to select an
9045 *Ian Lister (tweaked by Geoff Thorpe)*
9047 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9048 implementation in the following ways:
9050 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9053 Lack of BER streaming support means one pass streaming processing is
9054 only supported if data is detached: setting the streaming flag is
9055 ignored for embedded content.
9057 CMS support is disabled by default and must be explicitly enabled
9058 with the enable-cms configuration option.
9062 * Update the GMP engine glue to do direct copies between BIGNUM and
9063 mpz_t when openssl and GMP use the same limb size. Otherwise the
9064 existing "conversion via a text string export" trick is still used.
9066 *Paul Sheer <paulsheer@gmail.com>*
9068 * Zlib compression BIO. This is a filter BIO which compressed and
9069 uncompresses any data passed through it.
9073 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9074 RFC3394 compatible AES key wrapping.
9078 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9079 sets string data without copying. X509_ALGOR_set0() and
9080 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9081 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9082 from an X509_ATTRIBUTE structure optionally checking it occurs only
9083 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9088 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9089 to get the expected BN_FLG_CONSTTIME behavior.
9091 *Bodo Moeller (Google)*
9095 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9096 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9097 - added some more tests to do_tests.pl
9098 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9099 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9100 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9101 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9102 - various changes to netware.pl to enable gcc-cross builds on Win32
9104 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9105 - various changes to fix missing prototype warnings
9106 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9107 - added AES, WHIRLPOOL and CPUID assembler code to build files
9108 - added missing AES assembler make rules to mk1mf.pl
9109 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9111 *Guenter Knauf <eflash@gmx.net>*
9113 * Implement certificate status request TLS extension defined in RFC3546.
9114 A client can set the appropriate parameters and receive the encoded
9115 OCSP response via a callback. A server can query the supplied parameters
9116 and set the encoded OCSP response in the callback. Add simplified examples
9117 to s_client and s_server.
9121 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9124 + Binary incompatibility of ssl_ctx_st structure
9125 + DTLS interoperation with non-compliant servers
9126 + Don't call get_session_cb() without proposed session
9127 + Fix ia64 assembler code
9129 *Andy Polyakov, Steve Henson*
9131 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9133 * DTLS Handshake overhaul. There were longstanding issues with
9134 OpenSSL DTLS implementation, which were making it impossible for
9135 RFC 4347 compliant client to communicate with OpenSSL server.
9136 Unfortunately just fixing these incompatibilities would "cut off"
9137 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9138 server keeps tolerating non RFC compliant syntax. The opposite is
9139 not true, 0.9.8f client can not communicate with earlier server.
9140 This update even addresses CVE-2007-4995.
9144 * Changes to avoid need for function casts in OpenSSL: some compilers
9145 (gcc 4.2 and later) reject their use.
9146 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9149 * Add RFC4507 support to OpenSSL. This includes the corrections in
9150 RFC4507bis. The encrypted ticket format is an encrypted encoded
9151 SSL_SESSION structure, that way new session features are automatically
9154 If a client application caches session in an SSL_SESSION structure
9155 support is transparent because tickets are now stored in the encoded
9158 The SSL_CTX structure automatically generates keys for ticket
9159 protection in servers so again support should be possible
9160 with no application modification.
9162 If a client or server wishes to disable RFC4507 support then the option
9163 SSL_OP_NO_TICKET can be set.
9165 Add a TLS extension debugging callback to allow the contents of any client
9166 or server extensions to be examined.
9168 This work was sponsored by Google.
9172 * Add initial support for TLS extensions, specifically for the server_name
9173 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9174 have new members for a host name. The SSL data structure has an
9175 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9176 stored in that context to allow for session resumption, even after the
9177 SSL has been switched to a new SSL_CTX in reaction to a client's
9178 server_name extension.
9180 New functions (subject to change):
9182 SSL_get_servername()
9183 SSL_get_servername_type()
9186 New CTRL codes and macros (subject to change):
9188 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9189 - SSL_CTX_set_tlsext_servername_callback()
9190 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9191 - SSL_CTX_set_tlsext_servername_arg()
9192 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9194 openssl s_client has a new '-servername ...' option.
9196 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9197 '-key2 ...', '-servername_fatal' (subject to change). This allows
9198 testing the HostName extension for a specific single host name ('-cert'
9199 and '-key' remain fallbacks for handshakes without HostName
9200 negotiation). If the unrecognized_name alert has to be sent, this by
9201 default is a warning; it becomes fatal with the '-servername_fatal'
9204 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9206 * Add AES and SSE2 assembly language support to VC++ build.
9210 * Mitigate attack on final subtraction in Montgomery reduction.
9214 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9215 (which previously caused an internal error).
9219 * Squeeze another 10% out of IGE mode when in != out.
9223 * AES IGE mode speedup.
9225 *Dean Gaudet (Google)*
9227 * Add the Korean symmetric 128-bit cipher SEED (see
9228 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9229 add SEED ciphersuites from RFC 4162:
9231 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9232 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9233 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9234 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9236 To minimize changes between patchlevels in the OpenSSL 0.9.8
9237 series, SEED remains excluded from compilation unless OpenSSL
9238 is configured with 'enable-seed'.
9240 *KISA, Bodo Moeller*
9242 * Mitigate branch prediction attacks, which can be practical if a
9243 single processor is shared, allowing a spy process to extract
9244 information. For detailed background information, see
9245 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9246 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9247 and Necessary Software Countermeasures"). The core of the change
9248 are new versions BN_div_no_branch() and
9249 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9250 respectively, which are slower, but avoid the security-relevant
9251 conditional branches. These are automatically called by BN_div()
9252 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9253 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9254 remove a conditional branch.
9256 BN_FLG_CONSTTIME is the new name for the previous
9257 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9258 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9259 in the exponent causes BN_mod_exp_mont() to use the alternative
9260 implementation in BN_mod_exp_mont_consttime().) The old name
9261 remains as a deprecated alias.
9263 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9264 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9265 constant-time implementations for more than just exponentiation.
9266 Here too the old name is kept as a deprecated alias.
9268 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9269 the BN_BLINDING structure gets an independent copy of the
9270 modulus. This means that the previous `BIGNUM *m` argument to
9271 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9272 essentially becomes `const BIGNUM *m`, although we can't actually
9273 change this in the header file before 0.9.9. It allows
9274 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9275 enable BN_FLG_CONSTTIME.
9277 *Matthew D Wood (Intel Corp)*
9279 * In the SSL/TLS server implementation, be strict about session ID
9280 context matching (which matters if an application uses a single
9281 external cache for different purposes). Previously,
9282 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9283 set. This did ensure strict client verification, but meant that,
9284 with applications using a single external cache for quite
9285 different requirements, clients could circumvent ciphersuite
9286 restrictions for a given session ID context by starting a session
9287 in a different context.
9291 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9292 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9293 authentication-only ciphersuites.
9297 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9298 not complete and could lead to a possible single byte overflow
9299 [CVE-2007-5135][] [Ben Laurie]
9301 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9303 * Since AES128 and AES256 (and similarly Camellia128 and
9304 Camellia256) share a single mask bit in the logic of
9305 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9306 kludge to work properly if AES128 is available and AES256 isn't
9307 (or if Camellia128 is available and Camellia256 isn't).
9311 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9312 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9313 When a point or a seed is encoded in a BIT STRING, we need to
9314 prevent the removal of trailing zero bits to get the proper DER
9315 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9316 of a NamedBitList, for which trailing 0 bits need to be removed.)
9320 * Have SSL/TLS server implementation tolerate "mismatched" record
9321 protocol version while receiving ClientHello even if the
9322 ClientHello is fragmented. (The server can't insist on the
9323 particular protocol version it has chosen before the ServerHello
9324 message has informed the client about his choice.)
9328 * Add RFC 3779 support.
9330 *Rob Austein for ARIN, Ben Laurie*
9332 * Load error codes if they are not already present instead of using a
9333 static variable. This allows them to be cleanly unloaded and reloaded.
9334 Improve header file function name parsing.
9338 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9339 or CAPABILITY handshake as required by RFCs.
9343 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9345 * Introduce limits to prevent malicious keys being able to
9346 cause a denial of service. [CVE-2006-2940][]
9348 *Steve Henson, Bodo Moeller*
9350 * Fix ASN.1 parsing of certain invalid structures that can result
9351 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9353 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9354 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9356 * Fix SSL client code which could crash if connecting to a
9357 malicious SSLv2 server. [CVE-2006-4343][]
9359 *Tavis Ormandy and Will Drewry, Google Security Team*
9361 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9362 match only those. Before that, "AES256-SHA" would be interpreted
9363 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9364 the same strength classification in 0.9.7h) as we currently only
9365 have a single AES bit in the ciphersuite description bitmap.
9366 That change, however, also applied to ciphersuite strings such as
9367 "RC4-MD5" that intentionally matched multiple ciphersuites --
9368 namely, SSL 2.0 ciphersuites in addition to the more common ones
9369 from SSL 3.0/TLS 1.0.
9371 So we change the selection algorithm again: Naming an explicit
9372 ciphersuite selects this one ciphersuite, and any other similar
9373 ciphersuite (same bitmap) from *other* protocol versions.
9374 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9375 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9377 Since SSL 2.0 does not have any ciphersuites for which the
9378 128/256 bit distinction would be relevant, this works for now.
9379 The proper fix will be to use different bits for AES128 and
9380 AES256, which would have avoided the problems from the beginning;
9381 however, bits are scarce, so we can only do this in a new release
9382 (not just a patchlevel) when we can change the SSL_CIPHER
9383 definition to split the single 'unsigned long mask' bitmap into
9384 multiple values to extend the available space.
9388 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9390 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9391 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9393 * Add AES IGE and biIGE modes.
9397 * Change the Unix randomness entropy gathering to use poll() when
9398 possible instead of select(), since the latter has some
9399 undesirable limitations.
9401 *Darryl Miles via Richard Levitte and Bodo Moeller*
9403 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9404 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9405 cannot be implicitly activated as part of, e.g., the "AES" alias.
9406 However, please upgrade to OpenSSL 0.9.9[-dev] for
9407 non-experimental use of the ECC ciphersuites to get TLS extension
9408 support, which is required for curve and point format negotiation
9409 to avoid potential handshake problems.
9413 * Disable rogue ciphersuites:
9415 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9416 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9417 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9419 The latter two were purportedly from
9420 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9423 Also deactivate the remaining ciphersuites from
9424 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9425 unofficial, and the ID has long expired.
9429 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9430 dual-core machines) and other potential thread-safety issues.
9434 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9435 versions), which is now available for royalty-free use
9436 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9437 Also, add Camellia TLS ciphersuites from RFC 4132.
9439 To minimize changes between patchlevels in the OpenSSL 0.9.8
9440 series, Camellia remains excluded from compilation unless OpenSSL
9441 is configured with 'enable-camellia'.
9445 * Disable the padding bug check when compression is in use. The padding
9446 bug check assumes the first packet is of even length, this is not
9447 necessarily true if compression is enabled and can result in false
9448 positives causing handshake failure. The actual bug test is ancient
9449 code so it is hoped that implementations will either have fixed it by
9450 now or any which still have the bug do not support compression.
9454 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9456 * When applying a cipher rule check to see if string match is an explicit
9457 cipher suite and only match that one cipher suite if it is.
9461 * Link in manifests for VC++ if needed.
9463 *Austin Ziegler <halostatue@gmail.com>*
9465 * Update support for ECC-based TLS ciphersuites according to
9466 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9467 TLS extensions, which are supported starting with the 0.9.9
9468 branch, not in the OpenSSL 0.9.8 branch).
9472 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9473 opaque EVP_CIPHER_CTX handling.
9477 * Fixes and enhancements to zlib compression code. We now only use
9478 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9479 to conform with the standards mentioned here:
9480 <http://www.zlib.net/DLL_FAQ.txt>
9481 Static zlib linking now works on Windows and the new --with-zlib-include
9482 --with-zlib-lib options to Configure can be used to supply the location
9483 of the headers and library. Gracefully handle case where zlib library
9488 * Several fixes and enhancements to the OID generation code. The old code
9489 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9490 handle numbers larger than ULONG_MAX, truncated printing and had a
9491 non standard OBJ_obj2txt() behaviour.
9495 * Add support for building of engines under engine/ as shared libraries
9496 under VC++ build system.
9500 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9501 Hopefully, we will not see any false combination of paths any more.
9505 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9507 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9508 (part of SSL_OP_ALL). This option used to disable the
9509 countermeasure against man-in-the-middle protocol-version
9510 rollback in the SSL 2.0 server implementation, which is a bad
9511 idea. [CVE-2005-2969][]
9513 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9514 for Information Security, National Institute of Advanced Industrial
9515 Science and Technology [AIST], Japan)*
9517 * Add two function to clear and return the verify parameter flags.
9521 * Keep cipherlists sorted in the source instead of sorting them at
9522 runtime, thus removing the need for a lock.
9526 * Avoid some small subgroup attacks in Diffie-Hellman.
9528 *Nick Mathewson and Ben Laurie*
9530 * Add functions for well-known primes.
9534 * Extended Windows CE support.
9536 *Satoshi Nakamura and Andy Polyakov*
9538 * Initialize SSL_METHOD structures at compile time instead of during
9539 runtime, thus removing the need for a lock.
9543 * Make PKCS7_decrypt() work even if no certificate is supplied by
9544 attempting to decrypt each encrypted key in turn. Add support to
9549 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9551 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9554 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9558 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9559 key into the same file any more.
9563 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9567 * Add -utf8 command line and config file option to 'ca'.
9569 *Stefan <stf@udoma.org*
9571 * Removed the macro des_crypt(), as it seems to conflict with some
9572 libraries. Use DES_crypt().
9576 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9577 involves renaming the source and generated shared-libs for
9578 both. The engines will accept the corrected or legacy ids
9579 ('ncipher' and '4758_cca' respectively) when binding. NB,
9580 this only applies when building 'shared'.
9582 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9584 * Add attribute functions to EVP_PKEY structure. Modify
9585 PKCS12_create() to recognize a CSP name attribute and
9586 use it. Make -CSP option work again in pkcs12 utility.
9590 * Add new functionality to the bn blinding code:
9591 - automatic re-creation of the BN_BLINDING parameters after
9592 a fixed number of uses (currently 32)
9593 - add new function for parameter creation
9594 - introduce flags to control the update behaviour of the
9595 BN_BLINDING parameters
9596 - hide BN_BLINDING structure
9597 Add a second BN_BLINDING slot to the RSA structure to improve
9598 performance when a single RSA object is shared among several
9603 * Add support for DTLS.
9605 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9607 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9608 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9612 * Remove buggy and incomplete DH cert support from
9613 ssl/ssl_rsa.c and ssl/s3_both.c
9617 * Use SHA-1 instead of MD5 as the default digest algorithm for
9618 the apps/openssl applications.
9622 * Compile clean with "-Wall -Wmissing-prototypes
9623 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9624 DEBUG_SAFESTACK must also be set.
9628 * Change ./Configure so that certain algorithms can be disabled by default.
9629 The new counterpiece to "no-xxx" is "enable-xxx".
9631 The patented RC5 and MDC2 algorithms will now be disabled unless
9632 "enable-rc5" and "enable-mdc2", respectively, are specified.
9634 (IDEA remains enabled despite being patented. This is because IDEA
9635 is frequently required for interoperability, and there is no license
9636 fee for non-commercial use. As before, "no-idea" can be used to
9637 avoid this algorithm.)
9641 * Add processing of proxy certificates (see RFC 3820). This work was
9642 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9643 EGEE (Enabling Grids for E-science in Europe).
9647 * RC4 performance overhaul on modern architectures/implementations, such
9648 as Intel P4, IA-64 and AMD64.
9652 * New utility extract-section.pl. This can be used specify an alternative
9653 section number in a pod file instead of having to treat each file as
9654 a separate case in Makefile. This can be done by adding two lines to the
9657 =for comment openssl_section:XXX
9659 The blank line is mandatory.
9663 * New arguments -certform, -keyform and -pass for s_client and s_server
9664 to allow alternative format key and certificate files and passphrase
9669 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9670 update associated structures and add various utility functions.
9672 Add new policy related verify parameters, include policy checking in
9673 standard verify code. Enhance 'smime' application with extra parameters
9674 to support policy checking and print out.
9678 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9679 Nehemiah processors. These extensions support AES encryption in hardware
9680 as well as RNG (though RNG support is currently disabled).
9682 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9684 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9688 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9690 *Andy Polyakov and a number of other people*
9692 * Improved PowerPC platform support. Most notably BIGNUM assembler
9693 implementation contributed by IBM.
9695 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9697 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9698 exponent rather than 'unsigned long'. There is a corresponding change to
9699 the new 'rsa_keygen' element of the RSA_METHOD structure.
9701 *Jelte Jansen, Geoff Thorpe*
9703 * Functionality for creating the initial serial number file is now
9704 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9706 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9707 number file to 1, which is bound to cause problems. To avoid
9708 the problems while respecting compatibility between different 0.9.7
9709 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9710 CA.pl for serial number initialization. With the new release 0.9.8,
9711 we can fix the problem directly in the 'ca' utility.)
9715 * Reduced header interdependencies by declaring more opaque objects in
9716 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9717 give fewer recursive includes, which could break lazy source code - so
9718 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9719 developers should define this symbol when building and using openssl to
9720 ensure they track the recommended behaviour, interfaces, [etc], but
9721 backwards-compatible behaviour prevails when this isn't defined.
9725 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9729 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9730 This will generate a random key of the appropriate length based on the
9731 cipher context. The EVP_CIPHER can provide its own random key generation
9732 routine to support keys of a specific form. This is used in the des and
9733 3des routines to generate a key of the correct parity. Update S/MIME
9734 code to use new functions and hence generate correct parity DES keys.
9735 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9736 valid (weak or incorrect parity).
9740 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9741 as looking them up. This is useful when the verified structure may contain
9742 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9743 present unless the new PKCS7_NO_CRL flag is asserted.
9747 * Extend ASN1 oid configuration module. It now additionally accepts the
9750 shortName = some long name, 1.2.3.4
9754 * Reimplemented the BN_CTX implementation. There is now no more static
9755 limitation on the number of variables it can handle nor the depth of the
9756 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9757 information can now expand as required, and rather than having a single
9758 static array of bignums, BN_CTX now uses a linked-list of such arrays
9759 allowing it to expand on demand whilst maintaining the usefulness of
9760 BN_CTX's "bundling".
9764 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9765 to allow all RSA operations to function using a single BN_CTX.
9769 * Preliminary support for certificate policy evaluation and checking. This
9770 is initially intended to pass the tests outlined in "Conformance Testing
9771 of Relying Party Client Certificate Path Processing Logic" v1.07.
9775 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9776 remained unused and not that useful. A variety of other little bignum
9777 tweaks and fixes have also been made continuing on from the audit (see
9782 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9783 associated ASN1, EVP and SSL functions and old ASN1 macros.
9787 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9788 and this should never fail. So the return value from the use of
9789 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9790 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9794 * BN_CTX_get() should return zero-valued bignums, providing the same
9795 initialised value as BN_new().
9797 *Geoff Thorpe, suggested by Ulf Möller*
9799 * Support for inhibitAnyPolicy certificate extension.
9803 * An audit of the BIGNUM code is underway, for which debugging code is
9804 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9805 is considered valid when processing BIGNUMs, and causes execution to
9806 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9807 further steps are taken to deliberately pollute unused data in BIGNUM
9808 structures to try and expose faulty code further on. For now, openssl will
9809 (in its default mode of operation) continue to tolerate the inconsistent
9810 forms that it has tolerated in the past, but authors and packagers should
9811 consider trying openssl and their own applications when compiled with
9812 these debugging symbols defined. It will help highlight potential bugs in
9813 their own code, and will improve the test coverage for OpenSSL itself. At
9814 some point, these tighter rules will become openssl's default to improve
9815 maintainability, though the assert()s and other overheads will remain only
9816 in debugging configurations. See bn.h for more details.
9818 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9820 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9821 that can only be obtained through BN_CTX_new() (which implicitly
9822 initialises it). The presence of this function only made it possible
9823 to overwrite an existing structure (and cause memory leaks).
9827 * Because of the callback-based approach for implementing LHASH as a
9828 template type, lh_insert() adds opaque objects to hash-tables and
9829 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9830 to clean up those corresponding objects before destroying the hash table
9831 (and losing the object pointers). So some over-zealous constifications in
9832 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9833 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9834 prototyped to have "const" restrictions on the object pointers they are
9835 given (and so aren't required to cast them away any more).
9839 * The tmdiff.h API was so ugly and minimal that our own timing utility
9840 (speed) prefers to use its own implementation. The two implementations
9841 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9842 its object type properly exposed (MS_TM) instead of casting to/from
9843 `char *`. This may still change yet if someone realises MS_TM and
9845 aren't necessarily the greatest nomenclatures - but this is what was used
9846 internally to the implementation so I've used that for now.
9850 * Ensure that deprecated functions do not get compiled when
9851 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9852 the self-tests were still using deprecated key-generation functions so
9853 these have been updated also.
9857 * Reorganise PKCS#7 code to separate the digest location functionality
9858 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9859 New function PKCS7_set_digest() to set the digest type for PKCS#7
9860 digestedData type. Add additional code to correctly generate the
9861 digestedData type and add support for this type in PKCS7 initialization
9866 * New function PKCS7_set0_type_other() this initializes a PKCS7
9867 structure of type "other".
9871 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9872 sure the loop does correctly stop and breaking ("division by zero")
9873 modulus operations are not performed. The (pre-generated) prime
9874 table crypto/bn/bn_prime.h was already correct, but it could not be
9875 re-generated on some platforms because of the "division by zero"
9876 situation in the script.
9878 *Ralf S. Engelschall*
9880 * Update support for ECC-based TLS ciphersuites according to
9881 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9882 SHA-1 now is only used for "small" curves (where the
9883 representation of a field element takes up to 24 bytes); for
9884 larger curves, the field element resulting from ECDH is directly
9885 used as premaster secret.
9887 *Douglas Stebila (Sun Microsystems Laboratories)*
9889 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9890 curve secp160r1 to the tests.
9892 *Douglas Stebila (Sun Microsystems Laboratories)*
9894 * Add the possibility to load symbols globally with DSO.
9896 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9898 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9899 control of the error stack.
9903 * Add support for STORE in ENGINE.
9907 * Add the STORE type. The intention is to provide a common interface
9908 to certificate and key stores, be they simple file-based stores, or
9909 HSM-type store, or LDAP stores, or...
9910 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9914 * Add a generic structure called OPENSSL_ITEM. This can be used to
9915 pass a list of arguments to any function as well as provide a way
9916 for a function to pass data back to the caller.
9920 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9921 works like BUF_strdup() but can be used to duplicate a portion of
9922 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9927 * Add the function sk_find_ex() which works like sk_find(), but will
9928 return an index to an element even if an exact match couldn't be
9929 found. The index is guaranteed to point at the element where the
9930 searched-for key would be inserted to preserve sorting order.
9934 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9935 takes an extra flags argument for optional functionality. Currently,
9936 the following flags are defined:
9938 OBJ_BSEARCH_VALUE_ON_NOMATCH
9939 This one gets OBJ_bsearch_ex() to return a pointer to the first
9940 element where the comparing function returns a negative or zero
9943 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9944 This one gets OBJ_bsearch_ex() to return a pointer to the first
9945 element where the comparing function returns zero. This is useful
9946 if there are more than one element where the comparing function
9951 * Make it possible to create self-signed certificates with 'openssl ca'
9952 in such a way that the self-signed certificate becomes part of the
9953 CA database and uses the same mechanisms for serial number generation
9954 as all other certificate signing. The new flag '-selfsign' enables
9955 this functionality. Adapt CA.sh and CA.pl.in.
9959 * Add functionality to check the public key of a certificate request
9960 against a given private. This is useful to check that a certificate
9961 request can be signed by that key (self-signing).
9965 * Make it possible to have multiple active certificates with the same
9966 subject in the CA index file. This is done only if the keyword
9967 'unique_subject' is set to 'no' in the main CA section (default
9968 if 'CA_default') of the configuration file. The value is saved
9969 with the database itself in a separate index attribute file,
9970 named like the index file with '.attr' appended to the name.
9974 * Generate multi-valued AVAs using '+' notation in config files for
9979 * Support for nameConstraints certificate extension.
9983 * Support for policyConstraints certificate extension.
9987 * Support for policyMappings certificate extension.
9991 * Make sure the default DSA_METHOD implementation only uses its
9992 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9993 and change its own handlers to be NULL so as to remove unnecessary
9994 indirection. This lets alternative implementations fallback to the
9995 default implementation more easily.
9999 * Support for directoryName in GeneralName related extensions
10004 * Make it possible to link applications using Makefile.shared.
10005 Make that possible even when linking against static libraries!
10009 * Support for single pass processing for S/MIME signing. This now
10010 means that S/MIME signing can be done from a pipe, in addition
10011 cleartext signing (multipart/signed type) is effectively streaming
10012 and the signed data does not need to be all held in memory.
10014 This is done with a new flag PKCS7_STREAM. When this flag is set
10015 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10016 is done after the data is output (and digests calculated) in
10017 SMIME_write_PKCS7().
10021 * Add full support for -rpath/-R, both in shared libraries and
10022 applications, at least on the platforms where it's known how
10027 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10028 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10029 will now compute a table of multiples of the generator that
10030 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10031 faster (notably in the case of a single point multiplication,
10032 scalar * generator).
10034 *Nils Larsch, Bodo Moeller*
10036 * IPv6 support for certificate extensions. The various extensions
10037 which use the IP:a.b.c.d can now take IPv6 addresses using the
10038 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10043 * Added an ENGINE that implements RSA by performing private key
10044 exponentiations with the GMP library. The conversions to and from
10045 GMP's mpz_t format aren't optimised nor are any montgomery forms
10046 cached, and on x86 it appears OpenSSL's own performance has caught up.
10047 However there are likely to be other architectures where GMP could
10048 provide a boost. This ENGINE is not built in by default, but it can be
10049 specified at Configure time and should be accompanied by the necessary
10050 linker additions, eg;
10051 ./config -DOPENSSL_USE_GMP -lgmp
10055 * "openssl engine" will not display ENGINE/DSO load failure errors when
10056 testing availability of engines with "-t" - the old behaviour is
10057 produced by increasing the feature's verbosity with "-tt".
10061 * ECDSA routines: under certain error conditions uninitialized BN objects
10062 could be freed. Solution: make sure initialization is performed early
10063 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10068 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10069 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10070 software implementations. For DSA and DH, parameter generation can
10071 also be overridden by providing the appropriate method callbacks.
10075 * Change the "progress" mechanism used in key-generation and
10076 primality testing to functions that take a new BN_GENCB pointer in
10077 place of callback/argument pairs. The new API functions have `_ex`
10078 postfixes and the older functions are reimplemented as wrappers for
10079 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10080 declarations of the old functions to help (graceful) attempts to
10081 migrate to the new functions. Also, the new key-generation API
10082 functions operate on a caller-supplied key-structure and return
10083 success/failure rather than returning a key or NULL - this is to
10084 help make "keygen" another member function of RSA_METHOD etc.
10086 Example for using the new callback interface:
10088 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10089 void *my_arg = ...;
10092 BN_GENCB_set(&my_cb, my_callback, my_arg);
10094 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10095 /* For the meaning of a, b in calls to my_callback(), see the
10096 * documentation of the function that calls the callback.
10097 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10098 * my_callback should return 1 if it wants BN_is_prime_ex()
10099 * to continue, or 0 to stop.
10104 * Change the ZLIB compression method to be stateful, and make it
10105 available to TLS with the number defined in
10106 draft-ietf-tls-compression-04.txt.
10110 * Add the ASN.1 structures and functions for CertificatePair, which
10111 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10113 CertificatePair ::= SEQUENCE {
10114 forward [0] Certificate OPTIONAL,
10115 reverse [1] Certificate OPTIONAL,
10116 -- at least one of the pair shall be present -- }
10118 Also implement the PEM functions to read and write certificate
10119 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10121 This needed to be defined, mostly for the sake of the LDAP
10122 attribute crossCertificatePair, but may prove useful elsewhere as
10127 * Make it possible to inhibit symlinking of shared libraries in
10128 Makefile.shared, for Cygwin's sake.
10132 * Extend the BIGNUM API by creating a function
10133 void BN_set_negative(BIGNUM *a, int neg);
10134 and a macro that behave like
10135 int BN_is_negative(const BIGNUM *a);
10137 to avoid the need to access 'a->neg' directly in applications.
10141 * Implement fast modular reduction for pseudo-Mersenne primes
10142 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10143 EC_GROUP_new_curve_GFp() will now automatically use this
10146 *Nils Larsch <nla@trustcenter.de>*
10148 * Add new lock type (CRYPTO_LOCK_BN).
10152 * Change the ENGINE framework to automatically load engines
10153 dynamically from specific directories unless they could be
10154 found to already be built in or loaded. Move all the
10155 current engines except for the cryptodev one to a new
10156 directory engines/.
10157 The engines in engines/ are built as shared libraries if
10158 the "shared" options was given to ./Configure or ./config.
10159 Otherwise, they are inserted in libcrypto.a.
10160 /usr/local/ssl/engines is the default directory for dynamic
10161 engines, but that can be overridden at configure time through
10162 the usual use of --prefix and/or --openssldir, and at run
10163 time with the environment variable OPENSSL_ENGINES.
10165 *Geoff Thorpe and Richard Levitte*
10167 * Add Makefile.shared, a helper makefile to build shared
10168 libraries. Adapt Makefile.org.
10172 * Add version info to Win32 DLLs.
10174 *Peter 'Luna' Runestig" <peter@runestig.com>*
10176 * Add new 'medium level' PKCS#12 API. Certificates and keys
10177 can be added using this API to created arbitrary PKCS#12
10178 files while avoiding the low level API.
10180 New options to PKCS12_create(), key or cert can be NULL and
10181 will then be omitted from the output file. The encryption
10182 algorithm NIDs can be set to -1 for no encryption, the mac
10183 iteration count can be set to 0 to omit the mac.
10185 Enhance pkcs12 utility by making the -nokeys and -nocerts
10186 options work when creating a PKCS#12 file. New option -nomac
10187 to omit the mac, NONE can be set for an encryption algorithm.
10188 New code is modified to use the enhanced PKCS12_create()
10189 instead of the low level API.
10193 * Extend ASN1 encoder to support indefinite length constructed
10194 encoding. This can output sequences tags and octet strings in
10195 this form. Modify pk7_asn1.c to support indefinite length
10196 encoding. This is experimental and needs additional code to
10197 be useful, such as an ASN1 bio and some enhanced streaming
10200 Extend template encode functionality so that tagging is passed
10201 down to the template encoder.
10205 * Let 'openssl req' fail if an argument to '-newkey' is not
10206 recognized instead of using RSA as a default.
10210 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10211 As these are not official, they are not included in "ALL";
10212 the "ECCdraft" ciphersuite group alias can be used to select them.
10214 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10216 * Add ECDH engine support.
10218 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10220 * Add ECDH in new directory crypto/ecdh/.
10222 *Douglas Stebila (Sun Microsystems Laboratories)*
10224 * Let BN_rand_range() abort with an error after 100 iterations
10225 without success (which indicates a broken PRNG).
10229 * Change BN_mod_sqrt() so that it verifies that the input value
10230 is really the square of the return value. (Previously,
10231 BN_mod_sqrt would show GIGO behaviour.)
10235 * Add named elliptic curves over binary fields from X9.62, SECG,
10236 and WAP/WTLS; add OIDs that were still missing.
10238 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10240 * Extend the EC library for elliptic curves over binary fields
10241 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10244 EC_GF2m_simple_method
10248 EC_GROUP_new_curve_GF2m
10249 EC_GROUP_set_curve_GF2m
10250 EC_GROUP_get_curve_GF2m
10251 EC_POINT_set_affine_coordinates_GF2m
10252 EC_POINT_get_affine_coordinates_GF2m
10253 EC_POINT_set_compressed_coordinates_GF2m
10255 Point compression for binary fields is disabled by default for
10256 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10259 As binary polynomials are represented as BIGNUMs, various members
10260 of the EC_GROUP and EC_POINT data structures can be shared
10261 between the implementations for prime fields and binary fields;
10262 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10263 are essentially identical to their `..._GFp` counterparts.
10264 (For simplicity, the `..._GFp` prefix has been dropped from
10265 various internal method names.)
10267 An internal 'field_div' method (similar to 'field_mul' and
10268 'field_sqr') has been added; this is used only for binary fields.
10270 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10272 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10273 through methods ('mul', 'precompute_mult').
10275 The generic implementations (now internally called 'ec_wNAF_mul'
10276 and 'ec_wNAF_precomputed_mult') remain the default if these
10277 methods are undefined.
10279 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10281 * New function EC_GROUP_get_degree, which is defined through
10282 EC_METHOD. For curves over prime fields, this returns the bit
10283 length of the modulus.
10285 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10287 * New functions EC_GROUP_dup, EC_POINT_dup.
10288 (These simply call ..._new and ..._copy).
10290 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10292 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10293 Polynomials are represented as BIGNUMs (where the sign bit is not
10294 used) in the following functions [macros]:
10297 BN_GF2m_sub [= BN_GF2m_add]
10298 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10299 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10300 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10302 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10303 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10304 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10305 BN_GF2m_cmp [= BN_ucmp]
10307 (Note that only the 'mod' functions are actually for fields GF(2^m).
10308 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10310 For some functions, an the irreducible polynomial defining a
10311 field can be given as an 'unsigned int[]' with strictly
10312 decreasing elements giving the indices of those bits that are set;
10313 i.e., p[] represents the polynomial
10314 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10316 p[0] > p[1] > ... > p[k] = 0.
10317 This applies to the following functions:
10320 BN_GF2m_mod_mul_arr
10321 BN_GF2m_mod_sqr_arr
10322 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10323 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10324 BN_GF2m_mod_exp_arr
10325 BN_GF2m_mod_sqrt_arr
10326 BN_GF2m_mod_solve_quad_arr
10330 Conversion can be performed by the following functions:
10335 bntest.c has additional tests for binary polynomial arithmetic.
10337 Two implementations for BN_GF2m_mod_div() are available.
10338 The default algorithm simply uses BN_GF2m_mod_inv() and
10339 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10340 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10341 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10343 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10345 * Add new error code 'ERR_R_DISABLED' that can be used when some
10346 functionality is disabled at compile-time.
10348 *Douglas Stebila <douglas.stebila@sun.com>*
10350 * Change default behaviour of 'openssl asn1parse' so that more
10351 information is visible when viewing, e.g., a certificate:
10353 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10354 mode the content of non-printable OCTET STRINGs is output in a
10355 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10356 avoid the appearance of a printable string.
10358 *Nils Larsch <nla@trustcenter.de>*
10360 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10362 EC_GROUP_set_asn1_flag()
10363 EC_GROUP_get_asn1_flag()
10364 EC_GROUP_set_point_conversion_form()
10365 EC_GROUP_get_point_conversion_form()
10366 These control ASN1 encoding details:
10367 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10368 has been set to OPENSSL_EC_NAMED_CURVE.
10369 - Points are encoded in uncompressed form by default; options for
10370 asn1_for are as for point2oct, namely
10371 POINT_CONVERSION_COMPRESSED
10372 POINT_CONVERSION_UNCOMPRESSED
10373 POINT_CONVERSION_HYBRID
10375 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10377 EC_GROUP_set_seed()
10378 EC_GROUP_get0_seed()
10379 EC_GROUP_get_seed_len()
10380 This is used only for ASN1 purposes (so far).
10382 *Nils Larsch <nla@trustcenter.de>*
10384 * Add 'field_type' member to EC_METHOD, which holds the NID
10385 of the appropriate field type OID. The new function
10386 EC_METHOD_get_field_type() returns this value.
10388 *Nils Larsch <nla@trustcenter.de>*
10391 EC_POINT_point2bn()
10392 EC_POINT_bn2point()
10393 EC_POINT_point2hex()
10394 EC_POINT_hex2point()
10395 providing useful interfaces to EC_POINT_point2oct() and
10396 EC_POINT_oct2point().
10398 *Nils Larsch <nla@trustcenter.de>*
10400 * Change internals of the EC library so that the functions
10401 EC_GROUP_set_generator()
10402 EC_GROUP_get_generator()
10403 EC_GROUP_get_order()
10404 EC_GROUP_get_cofactor()
10405 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10406 to methods, which would lead to unnecessary code duplication when
10407 adding different types of curves.
10409 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10411 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10412 arithmetic, and such that modified wNAFs are generated
10413 (which avoid length expansion in many cases).
10417 * Add a function EC_GROUP_check_discriminant() (defined via
10418 EC_METHOD) that verifies that the curve discriminant is non-zero.
10420 Add a function EC_GROUP_check() that makes some sanity tests
10421 on a EC_GROUP, its generator and order. This includes
10422 EC_GROUP_check_discriminant().
10424 *Nils Larsch <nla@trustcenter.de>*
10426 * Add ECDSA in new directory crypto/ecdsa/.
10428 Add applications 'openssl ecparam' and 'openssl ecdsa'
10429 (these are based on 'openssl dsaparam' and 'openssl dsa').
10431 ECDSA support is also included in various other files across the
10432 library. Most notably,
10433 - 'openssl req' now has a '-newkey ecdsa:file' option;
10434 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10435 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10436 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10437 them suitable for ECDSA where domain parameters must be
10438 extracted before the specific public key;
10439 - ECDSA engine support has been added.
10441 *Nils Larsch <nla@trustcenter.de>*
10443 * Include some named elliptic curves, and add OIDs from X9.62,
10444 SECG, and WAP/WTLS. Each curve can be obtained from the new
10446 EC_GROUP_new_by_curve_name(),
10447 and the list of available named curves can be obtained with
10448 EC_get_builtin_curves().
10449 Also add a 'curve_name' member to EC_GROUP objects, which can be
10451 EC_GROUP_set_curve_name()
10452 EC_GROUP_get_curve_name()
10454 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10456 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10457 was actually never needed) and in BN_mul(). The removal in BN_mul()
10458 required a small change in bn_mul_part_recursive() and the addition
10459 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10460 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10461 bn_sub_words() and bn_add_words() except they take arrays with
10466 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10468 * Cleanse PEM buffers before freeing them since they may contain
10471 *Benjamin Bennett <ben@psc.edu>*
10473 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10474 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10475 authentication-only ciphersuites.
10479 * Since AES128 and AES256 share a single mask bit in the logic of
10480 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10481 kludge to work properly if AES128 is available and AES256 isn't.
10485 * Expand security boundary to match 1.1.1 module.
10489 * Remove redundant features: hash file source, editing of test vectors
10490 modify fipsld to use external fips_premain.c signature.
10494 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10495 run algorithm test programs.
10499 * Make algorithm test programs more tolerant of whitespace.
10503 * Have SSL/TLS server implementation tolerate "mismatched" record
10504 protocol version while receiving ClientHello even if the
10505 ClientHello is fragmented. (The server can't insist on the
10506 particular protocol version it has chosen before the ServerHello
10507 message has informed the client about his choice.)
10511 * Load error codes if they are not already present instead of using a
10512 static variable. This allows them to be cleanly unloaded and reloaded.
10516 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10518 * Introduce limits to prevent malicious keys being able to
10519 cause a denial of service. [CVE-2006-2940][]
10521 *Steve Henson, Bodo Moeller*
10523 * Fix ASN.1 parsing of certain invalid structures that can result
10524 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10526 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10527 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10529 * Fix SSL client code which could crash if connecting to a
10530 malicious SSLv2 server. [CVE-2006-4343][]
10532 *Tavis Ormandy and Will Drewry, Google Security Team*
10534 * Change ciphersuite string processing so that an explicit
10535 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10536 will no longer include "AES128-SHA"), and any other similar
10537 ciphersuite (same bitmap) from *other* protocol versions (so that
10538 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10539 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10540 changes from 0.9.8b and 0.9.8d.
10544 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10546 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10547 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10549 * Change the Unix randomness entropy gathering to use poll() when
10550 possible instead of select(), since the latter has some
10551 undesirable limitations.
10553 *Darryl Miles via Richard Levitte and Bodo Moeller*
10555 * Disable rogue ciphersuites:
10557 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10558 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10559 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10561 The latter two were purportedly from
10562 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10565 Also deactivate the remaining ciphersuites from
10566 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10567 unofficial, and the ID has long expired.
10571 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10572 dual-core machines) and other potential thread-safety issues.
10576 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10578 * Adapt fipsld and the build system to link against the validated FIPS
10579 module in FIPS mode.
10583 * Fixes for VC++ 2005 build under Windows.
10587 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10588 from a Windows bash shell such as MSYS. It is autodetected from the
10589 "config" script when run from a VC++ environment. Modify standard VC++
10590 build to use fipscanister.o from the GNU make build.
10594 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10596 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10597 The value now differs depending on if you build for FIPS or not.
10598 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10599 safely run with a non-FIPSed libcrypto, as it may crash because of
10600 the difference induced by this change.
10604 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10606 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10607 (part of SSL_OP_ALL). This option used to disable the
10608 countermeasure against man-in-the-middle protocol-version
10609 rollback in the SSL 2.0 server implementation, which is a bad
10610 idea. [CVE-2005-2969][]
10612 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10613 for Information Security, National Institute of Advanced Industrial
10614 Science and Technology [AIST, Japan)]*
10616 * Minimal support for X9.31 signatures and PSS padding modes. This is
10617 mainly for FIPS compliance and not fully integrated at this stage.
10621 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10622 the exponentiation using a fixed-length exponent. (Otherwise,
10623 the information leaked through timing could expose the secret key
10624 after many signatures; cf. Bleichenbacher's attack on DSA with
10629 * Make a new fixed-window mod_exp implementation the default for
10630 RSA, DSA, and DH private-key operations so that the sequence of
10631 squares and multiplies and the memory access pattern are
10632 independent of the particular secret key. This will mitigate
10633 cache-timing and potential related attacks.
10635 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10636 and this is automatically used by BN_mod_exp_mont() if the new flag
10637 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10638 will use this BN flag for private exponents unless the flag
10639 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10640 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10642 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10644 * Change the client implementation for SSLv23_method() and
10645 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10646 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10647 (Previously, the SSL 2.0 backwards compatible Client Hello
10648 message format would be used even with SSL_OP_NO_SSLv2.)
10652 * Add support for smime-type MIME parameter in S/MIME messages which some
10657 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10658 a threadsafe manner. Modify rsa code to use new function and add calls
10659 to dsa and dh code (which had race conditions before).
10663 * Include the fixed error library code in the C error file definitions
10664 instead of fixing them up at runtime. This keeps the error code
10665 structures constant.
10669 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10671 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10674 * Fixes for newer kerberos headers. NB: the casts are needed because
10675 the 'length' field is signed on one version and unsigned on another
10676 with no (?) obvious way to tell the difference, without these VC++
10677 complains. Also the "definition" of FAR (blank) is no longer included
10678 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10679 some needed definitions.
10683 * Undo Cygwin change.
10687 * Added support for proxy certificates according to RFC 3820.
10688 Because they may be a security thread to unaware applications,
10689 they must be explicitly allowed in run-time. See
10690 docs/HOWTO/proxy_certificates.txt for further information.
10694 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10696 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10697 server and client random values. Previously
10698 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10699 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10701 This change has negligible security impact because:
10703 1. Server and client random values still have 24 bytes of pseudo random
10706 2. Server and client random values are sent in the clear in the initial
10709 3. The master secret is derived using the premaster secret (48 bytes in
10710 size for static RSA ciphersuites) as well as client server and random
10713 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10716 *Stephen Henson, reported by UK NISCC*
10718 * Use Windows randomness collection on Cygwin.
10722 * Fix hang in EGD/PRNGD query when communication socket is closed
10723 prematurely by EGD/PRNGD.
10725 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10727 * Prompt for pass phrases when appropriate for PKCS12 input format.
10731 * Back-port of selected performance improvements from development
10732 branch, as well as improved support for PowerPC platforms.
10736 * Add lots of checks for memory allocation failure, error codes to indicate
10737 failure and freeing up memory if a failure occurs.
10739 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10741 * Add new -passin argument to dgst.
10745 * Perform some character comparisons of different types in X509_NAME_cmp:
10746 this is needed for some certificates that re-encode DNs into UTF8Strings
10747 (in violation of RFC3280) and can't or won't issue name rollover
10752 * Make an explicit check during certificate validation to see that
10753 the CA setting in each certificate on the chain is correct. As a
10754 side effect always do the following basic checks on extensions,
10755 not just when there's an associated purpose to the check:
10757 - if there is an unhandled critical extension (unless the user
10758 has chosen to ignore this fault)
10759 - if the path length has been exceeded (if one is set at all)
10760 - that certain extensions fit the associated purpose (if one has
10765 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10767 * Avoid a race condition when CRLs are checked in a multi threaded
10768 environment. This would happen due to the reordering of the revoked
10769 entries during signature checking and serial number lookup. Now the
10770 encoding is cached and the serial number sort performed under a lock.
10771 Add new STACK function sk_is_sorted().
10775 * Add Delta CRL to the extension code.
10779 * Various fixes to s3_pkt.c so alerts are sent properly.
10781 *David Holmes <d.holmes@f5.com>*
10783 * Reduce the chances of duplicate issuer name and serial numbers (in
10784 violation of RFC3280) using the OpenSSL certificate creation utilities.
10785 This is done by creating a random 64 bit value for the initial serial
10786 number when a serial number file is created or when a self signed
10787 certificate is created using 'openssl req -x509'. The initial serial
10788 number file is created using 'openssl x509 -next_serial' in CA.pl
10789 rather than being initialized to 1.
10793 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10795 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10796 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10798 *Joe Orton, Steve Henson*
10800 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10803 *Joe Orton, Steve Henson*
10805 * Make it possible to have multiple active certificates with the same
10806 subject in the CA index file. This is done only if the keyword
10807 'unique_subject' is set to 'no' in the main CA section (default
10808 if 'CA_default') of the configuration file. The value is saved
10809 with the database itself in a separate index attribute file,
10810 named like the index file with '.attr' appended to the name.
10814 * X509 verify fixes. Disable broken certificate workarounds when
10815 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10816 keyUsage extension present. Don't accept CRLs with unhandled critical
10817 extensions: since verify currently doesn't process CRL extensions this
10818 rejects a CRL with *any* critical extensions. Add new verify error codes
10823 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10824 A clarification of RFC2560 will require the use of OCTET STRINGs and
10825 some implementations cannot handle the current raw format. Since OpenSSL
10826 copies and compares OCSP nonces as opaque blobs without any attempt at
10827 parsing them this should not create any compatibility issues.
10831 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10832 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10833 this HMAC (and other) operations are several times slower than OpenSSL
10838 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10840 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10842 * Use the correct content when signing type "other".
10846 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10848 * Fix various bugs revealed by running the NISCC test suite:
10850 Stop out of bounds reads in the ASN1 code when presented with
10851 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10853 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10855 If verify callback ignores invalid public key errors don't try to check
10856 certificate signature with the NULL public key.
10860 * New -ignore_err option in ocsp application to stop the server
10861 exiting on the first error in a request.
10865 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10866 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10871 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10872 extra data after the compression methods not only for TLS 1.0
10873 but also for SSL 3.0 (as required by the specification).
10875 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10877 * Change X509_certificate_type() to mark the key as exported/exportable
10878 when it's 512 *bits* long, not 512 bytes.
10882 * Change AES_cbc_encrypt() so it outputs exact multiple of
10883 blocks during encryption.
10887 * Various fixes to base64 BIO and non blocking I/O. On write
10888 flushes were not handled properly if the BIO retried. On read
10889 data was not being buffered properly and had various logic bugs.
10890 This also affects blocking I/O when the data being decoded is a
10895 * Various S/MIME bugfixes and compatibility changes:
10896 output correct application/pkcs7 MIME type if
10897 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10898 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10899 of files as .eml work). Correctly handle very long lines in MIME
10904 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10906 * Countermeasure against the Klima-Pokorny-Rosa extension of
10907 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10908 a protocol version number mismatch like a decryption error
10909 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10913 * Turn on RSA blinding by default in the default implementation
10914 to avoid a timing attack. Applications that don't want it can call
10915 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10916 They would be ill-advised to do so in most cases.
10918 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10920 * Change RSA blinding code so that it works when the PRNG is not
10921 seeded (in this case, the secret RSA exponent is abused as
10922 an unpredictable seed -- if it is not unpredictable, there
10923 is no point in blinding anyway). Make RSA blinding thread-safe
10924 by remembering the creator's thread ID in rsa->blinding and
10925 having all other threads use local one-time blinding factors
10926 (this requires more computation than sharing rsa->blinding, but
10927 avoids excessive locking; and if an RSA object is not shared
10928 between threads, blinding will still be very fast).
10932 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10933 ENGINE as defaults for all supported algorithms irrespective of
10934 the 'flags' parameter. 'flags' is now honoured, so applications
10935 should make sure they are passing it correctly.
10939 * Target "mingw" now allows native Windows code to be generated in
10940 the Cygwin environment as well as with the MinGW compiler.
10944 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10946 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10947 via timing by performing a MAC computation even if incorrect
10948 block cipher padding has been found. This is a countermeasure
10949 against active attacks where the attacker has to distinguish
10950 between bad padding and a MAC verification error. [CVE-2003-0078][]
10952 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10953 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10954 Martin Vuagnoux (EPFL, Ilion)*
10956 * Make the no-err option work as intended. The intention with no-err
10957 is not to have the whole error stack handling routines removed from
10958 libcrypto, it's only intended to remove all the function name and
10959 reason texts, thereby removing some of the footprint that may not
10960 be interesting if those errors aren't displayed anyway.
10962 NOTE: it's still possible for any application or module to have its
10963 own set of error texts inserted. The routines are there, just not
10964 used by default when no-err is given.
10968 * Add support for FreeBSD on IA64.
10970 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10972 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10973 Kerberos function mit_des_cbc_cksum(). Before this change,
10974 the value returned by DES_cbc_cksum() was like the one from
10975 mit_des_cbc_cksum(), except the bytes were swapped.
10977 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10979 * Allow an application to disable the automatic SSL chain building.
10980 Before this a rather primitive chain build was always performed in
10981 ssl3_output_cert_chain(): an application had no way to send the
10982 correct chain if the automatic operation produced an incorrect result.
10984 Now the chain builder is disabled if either:
10986 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10988 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10990 The reasoning behind this is that an application would not want the
10991 auto chain building to take place if extra chain certificates are
10992 present and it might also want a means of sending no additional
10993 certificates (for example the chain has two certificates and the
10998 * Add the possibility to build without the ENGINE framework.
11000 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11002 * Under Win32 gmtime() can return NULL: check return value in
11003 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11007 * DSA routines: under certain error conditions uninitialized BN objects
11008 could be freed. Solution: make sure initialization is performed early
11009 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11010 Nils Larsch <nla@trustcenter.de> via PR#459)
11014 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11015 checked on reconnect on the client side, therefore session resumption
11016 could still fail with a "ssl session id is different" error. This
11017 behaviour is masked when SSL_OP_ALL is used due to
11018 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11019 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11020 followup to PR #377.
11024 * IA-32 assembler support enhancements: unified ELF targets, support
11025 for SCO/Caldera platforms, fix for Cygwin shared build.
11029 * Add support for FreeBSD on sparc64. As a consequence, support for
11030 FreeBSD on non-x86 processors is separate from x86 processors on
11031 the config script, much like the NetBSD support.
11033 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11035 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11037 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11040 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11041 code (06) was taken as the first octet of the session ID and the last
11042 octet was ignored consequently. As a result SSLv2 client side session
11043 caching could not have worked due to the session ID mismatch between
11045 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11050 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11051 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11056 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11057 seems that in spite of existing for more than a year, many application
11058 author have done nothing to provide the necessary callbacks, which
11059 means that this particular engine will not work properly anywhere.
11060 This is a very unfortunate situation which forces us, in the name
11061 of usability, to give the hw_ncipher.c a static lock, which is part
11063 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11064 appear in 0.9.8 or later. We EXPECT application authors to have
11065 dealt properly with this when 0.9.8 is released (unless we actually
11066 make such changes in the libcrypto locking code that changes will
11067 have to be made anyway).
11071 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11072 octets have been read, EOF or an error occurs. Without this change
11073 some truncated ASN1 structures will not produce an error.
11077 * Disable Heimdal support, since it hasn't been fully implemented.
11078 Still give the possibility to force the use of Heimdal, but with
11079 warnings and a request that patches get sent to openssl-dev.
11083 * Add the VC-CE target, introduce the WINCE sysname, and add
11084 INSTALL.WCE and appropriate conditionals to make it build.
11086 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11088 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11089 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11090 edit numbers of the version.
11092 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11094 * Introduce safe string copy and catenation functions
11095 (BUF_strlcpy() and BUF_strlcat()).
11097 *Ben Laurie (CHATS) and Richard Levitte*
11099 * Avoid using fixed-size buffers for one-line DNs.
11101 *Ben Laurie (CHATS)*
11103 * Add BUF_MEM_grow_clean() to avoid information leakage when
11104 resizing buffers containing secrets, and use where appropriate.
11106 *Ben Laurie (CHATS)*
11108 * Avoid using fixed size buffers for configuration file location.
11110 *Ben Laurie (CHATS)*
11112 * Avoid filename truncation for various CA files.
11114 *Ben Laurie (CHATS)*
11116 * Use sizeof in preference to magic numbers.
11118 *Ben Laurie (CHATS)*
11120 * Avoid filename truncation in cert requests.
11122 *Ben Laurie (CHATS)*
11124 * Add assertions to check for (supposedly impossible) buffer
11127 *Ben Laurie (CHATS)*
11129 * Don't cache truncated DNS entries in the local cache (this could
11130 potentially lead to a spoofing attack).
11132 *Ben Laurie (CHATS)*
11134 * Fix various buffers to be large enough for hex/decimal
11135 representations in a platform independent manner.
11137 *Ben Laurie (CHATS)*
11139 * Add CRYPTO_realloc_clean() to avoid information leakage when
11140 resizing buffers containing secrets, and use where appropriate.
11142 *Ben Laurie (CHATS)*
11144 * Add BIO_indent() to avoid much slightly worrying code to do
11147 *Ben Laurie (CHATS)*
11149 * Convert sprintf()/BIO_puts() to BIO_printf().
11151 *Ben Laurie (CHATS)*
11153 * buffer_gets() could terminate with the buffer only half
11156 *Ben Laurie (CHATS)*
11158 * Add assertions to prevent user-supplied crypto functions from
11159 overflowing internal buffers by having large block sizes, etc.
11161 *Ben Laurie (CHATS)*
11163 * New OPENSSL_assert() macro (similar to assert(), but enabled
11166 *Ben Laurie (CHATS)*
11168 * Eliminate unused copy of key in RC4.
11170 *Ben Laurie (CHATS)*
11172 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11174 *Ben Laurie (CHATS)*
11176 * Fix off-by-one error in EGD path.
11178 *Ben Laurie (CHATS)*
11180 * If RANDFILE path is too long, ignore instead of truncating.
11182 *Ben Laurie (CHATS)*
11184 * Eliminate unused and incorrectly sized X.509 structure
11187 *Ben Laurie (CHATS)*
11189 * Eliminate unused and dangerous function knumber().
11191 *Ben Laurie (CHATS)*
11193 * Eliminate unused and dangerous structure, KSSL_ERR.
11195 *Ben Laurie (CHATS)*
11197 * Protect against overlong session ID context length in an encoded
11198 session object. Since these are local, this does not appear to be
11201 *Ben Laurie (CHATS)*
11203 * Change from security patch (see 0.9.6e below) that did not affect
11204 the 0.9.6 release series:
11206 Remote buffer overflow in SSL3 protocol - an attacker could
11207 supply an oversized master key in Kerberos-enabled versions.
11210 *Ben Laurie (CHATS)*
11212 * Change the SSL kerb5 codes to match RFC 2712.
11216 * Make -nameopt work fully for req and add -reqopt switch.
11218 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11220 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11222 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11224 * Make sure tests can be performed even if the corresponding algorithms
11225 have been removed entirely. This was also the last step to make
11226 OpenSSL compilable with DJGPP under all reasonable conditions.
11228 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11230 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11231 to allow version independent disabling of normally unselected ciphers,
11232 which may be activated as a side-effect of selecting a single cipher.
11234 (E.g., cipher list string "RSA" enables ciphersuites that are left
11235 out of "ALL" because they do not provide symmetric encryption.
11236 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11238 *Lutz Jaenicke, Bodo Moeller*
11240 * Add appropriate support for separate platform-dependent build
11241 directories. The recommended way to make a platform-dependent
11242 build directory is the following (tested on Linux), maybe with
11245 # Place yourself outside of the OpenSSL source tree. In
11246 # this example, the environment variable OPENSSL_SOURCE
11247 # is assumed to contain the absolute OpenSSL source directory.
11248 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11249 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11250 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11251 mkdir -p `dirname $F`
11252 ln -s $OPENSSL_SOURCE/$F $F
11255 To be absolutely sure not to disturb the source tree, a "make clean"
11256 is a good thing. If it isn't successful, don't worry about it,
11257 it probably means the source directory is very clean.
11261 * Make sure any ENGINE control commands make local copies of string
11262 pointers passed to them whenever necessary. Otherwise it is possible
11263 the caller may have overwritten (or deallocated) the original string
11264 data when a later ENGINE operation tries to use the stored values.
11266 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11268 * Improve diagnostics in file reading and command-line digests.
11270 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11272 * Add AES modes CFB and OFB to the object database. Correct an
11273 error in AES-CFB decryption.
11277 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11278 allows existing EVP_CIPHER_CTX structures to be reused after
11279 calling `EVP_*Final()`. This behaviour is used by encryption
11280 BIOs and some applications. This has the side effect that
11281 applications must explicitly clean up cipher contexts with
11282 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11286 * Check the values of dna and dnb in bn_mul_recursive before calling
11287 bn_mul_comba (a non zero value means the a or b arrays do not contain
11288 n2 elements) and fallback to bn_mul_normal if either is not zero.
11292 * Fix escaping of non-ASCII characters when using the -subj option
11293 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11297 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11298 form for "surname", serialNumber has no short form.
11299 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11300 therefore remove "mail" short name for "internet 7".
11301 The OID for unique identifiers in X509 certificates is
11302 x500UniqueIdentifier, not uniqueIdentifier.
11303 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11307 * Add an "init" command to the ENGINE config module and auto initialize
11308 ENGINEs. Without any "init" command the ENGINE will be initialized
11309 after all ctrl commands have been executed on it. If init=1 the
11310 ENGINE is initialized at that point (ctrls before that point are run
11311 on the uninitialized ENGINE and after on the initialized one). If
11312 init=0 then the ENGINE will not be initialized at all.
11316 * Fix the 'app_verify_callback' interface so that the user-defined
11317 argument is actually passed to the callback: In the
11318 SSL_CTX_set_cert_verify_callback() prototype, the callback
11319 declaration has been changed from
11322 int (*cb)(X509_STORE_CTX *,void *);
11323 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11324 i=s->ctx->app_verify_callback(&ctx)
11325 has been changed into
11326 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11328 To update applications using SSL_CTX_set_cert_verify_callback(),
11329 a dummy argument can be added to their callback functions.
11331 *D. K. Smetters <smetters@parc.xerox.com>*
11333 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11335 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11337 * Add and OPENSSL_LOAD_CONF define which will cause
11338 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11339 This allows older applications to transparently support certain
11340 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11341 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11342 load the config file and OPENSSL_add_all_algorithms_conf() which will
11343 always load it have also been added.
11347 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11348 Adjust NIDs and EVP layer.
11350 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11352 * Config modules support in openssl utility.
11354 Most commands now load modules from the config file,
11355 though in a few (such as version) this isn't done
11356 because it couldn't be used for anything.
11358 In the case of ca and req the config file used is
11359 the same as the utility itself: that is the -config
11360 command line option can be used to specify an
11365 * Move default behaviour from OPENSSL_config(). If appname is NULL
11366 use "openssl_conf" if filename is NULL use default openssl config file.
11370 * Add an argument to OPENSSL_config() to allow the use of an alternative
11371 config section name. Add a new flag to tolerate a missing config file
11372 and move code to CONF_modules_load_file().
11376 * Support for crypto accelerator cards from Accelerated Encryption
11377 Processing, www.aep.ie. (Use engine 'aep')
11378 The support was copied from 0.9.6c [engine] and adapted/corrected
11379 to work with the new engine framework.
11381 *AEP Inc. and Richard Levitte*
11383 * Support for SureWare crypto accelerator cards from Baltimore
11384 Technologies. (Use engine 'sureware')
11385 The support was copied from 0.9.6c [engine] and adapted
11386 to work with the new engine framework.
11390 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11391 make the newer ENGINE framework commands for the CHIL engine work.
11393 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11395 * Make it possible to produce shared libraries on ReliantUNIX.
11397 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11399 * Add the configuration target debug-linux-ppro.
11400 Make 'openssl rsa' use the general key loading routines
11401 implemented in apps.c, and make those routines able to
11402 handle the key format FORMAT_NETSCAPE and the variant
11405 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11407 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11409 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11411 * Add -keyform to rsautl, and document -engine.
11413 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11415 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11416 BIO_R_NO_SUCH_FILE error code rather than the generic
11417 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11421 * Add new functions
11422 ERR_peek_last_error
11423 ERR_peek_last_error_line
11424 ERR_peek_last_error_line_data.
11425 These are similar to
11427 ERR_peek_error_line
11428 ERR_peek_error_line_data,
11429 but report on the latest error recorded rather than the first one
11430 still in the error queue.
11432 *Ben Laurie, Bodo Moeller*
11434 * default_algorithms option in ENGINE config module. This allows things
11436 default_algorithms = ALL
11437 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11441 * Preliminary ENGINE config module.
11445 * New experimental application configuration code.
11449 * Change the AES code to follow the same name structure as all other
11450 symmetric ciphers, and behave the same way. Move everything to
11451 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11453 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11455 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11457 *Ben Laurie and Theo de Raadt*
11459 * Add option to output public keys in req command.
11461 *Massimiliano Pala madwolf@openca.org*
11463 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11464 (up to about 10% better than before for P-192 and P-224).
11468 * New functions/macros
11470 SSL_CTX_set_msg_callback(ctx, cb)
11471 SSL_CTX_set_msg_callback_arg(ctx, arg)
11472 SSL_set_msg_callback(ssl, cb)
11473 SSL_set_msg_callback_arg(ssl, arg)
11475 to request calling a callback function
11477 void cb(int write_p, int version, int content_type,
11478 const void *buf, size_t len, SSL *ssl, void *arg)
11480 whenever a protocol message has been completely received
11481 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11482 protocol version according to which the SSL library interprets
11483 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11484 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11485 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11486 specification (change_cipher_spec(20), alert(21), handshake(22)).
11487 'buf' and 'len' point to the actual message, 'ssl' to the
11488 SSL object, and 'arg' is the application-defined value set by
11489 SSL[_CTX]_set_msg_callback_arg().
11491 'openssl s_client' and 'openssl s_server' have new '-msg' options
11492 to enable a callback that displays all protocol messages.
11496 * Change the shared library support so shared libraries are built as
11497 soon as the corresponding static library is finished, and thereby get
11498 openssl and the test programs linked against the shared library.
11499 This still only happens when the keyword "shard" has been given to
11500 the configuration scripts.
11502 NOTE: shared library support is still an experimental thing, and
11503 backward binary compatibility is still not guaranteed.
11505 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11507 * Add support for Subject Information Access extension.
11509 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11511 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11512 additional bytes when new memory had to be allocated, not just
11513 when reusing an existing buffer.
11517 * New command line and configuration option 'utf8' for the req command.
11518 This allows field values to be specified as UTF8 strings.
11522 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11523 runs for the former and machine-readable output for the latter.
11527 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11528 of the e-mail address in the DN (i.e., it will go into a certificate
11529 extension only). The new configuration file option 'email_in_dn = no'
11530 has the same effect.
11532 *Massimiliano Pala madwolf@openca.org*
11534 * Change all functions with names starting with `des_` to be starting
11535 with `DES_` instead. Add wrappers that are compatible with libdes,
11536 but are named `_ossl_old_des_*`. Finally, add macros that map the
11537 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11538 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11539 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11542 Since we provide two compatibility mappings, the user needs to
11543 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11544 compatibility is desired. The default (i.e., when that macro
11545 isn't defined) is OpenSSL 0.9.6c compatibility.
11547 There are also macros that enable and disable the support of old
11548 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11549 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11550 are defined, the default will apply: to support the old des routines.
11552 In either case, one must include openssl/des.h to get the correct
11553 definitions. Do not try to just include openssl/des_old.h, that
11556 NOTE: This is a major break of an old API into a new one. Software
11557 authors are encouraged to switch to the `DES_` style functions. Some
11558 time in the future, des_old.h and the libdes compatibility functions
11559 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11560 default), and then completely removed.
11564 * Test for certificates which contain unsupported critical extensions.
11565 If such a certificate is found during a verify operation it is
11566 rejected by default: this behaviour can be overridden by either
11567 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11568 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11569 X509_supported_extension() has also been added which returns 1 if a
11570 particular extension is supported.
11574 * Modify the behaviour of EVP cipher functions in similar way to digests
11575 to retain compatibility with existing code.
11579 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11580 compatibility with existing code. In particular the 'ctx' parameter does
11581 not have to be to be initialized before the call to EVP_DigestInit() and
11582 it is tidied up after a call to EVP_DigestFinal(). New function
11583 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11584 EVP_MD_CTX_copy() changed to not require the destination to be
11585 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11586 requires the destination to be valid.
11588 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11589 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11593 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11594 so that complete 'Handshake' protocol structures are kept in memory
11595 instead of overwriting 'msg_type' and 'length' with 'body' data.
11599 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11601 *Massimo Santin via Richard Levitte*
11603 * Major restructuring to the underlying ENGINE code. This includes
11604 reduction of linker bloat, separation of pure "ENGINE" manipulation
11605 (initialisation, etc) from functionality dealing with implementations
11606 of specific crypto interfaces. This change also introduces integrated
11607 support for symmetric ciphers and digest implementations - so ENGINEs
11608 can now accelerate these by providing EVP_CIPHER and EVP_MD
11609 implementations of their own. This is detailed in
11610 [crypto/engine/README.md](crypto/engine/README.md)
11611 as it couldn't be adequately described here. However, there are a few
11612 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11613 were changed in the original introduction of ENGINE code have now
11614 reverted back - the hooking from this code to ENGINE is now a good
11615 deal more passive and at run-time, operations deal directly with
11616 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11617 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11618 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11619 they were not being used by the framework as there is no concept of a
11620 BIGNUM_METHOD and they could not be generalised to the new
11621 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11622 ENGINE_cpy() has been removed as it cannot be consistently defined in
11627 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11631 * Change mkdef.pl to sort symbols that get the same entry number,
11632 and make sure the automatically generated functions `ERR_load_*`
11633 become part of libeay.num as well.
11637 * New function SSL_renegotiate_pending(). This returns true once
11638 renegotiation has been requested (either SSL_renegotiate() call
11639 or HelloRequest/ClientHello received from the peer) and becomes
11640 false once a handshake has been completed.
11641 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11642 sends a HelloRequest, but does not ensure that a handshake takes
11643 place. SSL_renegotiate_pending() is useful for checking if the
11644 client has followed the request.)
11648 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11649 By default, clients may request session resumption even during
11650 renegotiation (if session ID contexts permit); with this option,
11651 session resumption is possible only in the first handshake.
11653 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11654 more bits available for options that should not be part of
11655 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11659 * Add some demos for certificate and certificate request creation.
11663 * Make maximum certificate chain size accepted from the peer application
11664 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11665 "Douglas E. Engert" <deengert@anl.gov>.
11669 * Add support for shared libraries for Unixware-7
11670 (Boyd Lynn Gerber <gerberb@zenez.com>).
11674 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11675 be done prior to destruction. Use this to unload error strings from
11676 ENGINEs that load their own error strings. NB: This adds two new API
11677 functions to "get" and "set" this destroy handler in an ENGINE.
11681 * Alter all existing ENGINE implementations (except "openssl" and
11682 "openbsd") to dynamically instantiate their own error strings. This
11683 makes them more flexible to be built both as statically-linked ENGINEs
11684 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11685 Also, add stub code to each that makes building them as self-contained
11686 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11690 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11691 implementations into applications that are completely implemented in
11692 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11693 commands that can be used to configure what shared-library to load and
11694 to control aspects of the way it is handled. Also, made an update to
11695 the [README-Engine.md](README-Engine.md) file
11696 that brings its information up-to-date and
11697 provides some information and instructions on the "dynamic" ENGINE
11698 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11702 * Make it possible to unload ranges of ERR strings with a new
11703 "ERR_unload_strings" function.
11707 * Add a copy() function to EVP_MD.
11711 * Make EVP_MD routines take a context pointer instead of just the
11712 md_data void pointer.
11716 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11717 that the digest can only process a single chunk of data
11718 (typically because it is provided by a piece of
11719 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11720 is only going to provide a single chunk of data, and hence the
11721 framework needn't accumulate the data for oneshot drivers.
11725 * As with "ERR", make it possible to replace the underlying "ex_data"
11726 functions. This change also alters the storage and management of global
11727 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11728 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11729 index counters. The API functions that use this state have been changed
11730 to take a "class_index" rather than pointers to the class's local STACK
11731 and counter, and there is now an API function to dynamically create new
11732 classes. This centralisation allows us to (a) plug a lot of the
11733 thread-safety problems that existed, and (b) makes it possible to clean
11734 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11735 such data would previously have always leaked in application code and
11736 workarounds were in place to make the memory debugging turn a blind eye
11737 to it. Application code that doesn't use this new function will still
11738 leak as before, but their memory debugging output will announce it now
11739 rather than letting it slide.
11741 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11742 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11743 has a return value to indicate success or failure.
11747 * Make it possible to replace the underlying "ERR" functions such that the
11748 global state (2 LHASH tables and 2 locks) is only used by the "default"
11749 implementation. This change also adds two functions to "get" and "set"
11750 the implementation prior to it being automatically set the first time
11751 any other ERR function takes place. Ie. an application can call "get",
11752 pass the return value to a module it has just loaded, and that module
11753 can call its own "set" function using that value. This means the
11754 module's "ERR" operations will use (and modify) the error state in the
11755 application and not in its own statically linked copy of OpenSSL code.
11759 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11760 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11761 the operation, and provides a more encapsulated way for external code
11762 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11763 to use these functions rather than manually incrementing the counts.
11765 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11769 * Add EVP test program.
11773 * Add symmetric cipher support to ENGINE. Expect the API to change!
11777 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11778 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11779 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11780 These allow a CRL to be built without having to access X509_CRL fields
11781 directly. Modify 'ca' application to use new functions.
11785 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11786 bug workarounds. Rollback attack detection is a security feature.
11787 The problem will only arise on OpenSSL servers when TLSv1 is not
11788 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11789 Software authors not wanting to support TLSv1 will have special reasons
11790 for their choice and can explicitly enable this option.
11792 *Bodo Moeller, Lutz Jaenicke*
11794 * Rationalise EVP so it can be extended: don't include a union of
11795 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11796 (similar to those existing for EVP_CIPHER_CTX).
11801 EVP_MD_CTX_init(&md); /* new function call */
11802 EVP_DigestInit(&md, EVP_sha1());
11803 EVP_DigestUpdate(&md, in, len);
11804 EVP_DigestFinal(&md, out, NULL);
11805 EVP_MD_CTX_cleanup(&md); /* new function call */
11809 * Make DES key schedule conform to the usual scheme, as well as
11810 correcting its structure. This means that calls to DES functions
11811 now have to pass a pointer to a des_key_schedule instead of a
11812 plain des_key_schedule (which was actually always a pointer
11815 des_key_schedule ks;
11817 des_set_key_checked(..., &ks);
11818 des_ncbc_encrypt(..., &ks, ...);
11820 (Note that a later change renames 'des_...' into 'DES_...'.)
11824 * Initial reduction of linker bloat: the use of some functions, such as
11825 PEM causes large amounts of unused functions to be linked in due to
11826 poor organisation. For example pem_all.c contains every PEM function
11827 which has a knock on effect of linking in large amounts of (unused)
11828 ASN1 code. Grouping together similar functions and splitting unrelated
11829 functions prevents this.
11833 * Cleanup of EVP macros.
11837 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11838 correct `_ecb suffix`.
11842 * Add initial OCSP responder support to ocsp application. The
11843 revocation information is handled using the text based index
11844 use by the ca application. The responder can either handle
11845 requests generated internally, supplied in files (for example
11846 via a CGI script) or using an internal minimal server.
11850 * Add configuration choices to get zlib compression for TLS.
11854 * Changes to Kerberos SSL for RFC 2712 compliance:
11855 1. Implemented real KerberosWrapper, instead of just using
11856 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11857 2. Implemented optional authenticator field of KerberosWrapper.
11859 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11860 and authenticator structs; see crypto/krb5/.
11862 Generalized Kerberos calls to support multiple Kerberos libraries.
11863 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11864 via Richard Levitte*
11866 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11867 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11868 values for each of the key sizes rather than having just
11869 parameters (and 'speed' generating keys each time).
11873 * Speed up EVP routines.
11876 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11877 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11878 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11879 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11881 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11882 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11883 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11886 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11888 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11892 * Added the OS2-EMX target.
11894 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11896 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11897 to support NCONF routines in extension code. New function CONF_set_nconf()
11898 to allow functions which take an NCONF to also handle the old LHASH
11899 structure: this means that the old CONF compatible routines can be
11900 retained (in particular wrt extensions) without having to duplicate the
11901 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11905 * Enhance the general user interface with mechanisms for inner control
11906 and with possibilities to have yes/no kind of prompts.
11910 * Change all calls to low level digest routines in the library and
11911 applications to use EVP. Add missing calls to HMAC_cleanup() and
11912 don't assume HMAC_CTX can be copied using memcpy().
11914 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11916 * Add the possibility to control engines through control names but with
11917 arbitrary arguments instead of just a string.
11918 Change the key loaders to take a UI_METHOD instead of a callback
11919 function pointer. NOTE: this breaks binary compatibility with earlier
11920 versions of OpenSSL [engine].
11921 Adapt the nCipher code for these new conditions and add a card insertion
11926 * Enhance the general user interface with mechanisms to better support
11927 dialog box interfaces, application-defined prompts, the possibility
11928 to use defaults (for example default passwords from somewhere else)
11929 and interrupts/cancellations.
11933 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11934 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11938 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11939 tidy up some unnecessarily weird code in 'sk_new()').
11941 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11943 * Change the key loading routines for ENGINEs to use the same kind
11944 callback (pem_password_cb) as all other routines that need this
11949 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11950 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11951 than this minimum value is recommended.
11955 * New random seeder for OpenVMS, using the system process statistics
11956 that are easily reachable.
11960 * Windows apparently can't transparently handle global
11961 variables defined in DLLs. Initialisations such as:
11963 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11965 won't compile. This is used by the any applications that need to
11966 declare their own ASN1 modules. This was fixed by adding the option
11967 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11968 needed for static libraries under Win32.
11972 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11973 setting of purpose and trust fields. New X509_STORE trust and
11974 purpose functions and tidy up setting in other SSL functions.
11978 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11979 structure. These are inherited by X509_STORE_CTX when it is
11980 initialised. This allows various defaults to be set in the
11981 X509_STORE structure (such as flags for CRL checking and custom
11982 purpose or trust settings) for functions which only use X509_STORE_CTX
11983 internally such as S/MIME.
11985 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11986 trust settings if they are not set in X509_STORE. This allows X509_STORE
11987 purposes and trust (in S/MIME for example) to override any set by default.
11989 Add command line options for CRL checking to smime, s_client and s_server
11994 * Initial CRL based revocation checking. If the CRL checking flag(s)
11995 are set then the CRL is looked up in the X509_STORE structure and
11996 its validity and signature checked, then if the certificate is found
11997 in the CRL the verify fails with a revoked error.
11999 Various new CRL related callbacks added to X509_STORE_CTX structure.
12001 Command line options added to 'verify' application to support this.
12003 This needs some additional work, such as being able to handle multiple
12004 CRLs with different times, extension based lookup (rather than just
12005 by subject name) and ultimately more complete V2 CRL extension
12010 * Add a general user interface API (crypto/ui/). This is designed
12011 to replace things like des_read_password and friends (backward
12012 compatibility functions using this new API are provided).
12013 The purpose is to remove prompting functions from the DES code
12014 section as well as provide for prompting through dialog boxes in
12015 a window system and the like.
12019 * Add "ex_data" support to ENGINE so implementations can add state at a
12020 per-structure level rather than having to store it globally.
12024 * Make it possible for ENGINE structures to be copied when retrieved by
12025 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12026 This causes the "original" ENGINE structure to act like a template,
12027 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12028 operational state can be localised to each ENGINE structure, despite the
12029 fact they all share the same "methods". New ENGINE structures returned in
12030 this case have no functional references and the return value is the single
12031 structural reference. This matches the single structural reference returned
12032 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12037 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12038 needs to match any other type at all we need to manually clear the
12043 * Changes to the "openssl engine" utility to include;
12044 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12045 about an ENGINE's available control commands.
12046 - executing control commands from command line arguments using the
12047 '-pre' and '-post' switches. '-post' is only used if '-t' is
12048 specified and the ENGINE is successfully initialised. The syntax for
12049 the individual commands are colon-separated, for example;
12050 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12054 * New dynamic control command support for ENGINEs. ENGINEs can now
12055 declare their own commands (numbers), names (strings), descriptions,
12056 and input types for run-time discovery by calling applications. A
12057 subset of these commands are implicitly classed as "executable"
12058 depending on their input type, and only these can be invoked through
12059 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12060 can be based on user input, config files, etc). The distinction is
12061 that "executable" commands cannot return anything other than a boolean
12062 result and can only support numeric or string input, whereas some
12063 discoverable commands may only be for direct use through
12064 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12065 pointers, or other custom uses. The "executable" commands are to
12066 support parameterisations of ENGINE behaviour that can be
12067 unambiguously defined by ENGINEs and used consistently across any
12068 OpenSSL-based application. Commands have been added to all the
12069 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12070 control over shared-library paths without source code alterations.
12074 * Changed all ENGINE implementations to dynamically allocate their
12075 ENGINEs rather than declaring them statically. Apart from this being
12076 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12077 this also allows the implementations to compile without using the
12078 internal engine_int.h header.
12082 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12083 'const' value. Any code that should be able to modify a RAND_METHOD
12084 should already have non-const pointers to it (ie. they should only
12085 modify their own ones).
12089 * Made a variety of little tweaks to the ENGINE code.
12090 - "atalla" and "ubsec" string definitions were moved from header files
12091 to C code. "nuron" string definitions were placed in variables
12092 rather than hard-coded - allowing parameterisation of these values
12093 later on via ctrl() commands.
12094 - Removed unused "#if 0"'d code.
12095 - Fixed engine list iteration code so it uses ENGINE_free() to release
12096 structural references.
12097 - Constified the RAND_METHOD element of ENGINE structures.
12098 - Constified various get/set functions as appropriate and added
12099 missing functions (including a catch-all ENGINE_cpy that duplicates
12100 all ENGINE values onto a new ENGINE except reference counts/state).
12101 - Removed NULL parameter checks in get/set functions. Setting a method
12102 or function to NULL is a way of cancelling out a previously set
12103 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12104 and doesn't justify the extra error symbols and code.
12105 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12106 flags from engine_int.h to engine.h.
12107 - Changed prototypes for ENGINE handler functions (init(), finish(),
12108 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12112 * Implement binary inversion algorithm for BN_mod_inverse in addition
12113 to the algorithm using long division. The binary algorithm can be
12114 used only if the modulus is odd. On 32-bit systems, it is faster
12115 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12116 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12117 up to 450 bits. In 64-bit environments, the binary algorithm
12118 appears to be advantageous for much longer moduli; here we use it
12119 for moduli up to 2048 bits.
12123 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12124 could not support the combine flag in choice fields.
12128 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12129 extensions from a certificate request to the certificate.
12133 * Allow multiple 'certopt' and 'nameopt' options to be separated
12134 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12135 file: this allows the display of the certificate about to be
12136 signed to be customised, to allow certain fields to be included
12137 or excluded and extension details. The old system didn't display
12138 multicharacter strings properly, omitted fields not in the policy
12139 and couldn't display additional details such as extensions.
12143 * Function EC_POINTs_mul for multiple scalar multiplication
12144 of an arbitrary number of elliptic curve points
12145 \sum scalars[i]*points[i],
12146 optionally including the generator defined for the EC_GROUP:
12147 scalar*generator + \sum scalars[i]*points[i].
12149 EC_POINT_mul is a simple wrapper function for the typical case
12150 that the point list has just one item (besides the optional
12155 * First EC_METHODs for curves over GF(p):
12157 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12158 operations and provides various method functions that can also
12159 operate with faster implementations of modular arithmetic.
12161 EC_GFp_mont_method() reuses most functions that are part of
12162 EC_GFp_simple_method, but uses Montgomery arithmetic.
12164 *Bodo Moeller; point addition and point doubling
12165 implementation directly derived from source code provided by
12166 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12168 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12169 crypto/ec/ec_lib.c):
12171 Curves are EC_GROUP objects (with an optional group generator)
12172 based on EC_METHODs that are built into the library.
12174 Points are EC_POINT objects based on EC_GROUP objects.
12176 Most of the framework would be able to handle curves over arbitrary
12177 finite fields, but as there are no obvious types for fields other
12178 than GF(p), some functions are limited to that for now.
12182 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12183 that the file contains a complete HTTP response.
12187 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12188 change the def and num file printf format specifier from "%-40sXXX"
12189 to "%-39s XXX". The latter will always guarantee a space after the
12190 field while the former will cause them to run together if the field
12191 is 40 of more characters long.
12195 * Constify the cipher and digest 'method' functions and structures
12196 and modify related functions to take constant EVP_MD and EVP_CIPHER
12201 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12202 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12206 * Modify `EVP_Digest*()` routines so they now return values. Although the
12207 internal software routines can never fail additional hardware versions
12212 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12214 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12215 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12218 ERR_R_NESTED_ASN1_ERROR
12220 ERR_R_MISSING_ASN1_EOS
12221 were 4 .. 9, conflicting with
12222 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12224 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12225 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12227 Add new error code 'ERR_R_INTERNAL_ERROR'.
12231 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12236 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12237 sets the subject name for a new request or supersedes the
12238 subject name in a given request. Formats that can be parsed are
12239 'CN=Some Name, OU=myOU, C=IT'
12241 'CN=Some Name/OU=myOU/C=IT'.
12243 Add options '-batch' and '-verbose' to 'openssl req'.
12245 *Massimiliano Pala <madwolf@hackmasters.net>*
12247 * Introduce the possibility to access global variables through
12248 functions on platform were that's the best way to handle exporting
12249 global variables in shared libraries. To enable this functionality,
12250 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12251 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12252 is normally done by Configure or something similar).
12254 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12255 in the source file (foo.c) like this:
12257 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12258 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12260 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12261 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12263 OPENSSL_DECLARE_GLOBAL(int,foo);
12264 #define foo OPENSSL_GLOBAL_REF(foo)
12265 OPENSSL_DECLARE_GLOBAL(double,bar);
12266 #define bar OPENSSL_GLOBAL_REF(bar)
12268 The #defines are very important, and therefore so is including the
12269 header file everywhere where the defined globals are used.
12271 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12272 of ASN.1 items, but that structure is a bit different.
12274 The largest change is in util/mkdef.pl which has been enhanced with
12275 better and easier to understand logic to choose which symbols should
12276 go into the Windows .def files as well as a number of fixes and code
12277 cleanup (among others, algorithm keywords are now sorted
12278 lexicographically to avoid constant rewrites).
12282 * In BN_div() keep a copy of the sign of 'num' before writing the
12283 result to 'rm' because if rm==num the value will be overwritten
12284 and produce the wrong result if 'num' is negative: this caused
12285 problems with BN_mod() and BN_nnmod().
12289 * Function OCSP_request_verify(). This checks the signature on an
12290 OCSP request and verifies the signer certificate. The signer
12291 certificate is just checked for a generic purpose and OCSP request
12296 * Add OCSP_check_validity() function to check the validity of OCSP
12297 responses. OCSP responses are prepared in real time and may only
12298 be a few seconds old. Simply checking that the current time lies
12299 between thisUpdate and nextUpdate max reject otherwise valid responses
12300 caused by either OCSP responder or client clock inaccuracy. Instead
12301 we allow thisUpdate and nextUpdate to fall within a certain period of
12302 the current time. The age of the response can also optionally be
12303 checked. Two new options -validity_period and -status_age added to
12308 * If signature or public key algorithm is unrecognized print out its
12309 OID rather that just UNKNOWN.
12313 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12314 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12315 ID to be generated from the issuer certificate alone which can then be
12316 passed to OCSP_id_issuer_cmp().
12320 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12321 ASN1 modules to export functions returning ASN1_ITEM pointers
12322 instead of the ASN1_ITEM structures themselves. This adds several
12323 new macros which allow the underlying ASN1 function/structure to
12324 be accessed transparently. As a result code should not use ASN1_ITEM
12325 references directly (such as &X509_it) but instead use the relevant
12326 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12327 use of the new ASN1 code on platforms where exporting structures
12328 is problematical (for example in shared libraries) but exporting
12329 functions returning pointers to structures is not.
12333 * Add support for overriding the generation of SSL/TLS session IDs.
12334 These callbacks can be registered either in an SSL_CTX or per SSL.
12335 The purpose of this is to allow applications to control, if they wish,
12336 the arbitrary values chosen for use as session IDs, particularly as it
12337 can be useful for session caching in multiple-server environments. A
12338 command-line switch for testing this (and any client code that wishes
12339 to use such a feature) has been added to "s_server".
12341 *Geoff Thorpe, Lutz Jaenicke*
12343 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12344 of the form `#if defined(...) || defined(...) || ...` and
12345 `#if !defined(...) && !defined(...) && ...`. This also avoids
12346 the growing number of special cases it was previously handling.
12350 * Make all configuration macros available for application by making
12351 sure they are available in opensslconf.h, by giving them names starting
12352 with `OPENSSL_` to avoid conflicts with other packages and by making
12353 sure e_os2.h will cover all platform-specific cases together with
12355 Additionally, it is now possible to define configuration/platform-
12356 specific names (called "system identities"). In the C code, these
12357 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12358 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12359 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12364 * New option -set_serial to 'req' and 'x509' this allows the serial
12365 number to use to be specified on the command line. Previously self
12366 signed certificates were hard coded with serial number 0 and the
12367 CA options of 'x509' had to use a serial number in a file which was
12372 * New options to 'ca' utility to support V2 CRL entry extensions.
12373 Currently CRL reason, invalidity date and hold instruction are
12374 supported. Add new CRL extensions to V3 code and some new objects.
12378 * New function EVP_CIPHER_CTX_set_padding() this is used to
12379 disable standard block padding (aka PKCS#5 padding) in the EVP
12380 API, which was previously mandatory. This means that the data is
12381 not padded in any way and so the total length much be a multiple
12382 of the block size, otherwise an error occurs.
12386 * Initial (incomplete) OCSP SSL support.
12390 * New function OCSP_parse_url(). This splits up a URL into its host,
12391 port and path components: primarily to parse OCSP URLs. New -url
12392 option to ocsp utility.
12396 * New nonce behavior. The return value of OCSP_check_nonce() now
12397 reflects the various checks performed. Applications can decide
12398 whether to tolerate certain situations such as an absent nonce
12399 in a response when one was present in a request: the ocsp application
12400 just prints out a warning. New function OCSP_add1_basic_nonce()
12401 this is to allow responders to include a nonce in a response even if
12402 the request is nonce-less.
12406 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12407 skipped when using openssl x509 multiple times on a single input file,
12408 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12412 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12413 set string type: to handle setting ASN1_TIME structures. Fix ca
12414 utility to correctly initialize revocation date of CRLs.
12418 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12419 the clients preferred ciphersuites and rather use its own preferences.
12420 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12421 Internet Explorer by ensuring unchanged hash method during stepup.
12422 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12426 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12427 to aes and add a new 'exist' option to print out symbols that don't
12432 * Additional options to ocsp utility to allow flags to be set and
12433 additional certificates supplied.
12437 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12438 OCSP client a number of certificate to only verify the response
12443 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12444 handle the new API. Currently only ECB, CBC modes supported. Add new
12447 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12448 Encryption Standard (AES) Ciphersuites for Transport Layer
12449 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12450 not enabled by default and were not part of the "ALL" ciphersuite
12451 alias because they were not yet official; they could be
12452 explicitly requested by specifying the "AESdraft" ciphersuite
12453 group alias. In the final release of OpenSSL 0.9.7, the group
12454 alias is called "AES" and is part of "ALL".)
12456 *Ben Laurie, Steve Henson, Bodo Moeller*
12458 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12459 request to response.
12463 * Functions for OCSP responders. OCSP_request_onereq_count(),
12464 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12465 extract information from a certificate request. OCSP_response_create()
12466 creates a response and optionally adds a basic response structure.
12467 OCSP_basic_add1_status() adds a complete single response to a basic
12468 response and returns the OCSP_SINGLERESP structure just added (to allow
12469 extensions to be included for example). OCSP_basic_add1_cert() adds a
12470 certificate to a basic response and OCSP_basic_sign() signs a basic
12471 response with various flags. New helper functions ASN1_TIME_check()
12472 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12473 (converts ASN1_TIME to GeneralizedTime).
12477 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12478 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12479 structure from a certificate. X509_pubkey_digest() digests the public_key
12480 contents: this is used in various key identifiers.
12484 * Make sk_sort() tolerate a NULL argument.
12486 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12488 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12489 passed by the function are trusted implicitly. If any of them signed the
12490 response then it is assumed to be valid and is not verified.
12494 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12495 to data. This was previously part of the PKCS7 ASN1 code. This
12496 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12497 *Steve Henson, reported by Kenneth R. Robinette
12498 <support@securenetterm.com>*
12500 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12501 routines: without these tracing memory leaks is very painful.
12502 Fix leaks in PKCS12 and PKCS7 routines.
12506 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12507 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12508 effectively meant GeneralizedTime would never be used. Now it
12509 is initialised to -1 but X509_time_adj() now has to check the value
12510 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12511 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12512 *Steve Henson, reported by Kenneth R. Robinette
12513 <support@securenetterm.com>*
12515 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12516 result in a zero length in the ASN1_INTEGER structure which was
12517 not consistent with the structure when d2i_ASN1_INTEGER() was used
12518 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12519 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12520 where it did not print out a minus for negative ASN1_INTEGER.
12524 * Add summary printout to ocsp utility. The various functions which
12525 convert status values to strings have been renamed to:
12526 OCSP_response_status_str(), OCSP_cert_status_str() and
12527 OCSP_crl_reason_str() and are no longer static. New options
12528 to verify nonce values and to disable verification. OCSP response
12529 printout format cleaned up.
12533 * Add additional OCSP certificate checks. These are those specified
12534 in RFC2560. This consists of two separate checks: the CA of the
12535 certificate being checked must either be the OCSP signer certificate
12536 or the issuer of the OCSP signer certificate. In the latter case the
12537 OCSP signer certificate must contain the OCSP signing extended key
12538 usage. This check is performed by attempting to match the OCSP
12539 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12540 in the OCSP_CERTID structures of the response.
12544 * Initial OCSP certificate verification added to OCSP_basic_verify()
12545 and related routines. This uses the standard OpenSSL certificate
12546 verify routines to perform initial checks (just CA validity) and
12547 to obtain the certificate chain. Then additional checks will be
12548 performed on the chain. Currently the root CA is checked to see
12549 if it is explicitly trusted for OCSP signing. This is used to set
12550 a root CA as a global signing root: that is any certificate that
12551 chains to that CA is an acceptable OCSP signing certificate.
12555 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12556 extensions from a separate configuration file.
12557 As when reading extensions from the main configuration file,
12558 the '-extensions ...' option may be used for specifying the
12561 *Massimiliano Pala <madwolf@comune.modena.it>*
12563 * New OCSP utility. Allows OCSP requests to be generated or
12564 read. The request can be sent to a responder and the output
12565 parsed, outputted or printed in text form. Not complete yet:
12566 still needs to check the OCSP response validity.
12570 * New subcommands for 'openssl ca':
12571 `openssl ca -status <serial>` prints the status of the cert with
12572 the given serial number (according to the index file).
12573 `openssl ca -updatedb` updates the expiry status of certificates
12576 *Massimiliano Pala <madwolf@comune.modena.it>*
12578 * New '-newreq-nodes' command option to CA.pl. This is like
12579 '-newreq', but calls 'openssl req' with the '-nodes' option
12580 so that the resulting key is not encrypted.
12582 *Damien Miller <djm@mindrot.org>*
12584 * New configuration for the GNU Hurd.
12586 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12588 * Initial code to implement OCSP basic response verify. This
12589 is currently incomplete. Currently just finds the signer's
12590 certificate and verifies the signature on the response.
12594 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12595 value of OPENSSLDIR. This is available via the new '-d' option
12596 to 'openssl version', and is also included in 'openssl version -a'.
12600 * Allowing defining memory allocation callbacks that will be given
12601 file name and line number information in additional arguments
12602 (a `const char*` and an int). The basic functionality remains, as
12603 well as the original possibility to just replace malloc(),
12604 realloc() and free() by functions that do not know about these
12605 additional arguments. To register and find out the current
12606 settings for extended allocation functions, the following
12607 functions are provided:
12609 CRYPTO_set_mem_ex_functions
12610 CRYPTO_set_locked_mem_ex_functions
12611 CRYPTO_get_mem_ex_functions
12612 CRYPTO_get_locked_mem_ex_functions
12614 These work the same way as CRYPTO_set_mem_functions and friends.
12615 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12616 extended allocation function is enabled.
12617 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12618 a conventional allocation function is enabled.
12620 *Richard Levitte, Bodo Moeller*
12622 * Finish off removing the remaining LHASH function pointer casts.
12623 There should no longer be any prototype-casting required when using
12624 the LHASH abstraction, and any casts that remain are "bugs". See
12625 the callback types and macros at the head of lhash.h for details
12626 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12630 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12631 If /dev/[u]random devices are not available or do not return enough
12632 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12634 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12635 /etc/entropy will be queried once each in this sequence, querying stops
12636 when enough entropy was collected without querying more sockets.
12640 * Change the Unix RAND_poll() variant to be able to poll several
12641 random devices, as specified by DEVRANDOM, until a sufficient amount
12642 of data has been collected. We spend at most 10 ms on each file
12643 (select timeout) and read in non-blocking mode. DEVRANDOM now
12644 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12645 (previously it was just the string "/dev/urandom"), so on typical
12646 platforms the 10 ms delay will never occur.
12647 Also separate out the Unix variant to its own file, rand_unix.c.
12648 For VMS, there's a currently-empty rand_vms.c.
12652 * Move OCSP client related routines to ocsp_cl.c. These
12653 provide utility functions which an application needing
12654 to issue a request to an OCSP responder and analyse the
12655 response will typically need: as opposed to those which an
12656 OCSP responder itself would need which will be added later.
12658 OCSP_request_sign() signs an OCSP request with an API similar
12659 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12660 response. OCSP_response_get1_basic() extracts basic response
12661 from response. OCSP_resp_find_status(): finds and extracts status
12662 information from an OCSP_CERTID structure (which will be created
12663 when the request structure is built). These are built from lower
12664 level functions which work on OCSP_SINGLERESP structures but
12665 won't normally be used unless the application wishes to examine
12666 extensions in the OCSP response for example.
12668 Replace nonce routines with a pair of functions.
12669 OCSP_request_add1_nonce() adds a nonce value and optionally
12670 generates a random value. OCSP_check_nonce() checks the
12671 validity of the nonce in an OCSP response.
12675 * Change function OCSP_request_add() to OCSP_request_add0_id().
12676 This doesn't copy the supplied OCSP_CERTID and avoids the
12677 need to free up the newly created id. Change return type
12678 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12679 This can then be used to add extensions to the request.
12680 Deleted OCSP_request_new(), since most of its functionality
12681 is now in OCSP_REQUEST_new() (and the case insensitive name
12682 clash) apart from the ability to set the request name which
12683 will be added elsewhere.
12687 * Update OCSP API. Remove obsolete extensions argument from
12688 various functions. Extensions are now handled using the new
12689 OCSP extension code. New simple OCSP HTTP function which
12690 can be used to send requests and parse the response.
12694 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12695 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12696 uses the special reorder version of SET OF to sort the attributes
12697 and reorder them to match the encoded order. This resolves a long
12698 standing problem: a verify on a PKCS7 structure just after signing
12699 it used to fail because the attribute order did not match the
12700 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12701 it uses the received order. This is necessary to tolerate some broken
12702 software that does not order SET OF. This is handled by encoding
12703 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12704 to produce the required SET OF.
12708 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12709 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12710 files to get correct declarations of the ASN.1 item variables.
12714 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12715 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12716 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12717 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12718 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12719 ASN1_ITEM and no wrapper functions.
12723 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12724 replace the old function pointer based I/O routines. Change most of
12725 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12729 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12730 lines, recognize more "algorithms" that can be deselected, and make
12731 it complain about algorithm deselection that isn't recognised.
12735 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12736 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12737 to use new functions. Add NO_ASN1_OLD which can be set to remove
12738 some old style ASN1 functions: this can be used to determine if old
12739 code will still work when these eventually go away.
12743 * New extension functions for OCSP structures, these follow the
12744 same conventions as certificates and CRLs.
12748 * New function X509V3_add1_i2d(). This automatically encodes and
12749 adds an extension. Its behaviour can be customised with various
12750 flags to append, replace or delete. Various wrappers added for
12751 certificates and CRLs.
12755 * Fix to avoid calling the underlying ASN1 print routine when
12756 an extension cannot be parsed. Correct a typo in the
12757 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12761 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12762 entries for variables.
12766 * Add functionality to apps/openssl.c for detecting locking
12767 problems: As the program is single-threaded, all we have
12768 to do is register a locking callback using an array for
12769 storing which locks are currently held by the program.
12773 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12774 SSL_get_ex_data_X509_STORE_idx(), which is used in
12775 ssl_verify_cert_chain() and thus can be called at any time
12776 during TLS/SSL handshakes so that thread-safety is essential.
12777 Unfortunately, the ex_data design is not at all suited
12778 for multi-threaded use, so it probably should be abolished.
12782 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12784 *Broadcom, tweaked and integrated by Geoff Thorpe*
12786 * Move common extension printing code to new function
12787 X509V3_print_extensions(). Reorganise OCSP print routines and
12788 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12792 * New function X509_signature_print() to remove duplication in some
12797 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12798 set (this was treated exactly the same as SET OF previously). This
12799 is used to reorder the STACK representing the structure to match the
12800 encoding. This will be used to get round a problem where a PKCS7
12801 structure which was signed could not be verified because the STACK
12802 order did not reflect the encoded order.
12806 * Reimplement the OCSP ASN1 module using the new code.
12810 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12811 for its ASN1 operations. The old style function pointers still exist
12812 for now but they will eventually go away.
12816 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12817 completely replaces the old ASN1 functionality with a table driven
12818 encoder and decoder which interprets an ASN1_ITEM structure describing
12819 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12820 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12821 has also been converted to the new form.
12825 * Change BN_mod_exp_recp so that negative moduli are tolerated
12826 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12827 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12828 for negative moduli.
12832 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12833 of not touching the result's sign bit.
12837 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12842 * Changed the LHASH code to use prototypes for callbacks, and created
12843 macros to declare and implement thin (optionally static) functions
12844 that provide type-safety and avoid function pointer casting for the
12845 type-specific callbacks.
12849 * Added Kerberos Cipher Suites to be used with TLS, as written in
12851 *Veers Staats <staatsvr@asc.hpc.mil>,
12852 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12854 * Reformat the FAQ so the different questions and answers can be divided
12855 in sections depending on the subject.
12859 * Have the zlib compression code load ZLIB.DLL dynamically under
12864 * New function BN_mod_sqrt for computing square roots modulo a prime
12865 (using the probabilistic Tonelli-Shanks algorithm unless
12866 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12867 be handled deterministically).
12869 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12871 * Make BN_mod_inverse faster by explicitly handling small quotients
12872 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12873 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12877 * New function BN_kronecker.
12881 * Fix BN_gcd so that it works on negative inputs; the result is
12882 positive unless both parameters are zero.
12883 Previously something reasonably close to an infinite loop was
12884 possible because numbers could be growing instead of shrinking
12885 in the implementation of Euclid's algorithm.
12889 * Fix BN_is_word() and BN_is_one() macros to take into account the
12890 sign of the number in question.
12892 Fix BN_is_word(a,w) to work correctly for w == 0.
12894 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12895 because its test if the absolute value of 'a' equals 'w'.
12896 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12897 it exists mostly for use in the implementations of BN_is_zero(),
12898 BN_is_one(), and BN_is_word().
12902 * New function BN_swap.
12906 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12907 the exponentiation functions are more likely to produce reasonable
12908 results on negative inputs.
12912 * Change BN_mod_mul so that the result is always non-negative.
12913 Previously, it could be negative if one of the factors was negative;
12914 I don't think anyone really wanted that behaviour.
12918 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
12919 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
12920 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
12921 and add new functions:
12930 BN_mod_lshift1_quick
12932 BN_mod_lshift_quick
12934 These functions always generate non-negative results.
12936 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
12937 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
12939 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
12940 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
12941 be reduced modulo `m`.
12943 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12946 The following entry accidentally appeared in the CHANGES file
12947 distributed with OpenSSL 0.9.7. The modifications described in
12948 it do *not* apply to OpenSSL 0.9.7.
12950 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12951 was actually never needed) and in BN_mul(). The removal in BN_mul()
12952 required a small change in bn_mul_part_recursive() and the addition
12953 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12954 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12955 bn_sub_words() and bn_add_words() except they take arrays with
12961 * In 'openssl passwd', verify passwords read from the terminal
12962 unless the '-salt' option is used (which usually means that
12963 verification would just waste user's time since the resulting
12964 hash is going to be compared with some given password hash)
12965 or the new '-noverify' option is used.
12967 This is an incompatible change, but it does not affect
12968 non-interactive use of 'openssl passwd' (passwords on the command
12969 line, '-stdin' option, '-in ...' option) and thus should not
12970 cause any problems.
12974 * Remove all references to RSAref, since there's no more need for it.
12978 * Make DSO load along a path given through an environment variable
12979 (SHLIB_PATH) with shl_load().
12983 * Constify the ENGINE code as a result of BIGNUM constification.
12984 Also constify the RSA code and most things related to it. In a
12985 few places, most notable in the depth of the ASN.1 code, ugly
12986 casts back to non-const were required (to be solved at a later
12991 * Make it so the openssl application has all engines loaded by default.
12995 * Constify the BIGNUM routines a little more.
12999 * Add the following functions:
13001 ENGINE_load_cswift()
13003 ENGINE_load_atalla()
13004 ENGINE_load_nuron()
13005 ENGINE_load_builtin_engines()
13007 That way, an application can itself choose if external engines that
13008 are built-in in OpenSSL shall ever be used or not. The benefit is
13009 that applications won't have to be linked with libdl or other dso
13010 libraries unless it's really needed.
13012 Changed 'openssl engine' to load all engines on demand.
13013 Changed the engine header files to avoid the duplication of some
13014 declarations (they differed!).
13018 * 'openssl engine' can now list capabilities.
13022 * Better error reporting in 'openssl engine'.
13026 * Never call load_dh_param(NULL) in s_server.
13030 * Add engine application. It can currently list engines by name and
13031 identity, and test if they are actually available.
13035 * Improve RPM specification file by forcing symbolic linking and making
13036 sure the installed documentation is also owned by root.root.
13038 *Damien Miller <djm@mindrot.org>*
13040 * Give the OpenSSL applications more possibilities to make use of
13041 keys (public as well as private) handled by engines.
13045 * Add OCSP code that comes from CertCo.
13049 * Add VMS support for the Rijndael code.
13053 * Added untested support for Nuron crypto accelerator.
13057 * Add support for external cryptographic devices. This code was
13058 previously distributed separately as the "engine" branch.
13060 *Geoff Thorpe, Richard Levitte*
13062 * Rework the filename-translation in the DSO code. It is now possible to
13063 have far greater control over how a "name" is turned into a filename
13064 depending on the operating environment and any oddities about the
13065 different shared library filenames on each system.
13069 * Support threads on FreeBSD-elf in Configure.
13073 * Fix for SHA1 assembly problem with MASM: it produces
13074 warnings about corrupt line number information when assembling
13075 with debugging information. This is caused by the overlapping
13078 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13081 NCONF_get_number() has no error checking at all. As a replacement,
13082 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13083 promoted strongly. The old NCONF_get_number is kept around for
13084 binary backward compatibility.
13085 Make it possible for methods to load from something other than a BIO,
13086 by providing a function pointer that is given a name instead of a BIO.
13087 For example, this could be used to load configuration data from an
13092 * Fix for non blocking accept BIOs. Added new I/O special reason
13093 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13094 with non blocking I/O was not possible because no retry code was
13095 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13100 * Added the beginnings of Rijndael support.
13104 * Fix for bug in DirectoryString mask setting. Add support for
13105 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13106 to allow certificate printing to more controllable, additional
13107 'certopt' option to 'x509' to allow new printing options to be
13112 * Clean old EAY MD5 hack from e_os.h.
13116 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13118 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13119 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13121 *Joe Orton, Steve Henson*
13123 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13125 * Fix additional bug revealed by the NISCC test suite:
13127 Stop bug triggering large recursion when presented with
13128 certain ASN.1 tags [CVE-2003-0851][]
13132 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13134 * Fix various bugs revealed by running the NISCC test suite:
13136 Stop out of bounds reads in the ASN1 code when presented with
13137 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13139 If verify callback ignores invalid public key errors don't try to check
13140 certificate signature with the NULL public key.
13144 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13145 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13150 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13151 extra data after the compression methods not only for TLS 1.0
13152 but also for SSL 3.0 (as required by the specification).
13154 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13156 * Change X509_certificate_type() to mark the key as exported/exportable
13157 when it's 512 *bits* long, not 512 bytes.
13161 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13163 * Countermeasure against the Klima-Pokorny-Rosa extension of
13164 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13165 a protocol version number mismatch like a decryption error
13166 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13170 * Turn on RSA blinding by default in the default implementation
13171 to avoid a timing attack. Applications that don't want it can call
13172 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13173 They would be ill-advised to do so in most cases.
13175 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13177 * Change RSA blinding code so that it works when the PRNG is not
13178 seeded (in this case, the secret RSA exponent is abused as
13179 an unpredictable seed -- if it is not unpredictable, there
13180 is no point in blinding anyway). Make RSA blinding thread-safe
13181 by remembering the creator's thread ID in rsa->blinding and
13182 having all other threads use local one-time blinding factors
13183 (this requires more computation than sharing rsa->blinding, but
13184 avoids excessive locking; and if an RSA object is not shared
13185 between threads, blinding will still be very fast).
13189 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13191 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13192 via timing by performing a MAC computation even if incorrect
13193 block cipher padding has been found. This is a countermeasure
13194 against active attacks where the attacker has to distinguish
13195 between bad padding and a MAC verification error. [CVE-2003-0078][]
13197 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13198 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13199 Martin Vuagnoux (EPFL, Ilion)*
13201 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13203 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13204 memory from its contents. This is done with a counter that will
13205 place alternating values in each byte. This can be used to solve
13206 two issues: 1) the removal of calls to memset() by highly optimizing
13207 compilers, and 2) cleansing with other values than 0, since those can
13208 be read through on certain media, for example a swap space on disk.
13212 * Bugfix: client side session caching did not work with external caching,
13213 because the session->cipher setting was not restored when reloading
13214 from the external cache. This problem was masked, when
13215 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13216 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13220 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13221 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13223 *Zeev Lieber <zeev-l@yahoo.com>*
13225 * Undo an undocumented change introduced in 0.9.6e which caused
13226 repeated calls to OpenSSL_add_all_ciphers() and
13227 OpenSSL_add_all_digests() to be ignored, even after calling
13232 * Change the default configuration reader to deal with last line not
13233 being properly terminated.
13237 * Change X509_NAME_cmp() so it applies the special rules on handling
13238 DN values that are of type PrintableString, as well as RDNs of type
13239 emailAddress where the value has the type ia5String.
13241 *stefank@valicert.com via Richard Levitte*
13243 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13244 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13245 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13246 the bitwise-OR of the two for use by the majority of applications
13247 wanting this behaviour, and update the docs. The documented
13248 behaviour and actual behaviour were inconsistent and had been
13249 changing anyway, so this is more a bug-fix than a behavioural
13252 *Geoff Thorpe, diagnosed by Nadav Har'El*
13254 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13255 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13259 * Fix initialization code race conditions in
13260 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13261 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13262 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13263 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13264 ssl2_get_cipher_by_char(),
13265 ssl3_get_cipher_by_char().
13267 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13269 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13270 the cached sessions are flushed, as the remove_cb() might use ex_data
13271 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13272 (see [openssl.org #212]).
13274 *Geoff Thorpe, Lutz Jaenicke*
13276 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13277 length, instead of the encoding length to d2i_ASN1_OBJECT.
13281 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13283 * [In 0.9.6g-engine release:]
13284 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13286 *Lynn Gazis <lgazis@rainbow.com>*
13288 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13290 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13291 and get fix the header length calculation.
13292 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13293 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13295 * Use proper error handling instead of 'assertions' in buffer
13296 overflow checks added in 0.9.6e. This prevents DoS (the
13297 assertions could call abort()).
13299 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13301 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13303 * Add various sanity checks to asn1_get_length() to reject
13304 the ASN1 length bytes if they exceed sizeof(long), will appear
13305 negative or the content length exceeds the length of the
13308 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13310 * Fix cipher selection routines: ciphers without encryption had no flags
13311 for the cipher strength set and where therefore not handled correctly
13312 by the selection routines (PR #130).
13316 * Fix EVP_dsa_sha macro.
13321 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13322 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13323 that was added in OpenSSL 0.9.6d.
13325 As the countermeasure turned out to be incompatible with some
13326 broken SSL implementations, the new option is part of SSL_OP_ALL.
13327 SSL_OP_ALL is usually employed when compatibility with weird SSL
13328 implementations is desired (e.g. '-bugs' option to 's_client' and
13329 's_server'), so the new option is automatically set in many
13334 * Changes in security patch:
13336 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13337 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13338 Air Force Materiel Command, USAF, under agreement number
13341 * Add various sanity checks to asn1_get_length() to reject
13342 the ASN1 length bytes if they exceed sizeof(long), will appear
13343 negative or the content length exceeds the length of the
13344 supplied buffer. [CVE-2002-0659][]
13346 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13348 * Assertions for various potential buffer overflows, not known to
13349 happen in practice.
13351 *Ben Laurie (CHATS)*
13353 * Various temporary buffers to hold ASCII versions of integers were
13354 too small for 64 bit platforms. [CVE-2002-0655][]
13355 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13357 * Remote buffer overflow in SSL3 protocol - an attacker could
13358 supply an oversized session ID to a client. [CVE-2002-0656][]
13360 *Ben Laurie (CHATS)*
13362 * Remote buffer overflow in SSL2 protocol - an attacker could
13363 supply an oversized client master key. [CVE-2002-0656][]
13365 *Ben Laurie (CHATS)*
13367 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13369 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13370 encoded as NULL) with id-dsa-with-sha1.
13372 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13374 * Check various `X509_...()` return values in apps/req.c.
13376 *Nils Larsch <nla@trustcenter.de>*
13378 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13379 an end-of-file condition would erroneously be flagged, when the CRLF
13380 was just at the end of a processed block. The bug was discovered when
13381 processing data through a buffering memory BIO handing the data to a
13382 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13383 <ptsekov@syntrex.com> and Nedelcho Stanev.
13387 * Implement a countermeasure against a vulnerability recently found
13388 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13389 before application data chunks to avoid the use of known IVs
13390 with data potentially chosen by the attacker.
13394 * Fix length checks in ssl3_get_client_hello().
13398 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13399 to prevent ssl3_read_internal() from incorrectly assuming that
13400 ssl3_read_bytes() found application data while handshake
13401 processing was enabled when in fact s->s3->in_read_app_data was
13402 merely automatically cleared during the initial handshake.
13404 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13406 * Fix object definitions for Private and Enterprise: they were not
13407 recognized in their shortname (=lowercase) representation. Extend
13408 obj_dat.pl to issue an error when using undefined keywords instead
13409 of silently ignoring the problem (Svenning Sorensen
13410 <sss@sss.dnsalias.net>).
13414 * Fix DH_generate_parameters() so that it works for 'non-standard'
13415 generators, i.e. generators other than 2 and 5. (Previously, the
13416 code did not properly initialise the 'add' and 'rem' values to
13417 BN_generate_prime().)
13419 In the new general case, we do not insist that 'generator' is
13420 actually a primitive root: This requirement is rather pointless;
13421 a generator of the order-q subgroup is just as good, if not
13426 * Map new X509 verification errors to alerts. Discovered and submitted by
13427 Tom Wu <tom@arcot.com>.
13431 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13432 returning non-zero before the data has been completely received
13433 when using non-blocking I/O.
13435 *Bodo Moeller; problem pointed out by John Hughes*
13437 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13439 *Ben Laurie, Lutz Jaenicke*
13441 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13442 Yoram Zahavi <YoramZ@gilian.com>).
13446 * Add information about CygWin 1.3 and on, and preserve proper
13447 configuration for the versions before that.
13449 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13451 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13452 check whether we deal with a copy of a session and do not delete from
13453 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13454 <izhar@checkpoint.com>.
13458 * Do not store session data into the internal session cache, if it
13459 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13460 flag is set). Proposed by Aslam <aslam@funk.com>.
13464 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13469 * [In 0.9.6d-engine release:]
13470 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13472 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13474 * Add the configuration target linux-s390x.
13476 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13478 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13479 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13480 variable as an indication that a ClientHello message has been
13481 received. As the flag value will be lost between multiple
13482 invocations of ssl3_accept when using non-blocking I/O, the
13483 function may not be aware that a handshake has actually taken
13484 place, thus preventing a new session from being added to the
13487 To avoid this problem, we now set s->new_session to 2 instead of
13488 using a local variable.
13490 *Lutz Jaenicke, Bodo Moeller*
13492 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13493 if the SSL_R_LENGTH_MISMATCH error is detected.
13495 *Geoff Thorpe, Bodo Moeller*
13497 * New 'shared_ldflag' column in Configure platform table.
13501 * Fix EVP_CIPHER_mode macro.
13503 *"Dan S. Camper" <dan@bti.net>*
13505 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13506 type, we must throw them away by setting rr->length to 0.
13508 *D P Chang <dpc@qualys.com>*
13510 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13512 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13513 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13514 worked incorrectly for those cases where range = `10..._2` and
13515 `3*range` is two bits longer than range.)
13519 * Only add signing time to PKCS7 structures if it is not already
13524 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13525 OBJ_ld_ce should be OBJ_id_ce.
13526 Also some ip-pda OIDs in crypto/objects/objects.txt were
13527 incorrect (cf. RFC 3039).
13529 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13531 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13532 returns early because it has nothing to do.
13534 *Andy Schneider <andy.schneider@bjss.co.uk>*
13536 * [In 0.9.6c-engine release:]
13537 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13539 *Andy Schneider <andy.schneider@bjss.co.uk>*
13541 * [In 0.9.6c-engine release:]
13542 Add support for Cryptographic Appliance's keyserver technology.
13543 (Use engine 'keyclient')
13545 *Cryptographic Appliances and Geoff Thorpe*
13547 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13548 is called via tools/c89.sh because arguments have to be
13549 rearranged (all '-L' options must appear before the first object
13552 *Richard Shapiro <rshapiro@abinitio.com>*
13554 * [In 0.9.6c-engine release:]
13555 Add support for Broadcom crypto accelerator cards, backported
13558 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13560 * [In 0.9.6c-engine release:]
13561 Add support for SureWare crypto accelerator cards from
13562 Baltimore Technologies. (Use engine 'sureware')
13564 *Baltimore Technologies and Mark Cox*
13566 * [In 0.9.6c-engine release:]
13567 Add support for crypto accelerator cards from Accelerated
13568 Encryption Processing, www.aep.ie. (Use engine 'aep')
13570 *AEP Inc. and Mark Cox*
13572 * Add a configuration entry for gcc on UnixWare.
13574 *Gary Benson <gbenson@redhat.com>*
13576 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13577 messages are stored in a single piece (fixed-length part and
13578 variable-length part combined) and fix various bugs found on the way.
13582 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13583 instead. BIO_gethostbyname() does not know what timeouts are
13584 appropriate, so entries would stay in cache even when they have
13586 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13588 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13589 faced with a pathologically small ClientHello fragment that does
13590 not contain client_version: Instead of aborting with an error,
13591 simply choose the highest available protocol version (i.e.,
13592 TLS 1.0 unless it is disabled). In practice, ClientHello
13593 messages are never sent like this, but this change gives us
13594 strictly correct behaviour at least for TLS.
13598 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13599 never resets s->method to s->ctx->method when called from within
13600 one of the SSL handshake functions.
13602 *Bodo Moeller; problem pointed out by Niko Baric*
13604 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13605 (sent using the client's version number) if client_version is
13606 smaller than the protocol version in use. Also change
13607 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13608 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13609 the client will at least see that alert.
13613 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13618 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13619 client receives HelloRequest while in a handshake.
13621 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13623 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13624 should end in 'break', not 'goto end' which circumvents various
13625 cleanups done in state SSL_ST_OK. But session related stuff
13626 must be disabled for SSL_ST_OK in the case that we just sent a
13629 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13630 before just sending a HelloRequest.
13632 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13634 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13635 reveal whether illegal block cipher padding was found or a MAC
13636 verification error occurred. (Neither SSLerr() codes nor alerts
13637 are directly visible to potential attackers, but the information
13638 may leak via logfiles.)
13640 Similar changes are not required for the SSL 2.0 implementation
13641 because the number of padding bytes is sent in clear for SSL 2.0,
13642 and the extra bytes are just ignored. However ssl/s2_pkt.c
13643 failed to verify that the purported number of padding bytes is in
13648 * Add OpenUNIX-8 support including shared libraries
13649 (Boyd Lynn Gerber <gerberb@zenez.com>).
13653 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13654 'wristwatch attack' using huge encoding parameters (cf.
13655 James H. Manger's CRYPTO 2001 paper). Note that the
13656 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13657 encoding parameters and hence was not vulnerable.
13661 * BN_sqr() bug fix.
13663 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13665 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13666 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13667 followed by modular reduction.
13669 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13671 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13672 equivalent based on BN_pseudo_rand() instead of BN_rand().
13676 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13677 This function was broken, as the check for a new client hello message
13678 to handle SGC did not allow these large messages.
13679 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13683 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13687 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13688 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13692 * Rework the configuration and shared library support for Tru64 Unix.
13693 The configuration part makes use of modern compiler features and
13694 still retains old compiler behavior for those that run older versions
13695 of the OS. The shared library support part includes a variant that
13696 uses the RPATH feature, and is available through the special
13697 configuration target "alpha-cc-rpath", which will never be selected
13700 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13702 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13703 with the same message size as in ssl3_get_certificate_request().
13704 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13705 messages might inadvertently be reject as too long.
13707 *Petr Lampa <lampa@fee.vutbr.cz>*
13709 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13713 * Modified SSL library such that the verify_callback that has been set
13714 specifically for an SSL object with SSL_set_verify() is actually being
13715 used. Before the change, a verify_callback set with this function was
13716 ignored and the verify_callback() set in the SSL_CTX at the time of
13717 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13718 to allow the necessary settings.
13722 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13723 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13724 done automatically (in contradiction to the requirements of the C
13725 standard). This made problems when used from OpenSSH.
13729 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13730 dh->length and always used
13732 BN_rand_range(priv_key, dh->p).
13734 BN_rand_range() is not necessary for Diffie-Hellman, and this
13735 specific range makes Diffie-Hellman unnecessarily inefficient if
13736 dh->length (recommended exponent length) is much smaller than the
13737 length of dh->p. We could use BN_rand_range() if the order of
13738 the subgroup was stored in the DH structure, but we only have
13743 BN_rand(priv_key, l, ...)
13745 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13752 RSA_eay_public_encrypt
13753 RSA_eay_private_decrypt
13754 RSA_eay_private_encrypt (signing)
13755 RSA_eay_public_decrypt (signature verification)
13757 (default implementations for RSA_public_encrypt,
13758 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13759 always reject numbers >= n.
13763 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13764 to synchronize access to 'locking_thread'. This is necessary on
13765 systems where access to 'locking_thread' (an 'unsigned long'
13766 variable) is not atomic.
13770 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13771 *before* setting the 'crypto_lock_rand' flag. The previous code had
13772 a race condition if 0 is a valid thread ID.
13774 *Travis Vitek <vitek@roguewave.com>*
13776 * Add support for shared libraries under Irix.
13778 *Albert Chin-A-Young <china@thewrittenword.com>*
13780 * Add configuration option to build on Linux on both big-endian and
13781 little-endian MIPS.
13783 *Ralf Baechle <ralf@uni-koblenz.de>*
13785 * Add the possibility to create shared libraries on HP-UX.
13789 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13791 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13792 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13793 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13794 PRNG state recovery was possible based on the output of
13795 one PRNG request appropriately sized to gain knowledge on
13796 'md' followed by enough consecutive 1-byte PRNG requests
13797 to traverse all of 'state'.
13799 1. When updating 'md_local' (the current thread's copy of 'md')
13800 during PRNG output generation, hash all of the previous
13801 'md_local' value, not just the half used for PRNG output.
13803 2. Make the number of bytes from 'state' included into the hash
13804 independent from the number of PRNG bytes requested.
13806 The first measure alone would be sufficient to avoid
13807 Markku-Juhani's attack. (Actually it had never occurred
13808 to me that the half of 'md_local' used for chaining was the
13809 half from which PRNG output bytes were taken -- I had always
13810 assumed that the secret half would be used.) The second
13811 measure makes sure that additional data from 'state' is never
13812 mixed into 'md_local' in small portions; this heuristically
13813 further strengthens the PRNG.
13817 * Fix crypto/bn/asm/mips3.s.
13821 * When only the key is given to "enc", the IV is undefined. Print out
13822 an error message in this case.
13826 * Handle special case when X509_NAME is empty in X509 printing routines.
13830 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13831 positive and less than q.
13835 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13836 used: it isn't thread safe and the add_lock_callback should handle
13839 *Paul Rose <Paul.Rose@bridge.com>*
13841 * Verify that incoming data obeys the block size in
13842 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13848 *Ulf Möller, Bodo Möller*
13850 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13851 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13852 when fixing the server behaviour for backwards-compatible 'client
13853 hello' messages. (Note that the attack is impractical against
13854 SSL 3.0 and TLS 1.0 anyway because length and version checking
13855 means that the probability of guessing a valid ciphertext is
13856 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13859 Before 0.9.5, the countermeasure (hide the error by generating a
13860 random 'decryption result') did not work properly because
13861 ERR_clear_error() was missing, meaning that SSL_get_error() would
13862 detect the supposedly ignored error.
13864 Both problems are now fixed.
13868 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13869 (previously it was 1024).
13873 * Fix for compatibility mode trust settings: ignore trust settings
13874 unless some valid trust or reject settings are present.
13878 * Fix for blowfish EVP: its a variable length cipher.
13882 * Fix various bugs related to DSA S/MIME verification. Handle missing
13883 parameters in DSA public key structures and return an error in the
13884 DSA routines if parameters are absent.
13888 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13889 in the current directory if neither $RANDFILE nor $HOME was set.
13890 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13891 caused some confusion to Windows users who haven't defined $HOME.
13892 Thus RAND_file_name() is changed again: e_os.h can define a
13893 DEFAULT_HOME, which will be used if $HOME is not set.
13894 For Windows, we use "C:"; on other platforms, we still require
13895 environment variables.
13897 * Move 'if (!initialized) RAND_poll()' into regions protected by
13898 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13899 having multiple threads call RAND_poll() concurrently.
13903 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13904 combination of a flag and a thread ID variable.
13905 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13906 flag), *other* threads can enter ssleay_add_bytes without obeying
13907 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13908 that they do not hold after the first thread unsets add_do_not_lock).
13912 * Change bctest again: '-x' expressions are not available in all
13913 versions of 'test'.
13917 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13919 * Fix a couple of memory leaks in PKCS7_dataDecode()
13921 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13923 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13924 the default extension for executables, if any. Also, make the perl
13925 scripts that use symlink() to test if it really exists and use "cp"
13926 if it doesn't. All this made OpenSSL compilable and installable in
13931 * Fix for asn1_GetSequence() for indefinite length constructed data.
13932 If SEQUENCE is length is indefinite just set c->slen to the total
13933 amount of data available.
13935 *Steve Henson, reported by shige@FreeBSD.org*
13937 *This change does not apply to 0.9.7.*
13939 * Change bctest to avoid here-documents inside command substitution
13940 (workaround for FreeBSD /bin/sh bug).
13941 For compatibility with Ultrix, avoid shell functions (introduced
13942 in the bctest version that searches along $PATH).
13946 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13947 with des_encrypt() defined on some operating systems, like Solaris
13952 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13953 On the Importance of Eliminating Errors in Cryptographic
13954 Computations, J. Cryptology 14 (2001) 2, 101-119,
13955 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13959 * MIPS assembler BIGNUM division bug fix.
13963 * Disabled incorrect Alpha assembler code.
13967 * Fix PKCS#7 decode routines so they correctly update the length
13968 after reading an EOC for the EXPLICIT tag.
13972 *This change does not apply to 0.9.7.*
13974 * Fix bug in PKCS#12 key generation routines. This was triggered
13975 if a 3DES key was generated with a 0 initial byte. Include
13976 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13977 (but broken) behaviour.
13981 * Enhance bctest to search for a working bc along $PATH and print
13984 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13986 * Fix memory leaks in err.c: free err_data string if necessary;
13987 don't write to the wrong index in ERR_set_error_data.
13991 * Implement ssl23_peek (analogous to ssl23_read), which previously
13996 * Replace rdtsc with `_emit` statements for VC++ version 5.
13998 *Jeremy Cooper <jeremy@baymoo.org>*
14000 * Make it possible to reuse SSLv2 sessions.
14004 * In copy_email() check for >= 0 as a return value for
14005 X509_NAME_get_index_by_NID() since 0 is a valid index.
14007 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14009 * Avoid coredump with unsupported or invalid public keys by checking if
14010 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14011 PKCS7_verify() fails with non detached data.
14015 * Don't use getenv in library functions when run as setuid/setgid.
14016 New function OPENSSL_issetugid().
14020 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14021 due to incorrect handling of multi-threading:
14023 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14025 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14027 3. Count how many times MemCheck_off() has been called so that
14028 nested use can be treated correctly. This also avoids
14029 inband-signalling in the previous code (which relied on the
14030 assumption that thread ID 0 is impossible).
14034 * Add "-rand" option also to s_client and s_server.
14038 * Fix CPU detection on Irix 6.x.
14039 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14040 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14042 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14047 *This change does not apply to 0.9.7.*
14049 * Use the cached encoding of an X509_NAME structure rather than
14050 copying it. This is apparently the reason for the libsafe "errors"
14051 but the code is actually correct.
14055 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14056 Bleichenbacher's DSA attack.
14057 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14058 to be set and top=0 forces the highest bit to be set; top=-1 is new
14059 and leaves the highest bit random.
14061 *Ulf Moeller, Bodo Moeller*
14063 * In the `NCONF_...`-based implementations for `CONF_...` queries
14064 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14065 a temporary CONF structure with the data component set to NULL
14066 (which gives segmentation faults in lh_retrieve).
14067 Instead, use NULL for the CONF pointer in CONF_get_string and
14068 CONF_get_number (which may use environment variables) and directly
14069 return NULL from CONF_get_section.
14073 * Fix potential buffer overrun for EBCDIC.
14077 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14078 keyUsage if basicConstraints absent for a CA.
14082 * Make SMIME_write_PKCS7() write mail header values with a format that
14083 is more generally accepted (no spaces before the semicolon), since
14084 some programs can't parse those values properly otherwise. Also make
14085 sure BIO's that break lines after each write do not create invalid
14090 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14091 macros previously used would not encode an empty SEQUENCE OF
14092 and break the signature.
14096 *This change does not apply to 0.9.7.*
14098 * Zero the premaster secret after deriving the master secret in
14103 * Add some EVP_add_digest_alias registrations (as found in
14104 OpenSSL_add_all_digests()) to SSL_library_init()
14105 aka OpenSSL_add_ssl_algorithms(). This provides improved
14106 compatibility with peers using X.509 certificates
14107 with unconventional AlgorithmIdentifier OIDs.
14111 * Fix for Irix with NO_ASM.
14113 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14115 * ./config script fixes.
14117 *Ulf Moeller, Richard Levitte*
14119 * Fix 'openssl passwd -1'.
14123 * Change PKCS12_key_gen_asc() so it can cope with non null
14124 terminated strings whose length is passed in the passlen
14125 parameter, for example from PEM callbacks. This was done
14126 by adding an extra length parameter to asc2uni().
14128 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14130 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14131 call failed, free the DSA structure.
14135 * Fix to uni2asc() to cope with zero length Unicode strings.
14136 These are present in some PKCS#12 files.
14140 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14141 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14142 when writing a 32767 byte record.
14144 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14146 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14147 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14149 (RSA objects have a reference count access to which is protected
14150 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14151 so they are meant to be shared between threads.)
14152 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14153 "Reddie, Steven" <Steven.Reddie@ca.com>*
14155 * Fix a deadlock in CRYPTO_mem_leaks().
14159 * Use better test patterns in bntest.
14163 * rand_win.c fix for Borland C.
14167 * BN_rshift bugfix for n == 0.
14171 * Add a 'bctest' script that checks for some known 'bc' bugs
14172 so that 'make test' does not abort just because 'bc' is broken.
14176 * Store verify_result within SSL_SESSION also for client side to
14177 avoid potential security hole. (Re-used sessions on the client side
14178 always resulted in verify_result==X509_V_OK, not using the original
14179 result of the server certificate verification.)
14183 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14184 SSL3_RT_APPLICATION_DATA, return 0.
14185 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14190 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14191 releases, have been re-implemented by renaming the previous
14192 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14193 and ssl3_read_internal, respectively, and adding 'peek' parameters
14194 to them. The new ssl[23]_{read,peek} functions are calls to
14195 ssl[23]_read_internal with the 'peek' flag set appropriately.
14196 A 'peek' parameter has also been added to ssl3_read_bytes, which
14197 does the actual work for ssl3_read_internal.
14201 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14202 the method-specific "init()" handler. Also clean up ex_data after
14203 calling the method-specific "finish()" handler. Previously, this was
14204 happening the other way round.
14208 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14209 The previous value, 12, was not always sufficient for BN_mod_exp().
14213 * Make sure that shared libraries get the internal name engine with
14214 the full version number and not just 0. This should mark the
14215 shared libraries as not backward compatible. Of course, this should
14216 be changed again when we can guarantee backward binary compatibility.
14220 * Fix typo in get_cert_by_subject() in by_dir.c
14222 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14224 * Rework the system to generate shared libraries:
14226 - Make note of the expected extension for the shared libraries and
14227 if there is a need for symbolic links from for example libcrypto.so.0
14228 to libcrypto.so.0.9.7. There is extended info in Configure for
14231 - Make as few rebuilds of the shared libraries as possible.
14233 - Still avoid linking the OpenSSL programs with the shared libraries.
14235 - When installing, install the shared libraries separately from the
14240 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14242 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14243 and not in SSL_clear because the latter is also used by the
14244 accept/connect functions; previously, the settings made by
14245 SSL_set_read_ahead would be lost during the handshake.
14247 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14249 * Correct util/mkdef.pl to be selective about disabled algorithms.
14250 Previously, it would create entries for disabled algorithms no
14255 * Added several new manual pages for SSL_* function.
14259 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14261 * In ssl23_get_client_hello, generate an error message when faced
14262 with an initial SSL 3.0/TLS record that is too small to contain the
14263 first two bytes of the ClientHello message, i.e. client_version.
14264 (Note that this is a pathologic case that probably has never happened
14265 in real life.) The previous approach was to use the version number
14266 from the record header as a substitute; but our protocol choice
14267 should not depend on that one because it is not authenticated
14268 by the Finished messages.
14272 * More robust randomness gathering functions for Windows.
14274 *Jeffrey Altman <jaltman@columbia.edu>*
14276 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14277 not set then we don't setup the error code for issuer check errors
14278 to avoid possibly overwriting other errors which the callback does
14279 handle. If an application does set the flag then we assume it knows
14280 what it is doing and can handle the new informational codes
14285 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14286 a general "ANY" type, as such it should be able to decode anything
14287 including tagged types. However it didn't check the class so it would
14288 wrongly interpret tagged types in the same way as their universal
14289 counterpart and unknown types were just rejected. Changed so that the
14290 tagged and unknown types are handled in the same way as a SEQUENCE:
14291 that is the encoding is stored intact. There is also a new type
14292 "V_ASN1_OTHER" which is used when the class is not universal, in this
14293 case we have no idea what the actual type is so we just lump them all
14298 * On VMS, stdout may very well lead to a file that is written to
14299 in a record-oriented fashion. That means that every write() will
14300 write a separate record, which will be read separately by the
14301 programs trying to read from it. This can be very confusing.
14303 The solution is to put a BIO filter in the way that will buffer
14304 text until a linefeed is reached, and then write everything a
14305 line at a time, so every record written will be an actual line,
14306 not chunks of lines and not (usually doesn't happen, but I've
14307 seen it once) several lines in one record. BIO_f_linebuffer() is
14310 Currently, it's a VMS-only method, because that's where it has
14311 been tested well enough.
14315 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14316 it can return incorrect results.
14317 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14318 but it was in 0.9.6-beta[12].)
14322 * Disable the check for content being present when verifying detached
14323 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14324 include zero length content when signing messages.
14328 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14329 BIO_ctrl (for BIO pairs).
14333 * Add DSO method for VMS.
14337 * Bug fix: Montgomery multiplication could produce results with the
14342 * Add RPM specification openssl.spec and modify it to build three
14343 packages. The default package contains applications, application
14344 documentation and run-time libraries. The devel package contains
14345 include files, static libraries and function documentation. The
14346 doc package contains the contents of the doc directory. The original
14347 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14351 * Add a large number of documentation files for many SSL routines.
14353 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14355 * Add a configuration entry for Sony News 4.
14357 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14359 * Don't set the two most significant bits to one when generating a
14360 random number < q in the DSA library.
14364 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14365 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14366 the underlying transport is blocking) if a handshake took place.
14367 (The default behaviour is needed by applications such as s_client
14368 and s_server that use select() to determine when to use SSL_read;
14369 but for applications that know in advance when to expect data, it
14370 just makes things more complicated.)
14374 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14379 * Add a few more EBCDIC conditionals that make `req` and `x509`
14380 work better on such systems.
14382 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14384 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14385 Update PKCS12_parse() so it copies the friendlyName and the
14386 keyid to the certificates aux info.
14390 * Fix bug in PKCS7_verify() which caused an infinite loop
14391 if there was more than one signature.
14393 *Sven Uszpelkat <su@celocom.de>*
14395 * Major change in util/mkdef.pl to include extra information
14396 about each symbol, as well as presenting variables as well
14397 as functions. This change means that there's n more need
14398 to rebuild the .num files when some algorithms are excluded.
14402 * Allow the verify time to be set by an application,
14403 rather than always using the current time.
14407 * Phase 2 verify code reorganisation. The certificate
14408 verify code now looks up an issuer certificate by a
14409 number of criteria: subject name, authority key id
14410 and key usage. It also verifies self signed certificates
14411 by the same criteria. The main comparison function is
14412 X509_check_issued() which performs these checks.
14414 Lot of changes were necessary in order to support this
14415 without completely rewriting the lookup code.
14417 Authority and subject key identifier are now cached.
14419 The LHASH 'certs' is X509_STORE has now been replaced
14420 by a STACK_OF(X509_OBJECT). This is mainly because an
14421 LHASH can't store or retrieve multiple objects with
14422 the same hash value.
14424 As a result various functions (which were all internal
14425 use only) have changed to handle the new X509_STORE
14426 structure. This will break anything that messed round
14427 with X509_STORE internally.
14429 The functions X509_STORE_add_cert() now checks for an
14430 exact match, rather than just subject name.
14432 The X509_STORE API doesn't directly support the retrieval
14433 of multiple certificates matching a given criteria, however
14434 this can be worked round by performing a lookup first
14435 (which will fill the cache with candidate certificates)
14436 and then examining the cache for matches. This is probably
14437 the best we can do without throwing out X509_LOOKUP
14438 entirely (maybe later...).
14440 The X509_VERIFY_CTX structure has been enhanced considerably.
14442 All certificate lookup operations now go via a get_issuer()
14443 callback. Although this currently uses an X509_STORE it
14444 can be replaced by custom lookups. This is a simple way
14445 to bypass the X509_STORE hackery necessary to make this
14446 work and makes it possible to use more efficient techniques
14447 in future. A very simple version which uses a simple
14448 STACK for its trusted certificate store is also provided
14449 using X509_STORE_CTX_trusted_stack().
14451 The verify_cb() and verify() callbacks now have equivalents
14452 in the X509_STORE_CTX structure.
14454 X509_STORE_CTX also has a 'flags' field which can be used
14455 to customise the verify behaviour.
14459 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14460 excludes S/MIME capabilities.
14464 * When a certificate request is read in keep a copy of the
14465 original encoding of the signed data and use it when outputting
14466 again. Signatures then use the original encoding rather than
14467 a decoded, encoded version which may cause problems if the
14468 request is improperly encoded.
14472 * For consistency with other BIO_puts implementations, call
14473 buffer_write(b, ...) directly in buffer_puts instead of calling
14476 In BIO_puts, increment b->num_write as in BIO_write.
14478 *Peter.Sylvester@EdelWeb.fr*
14480 * Fix BN_mul_word for the case where the word is 0. (We have to use
14481 BN_zero, we may not return a BIGNUM with an array consisting of
14482 words set to zero.)
14486 * Avoid calling abort() from within the library when problems are
14487 detected, except if preprocessor symbols have been defined
14488 (such as REF_CHECK, BN_DEBUG etc.).
14492 * New openssl application 'rsautl'. This utility can be
14493 used for low level RSA operations. DER public key
14494 BIO/fp routines also added.
14498 * New Configure entry and patches for compiling on QNX 4.
14500 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14502 * A demo state-machine implementation was sponsored by
14503 Nuron (<http://www.nuron.com/>) and is now available in
14504 demos/state_machine.
14508 * New options added to the 'dgst' utility for signature
14509 generation and verification.
14513 * Unrecognized PKCS#7 content types are now handled via a
14514 catch all ASN1_TYPE structure. This allows unsupported
14515 types to be stored as a "blob" and an application can
14516 encode and decode it manually.
14520 * Fix various signed/unsigned issues to make a_strex.c
14521 compile under VC++.
14523 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14525 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14526 length if passed a buffer. ASN1_INTEGER_to_BN failed
14527 if passed a NULL BN and its argument was negative.
14529 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14531 * Modification to PKCS#7 encoding routines to output definite
14532 length encoding. Since currently the whole structures are in
14533 memory there's not real point in using indefinite length
14534 constructed encoding. However if OpenSSL is compiled with
14535 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14539 * Added BIO_vprintf() and BIO_vsnprintf().
14543 * Added more prefixes to parse for in the strings written
14544 through a logging bio, to cover all the levels that are available
14545 through syslog. The prefixes are now:
14547 PANIC, EMERG, EMR => LOG_EMERG
14548 ALERT, ALR => LOG_ALERT
14549 CRIT, CRI => LOG_CRIT
14550 ERROR, ERR => LOG_ERR
14551 WARNING, WARN, WAR => LOG_WARNING
14552 NOTICE, NOTE, NOT => LOG_NOTICE
14553 INFO, INF => LOG_INFO
14554 DEBUG, DBG => LOG_DEBUG
14556 and as before, if none of those prefixes are present at the
14557 beginning of the string, LOG_ERR is chosen.
14559 On Win32, the `LOG_*` levels are mapped according to this:
14561 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14562 LOG_WARNING => EVENTLOG_WARNING_TYPE
14563 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14567 * Made it possible to reconfigure with just the configuration
14568 argument "reconf" or "reconfigure". The command line arguments
14569 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14570 and are retrieved from there when reconfiguring.
14576 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14578 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14582 * The obj_dat.pl script was messing up the sorting of object
14583 names. The reason was that it compared the quoted version
14584 of strings as a result "OCSP" > "OCSP Signing" because
14585 " > SPACE. Changed script to store unquoted versions of
14586 names and add quotes on output. It was also omitting some
14587 names from the lookup table if they were given a default
14588 value (that is if SN is missing it is given the same
14589 value as LN and vice versa), these are now added on the
14590 grounds that if an object has a name we should be able to
14591 look it up. Finally added warning output when duplicate
14592 short or long names are found.
14596 * Changes needed for Tandem NSK.
14598 *Scott Uroff <scott@xypro.com>*
14600 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14601 RSA_padding_check_SSLv23(), special padding was never detected
14602 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14603 version rollback attacks was not effective.
14605 In s23_clnt.c, don't use special rollback-attack detection padding
14606 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14607 client; similarly, in s23_srvr.c, don't do the rollback check if
14608 SSL 2.0 is the only protocol enabled in the server.
14612 * Make it possible to get hexdumps of unprintable data with 'openssl
14613 asn1parse'. By implication, the functions ASN1_parse_dump() and
14614 BIO_dump_indent() are added.
14618 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14619 these print out strings and name structures based on various
14620 flags including RFC2253 support and proper handling of
14621 multibyte characters. Added options to the 'x509' utility
14622 to allow the various flags to be set.
14626 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14627 Also change the functions X509_cmp_current_time() and
14628 X509_gmtime_adj() work with an ASN1_TIME structure,
14629 this will enable certificates using GeneralizedTime in validity
14630 dates to be checked.
14634 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14635 negative public key encodings) on by default,
14636 NO_NEG_PUBKEY_BUG can be set to disable it.
14640 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14641 content octets. An i2c_ASN1_OBJECT is unnecessary because
14642 the encoding can be trivially obtained from the structure.
14646 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14647 not read locks (`CRYPTO_r_[un]lock`).
14651 * A first attempt at creating official support for shared
14652 libraries through configuration. I've kept it so the
14653 default is static libraries only, and the OpenSSL programs
14654 are always statically linked for now, but there are
14655 preparations for dynamic linking in place.
14656 This has been tested on Linux and Tru64.
14660 * Randomness polling function for Win9x, as described in:
14661 Peter Gutmann, Software Generation of Practically Strong
14666 * Fix so PRNG is seeded in req if using an already existing
14671 * New options to smime application. -inform and -outform
14672 allow alternative formats for the S/MIME message including
14673 PEM and DER. The -content option allows the content to be
14674 specified separately. This should allow things like Netscape
14675 form signing output easier to verify.
14679 * Fix the ASN1 encoding of tags using the 'long form'.
14683 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14684 STRING types. These convert content octets to and from the
14685 underlying type. The actual tag and length octets are
14686 already assumed to have been read in and checked. These
14687 are needed because all other string types have virtually
14688 identical handling apart from the tag. By having versions
14689 of the ASN1 functions that just operate on content octets
14690 IMPLICIT tagging can be handled properly. It also allows
14691 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14692 and ASN1_INTEGER are identical apart from the tag.
14696 * Change the handling of OID objects as follows:
14698 - New object identifiers are inserted in objects.txt, following
14699 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14700 - objects.pl is used to process obj_mac.num and create a new
14702 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14705 This is currently kind of a hack, and the perl code in objects.pl
14706 isn't very elegant, but it works as I intended. The simplest way
14707 to check that it worked correctly is to look in obj_dat.h and
14708 check the array nid_objs and make sure the objects haven't moved
14709 around (this is important!). Additions are OK, as well as
14710 consistent name changes.
14714 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14718 * Addition of the command line parameter '-rand file' to 'openssl req'.
14719 The given file adds to whatever has already been seeded into the
14720 random pool through the RANDFILE configuration file option or
14721 environment variable, or the default random state file.
14725 * mkstack.pl now sorts each macro group into lexical order.
14726 Previously the output order depended on the order the files
14727 appeared in the directory, resulting in needless rewriting
14732 * Patches to make OpenSSL compile under Win32 again. Mostly
14733 work arounds for the VC++ problem that it treats func() as
14734 func(void). Also stripped out the parts of mkdef.pl that
14735 added extra typesafe functions: these no longer exist.
14739 * Reorganisation of the stack code. The macros are now all
14740 collected in safestack.h . Each macro is defined in terms of
14741 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14742 DEBUG_SAFESTACK is now handled in terms of function casts,
14743 this has the advantage of retaining type safety without the
14744 use of additional functions. If DEBUG_SAFESTACK is not defined
14745 then the non typesafe macros are used instead. Also modified the
14746 mkstack.pl script to handle the new form. Needs testing to see
14747 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14748 the default if no major problems. Similar behaviour for ASN1_SET_OF
14749 and PKCS12_STACK_OF.
14753 * When some versions of IIS use the 'NET' form of private key the
14754 key derivation algorithm is different. Normally MD5(password) is
14755 used as a 128 bit RC4 key. In the modified case
14756 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14757 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14758 as the old Netscape_RSA functions except they have an additional
14759 'sgckey' parameter which uses the modified algorithm. Also added
14760 an -sgckey command line option to the rsa utility. Thanks to
14761 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14762 algorithm to openssl-dev.
14766 * The evp_local.h macros were using 'c.##kname' which resulted in
14767 invalid expansion on some systems (SCO 5.0.5 for example).
14768 Corrected to 'c.kname'.
14770 *Phillip Porch <root@theporch.com>*
14772 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14773 a STACK of email addresses from a certificate or request, these look
14774 in the subject name and the subject alternative name extensions and
14775 omit any duplicate addresses.
14779 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14780 This makes DSA verification about 2 % faster.
14784 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14785 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14786 plus overhead for 1024 bit moduli).
14787 This makes exponentiations about 0.5 % faster for 1024 bit
14788 exponents (as measured by "openssl speed rsa2048").
14792 * Rename memory handling macros to avoid conflicts with other
14794 Malloc => OPENSSL_malloc
14795 Malloc_locked => OPENSSL_malloc_locked
14796 Realloc => OPENSSL_realloc
14797 Free => OPENSSL_free
14801 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14802 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14806 * CygWin32 support.
14808 *John Jarvie <jjarvie@newsguy.com>*
14810 * The type-safe stack code has been rejigged. It is now only compiled
14811 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14812 by default all type-specific stack functions are "#define"d back to
14813 standard stack functions. This results in more streamlined output
14814 but retains the type-safety checking possibilities of the original
14819 * The STACK code has been cleaned up, and certain type declarations
14820 that didn't make a lot of sense have been brought in line. This has
14821 also involved a cleanup of sorts in safestack.h to more correctly
14822 map type-safe stack functions onto their plain stack counterparts.
14823 This work has also resulted in a variety of "const"ifications of
14824 lots of the code, especially `_cmp` operations which should normally
14825 be prototyped with "const" parameters anyway.
14829 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14830 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14831 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14832 where all of 'md' is used each time the PRNG is used, but 'state'
14833 is used only indexed by a cyclic counter. As entropy may not be
14834 well distributed from the beginning, 'md' is important as a
14835 chaining variable. However, the output function chains only half
14836 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14837 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14838 in all of 'state' being rewritten, with the new values depending
14839 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14843 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14844 the handshake is continued after ssl_verify_cert_chain();
14845 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14846 can lead to 'unexplainable' connection aborts later.
14848 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14850 * Major EVP API cipher revision.
14851 Add hooks for extra EVP features. This allows various cipher
14852 parameters to be set in the EVP interface. Support added for variable
14853 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14854 setting of RC2 and RC5 parameters.
14856 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14859 Remove lots of duplicated code from the EVP library. For example *every*
14860 cipher init() function handles the 'iv' in the same way according to the
14861 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14862 for CFB and OFB modes they zero ctx->num.
14864 New functionality allows removal of S/MIME code RC2 hack.
14866 Most of the routines have the same form and so can be declared in terms
14869 By shifting this to the top level EVP_CipherInit() it can be removed from
14870 all individual ciphers. If the cipher wants to handle IVs or keys
14871 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14874 Change lots of functions like EVP_EncryptUpdate() to now return a
14875 value: although software versions of the algorithms cannot fail
14876 any installed hardware versions can.
14880 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14881 this option is set, tolerate broken clients that send the negotiated
14882 protocol version number instead of the requested protocol version
14887 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14888 i.e. non-zero for export ciphersuites, zero otherwise.
14889 Previous versions had this flag inverted, inconsistent with
14890 rsa_tmp_cb (..._TMP_RSA_CB).
14892 *Bodo Moeller; problem reported by Amit Chopra*
14894 * Add missing DSA library text string. Work around for some IIS
14895 key files with invalid SEQUENCE encoding.
14899 * Add a document (doc/standards.txt) that list all kinds of standards
14900 and so on that are implemented in OpenSSL.
14904 * Enhance c_rehash script. Old version would mishandle certificates
14905 with the same subject name hash and wouldn't handle CRLs at all.
14906 Added -fingerprint option to crl utility, to support new c_rehash
14911 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14915 * Fix for SSL server purpose checking. Server checking was
14916 rejecting certificates which had extended key usage present
14917 but no ssl client purpose.
14919 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14921 * Make PKCS#12 code work with no password. The PKCS#12 spec
14922 is a little unclear about how a blank password is handled.
14923 Since the password in encoded as a BMPString with terminating
14924 double NULL a zero length password would end up as just the
14925 double NULL. However no password at all is different and is
14926 handled differently in the PKCS#12 key generation code. NS
14927 treats a blank password as zero length. MSIE treats it as no
14928 password on export: but it will try both on import. We now do
14929 the same: PKCS12_parse() tries zero length and no password if
14930 the password is set to "" or NULL (NULL is now a valid password:
14931 it wasn't before) as does the pkcs12 application.
14935 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14936 perror when PEM_read_bio_X509_REQ fails, the error message must
14937 be obtained from the error queue.
14941 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14942 it in ERR_remove_state if appropriate, and change ERR_get_state
14943 accordingly to avoid race conditions (this is necessary because
14944 thread_hash is no longer constant once set).
14948 * Bugfix for linux-elf makefile.one.
14952 * RSA_get_default_method() will now cause a default
14953 RSA_METHOD to be chosen if one doesn't exist already.
14954 Previously this was only set during a call to RSA_new()
14955 or RSA_new_method(NULL) meaning it was possible for
14956 RSA_get_default_method() to return NULL.
14960 * Added native name translation to the existing DSO code
14961 that will convert (if the flag to do so is set) filenames
14962 that are sufficiently small and have no path information
14963 into a canonical native form. Eg. "blah" converted to
14964 "libblah.so" or "blah.dll" etc.
14968 * New function ERR_error_string_n(e, buf, len) which is like
14969 ERR_error_string(e, buf), but writes at most 'len' bytes
14970 including the 0 terminator. For ERR_error_string_n, 'buf'
14973 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14975 * CONF library reworked to become more general. A new CONF
14976 configuration file reader "class" is implemented as well as a
14977 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14978 old `CONF_*` functions are still there, but are reimplemented to
14979 work in terms of the new functions. Also, a set of functions
14980 to handle the internal storage of the configuration data is
14981 provided to make it easier to write new configuration file
14982 reader "classes" (I can definitely see something reading a
14983 configuration file in XML format, for example), called `_CONF_*`,
14984 or "the configuration storage API"...
14986 The new configuration file reading functions are:
14988 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14989 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14991 NCONF_default, NCONF_WIN32
14993 NCONF_dump_fp, NCONF_dump_bio
14995 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14996 NCONF_new creates a new CONF object. This works in the same way
14997 as other interfaces in OpenSSL, like the BIO interface.
14998 `NCONF_dump_*` dump the internal storage of the configuration file,
14999 which is useful for debugging. All other functions take the same
15000 arguments as the old `CONF_*` functions with the exception of the
15001 first that must be a `CONF *` instead of a `LHASH *`.
15003 To make it easier to use the new classes with the old `CONF_*` functions,
15004 the function CONF_set_default_method is provided.
15008 * Add '-tls1' option to 'openssl ciphers', which was already
15009 mentioned in the documentation but had not been implemented.
15010 (This option is not yet really useful because even the additional
15011 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15015 * Initial DSO code added into libcrypto for letting OpenSSL (and
15016 OpenSSL-based applications) load shared libraries and bind to
15017 them in a portable way.
15019 *Geoff Thorpe, with contributions from Richard Levitte*
15021 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15023 * Make sure _lrotl and _lrotr are only used with MSVC.
15025 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15026 (the default implementation of RAND_status).
15028 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15029 to '-clrext' (= clear extensions), as intended and documented.
15030 *Bodo Moeller; inconsistency pointed out by Michael Attili
15031 <attili@amaxo.com>*
15033 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15034 was larger than the MD block size.
15036 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15038 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15039 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15040 using the passed key: if the passed key was a private key the result
15041 of X509_print(), for example, would be to print out all the private key
15046 * des_quad_cksum() byte order bug fix.
15047 *Ulf Möller, using the problem description in krb4-0.9.7, where
15048 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15050 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15053 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15055 * For easily testing in shell scripts whether some command
15056 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15057 returns with exit code 0 iff no command of the given name is available.
15058 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15059 the output goes to stdout and nothing is printed to stderr.
15060 Additional arguments are always ignored.
15062 Since for each cipher there is a command of the same name,
15063 the 'no-cipher' compilation switches can be tested this way.
15065 ('openssl no-XXX' is not able to detect pseudo-commands such
15066 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15070 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15074 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15075 is set; it will be thrown away anyway because each handshake creates
15077 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15078 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15079 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15080 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15084 * New s_client option -ign_eof: EOF at stdin is ignored, and
15085 'Q' and 'R' lose their special meanings (quit/renegotiate).
15086 This is part of what -quiet does; unlike -quiet, -ign_eof
15087 does not suppress any output.
15091 * Add compatibility options to the purpose and trust code. The
15092 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15093 accepts a certificate or CA, this was the previous behaviour,
15094 with all the associated security issues.
15096 X509_TRUST_COMPAT is the old trust behaviour: only and
15097 automatically trust self signed roots in certificate store. A
15098 new trust setting X509_TRUST_DEFAULT is used to specify that
15099 a purpose has no associated trust setting and it should instead
15100 use the value in the default purpose.
15104 * Fix the PKCS#8 DSA private key code so it decodes keys again
15105 and fix a memory leak.
15109 * In util/mkerr.pl (which implements 'make errors'), preserve
15110 reason strings from the previous version of the .c file, as
15111 the default to have only downcase letters (and digits) in
15112 automatically generated reasons codes is not always appropriate.
15116 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15117 using strerror. Previously, ERR_reason_error_string() returned
15118 library names as reason strings for SYSerr; but SYSerr is a special
15119 case where small numbers are errno values, not library numbers.
15123 * Add '-dsaparam' option to 'openssl dhparam' application. This
15124 converts DSA parameters into DH parameters. (When creating parameters,
15125 DSA_generate_parameters is used.)
15129 * Include 'length' (recommended exponent length) in C code generated
15130 by 'openssl dhparam -C'.
15134 * The second argument to set_label in perlasm was already being used
15135 so couldn't be used as a "file scope" flag. Moved to third argument
15140 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15141 instead of RAND_bytes for encryption IVs and salts.
15145 * Include RAND_status() into RAND_METHOD instead of implementing
15146 it only for md_rand.c Otherwise replacing the PRNG by calling
15147 RAND_set_rand_method would be impossible.
15151 * Don't let DSA_generate_key() enter an infinite loop if the random
15152 number generation fails.
15156 * New 'rand' application for creating pseudo-random output.
15160 * Added configuration support for Linux/IA64
15162 *Rolf Haberrecker <rolf@suse.de>*
15164 * Assembler module support for Mingw32.
15168 * Shared library support for HPUX (in shlib/).
15170 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15172 * Shared library support for Solaris gcc.
15174 *Lutz Behnke <behnke@trustcenter.de>*
15176 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15178 * PKCS7_encrypt() was adding text MIME headers twice because they
15179 were added manually and by SMIME_crlf_copy().
15183 * In bntest.c don't call BN_rand with zero bits argument.
15185 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15187 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15188 case was implemented. This caused BN_div_recp() to fail occasionally.
15192 * Add an optional second argument to the set_label() in the perl
15193 assembly language builder. If this argument exists and is set
15194 to 1 it signals that the assembler should use a symbol whose
15195 scope is the entire file, not just the current function. This
15196 is needed with MASM which uses the format label:: for this scope.
15198 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15200 * Change the ASN1 types so they are typedefs by default. Before
15201 almost all types were #define'd to ASN1_STRING which was causing
15202 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15207 * Change names of new functions to the new get1/get0 naming
15208 convention: After 'get1', the caller owns a reference count
15209 and has to call `..._free`; 'get0' returns a pointer to some
15210 data structure without incrementing reference counters.
15211 (Some of the existing 'get' functions increment a reference
15212 counter, some don't.)
15213 Similarly, 'set1' and 'add1' functions increase reference
15214 counters or duplicate objects.
15218 * Allow for the possibility of temp RSA key generation failure:
15219 the code used to assume it always worked and crashed on failure.
15223 * Fix potential buffer overrun problem in BIO_printf().
15224 *Ulf Möller, using public domain code by Patrick Powell; problem
15225 pointed out by David Sacerdote <das33@cornell.edu>*
15227 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15228 RAND_egd() and RAND_status(). In the command line application,
15229 the EGD socket can be specified like a seed file using RANDFILE
15234 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15235 Some CAs (e.g. Verisign) distribute certificates in this form.
15239 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15240 list to exclude them. This means that no special compilation option
15241 is needed to use anonymous DH: it just needs to be included in the
15246 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15247 EVP_MD_type. The old functionality is available in a new macro called
15248 EVP_MD_md(). Change code that uses it and update docs.
15252 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15253 where the `void *` argument is replaced by a function pointer argument.
15254 Previously `void *` was abused to point to functions, which works on
15255 many platforms, but is not correct. As these functions are usually
15256 called by macros defined in OpenSSL header files, most source code
15257 should work without changes.
15261 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15262 sections with information on -D... compiler switches used for
15263 compiling the library so that applications can see them. To enable
15264 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15265 must be defined. E.g.,
15266 #define OPENSSL_ALGORITHM_DEFINES
15267 #include <openssl/opensslconf.h>
15268 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15270 *Richard Levitte, Ulf and Bodo Möller*
15272 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15277 * Change the 'other' type in certificate aux info to a STACK_OF
15278 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15279 the required ASN1 format: arbitrary types determined by an OID.
15283 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15284 argument to 'req'. This is not because the function is newer or
15285 better than others it just uses the work 'NEW' in the certificate
15286 request header lines. Some software needs this.
15290 * Reorganise password command line arguments: now passwords can be
15291 obtained from various sources. Delete the PEM_cb function and make
15292 it the default behaviour: i.e. if the callback is NULL and the
15293 usrdata argument is not NULL interpret it as a null terminated pass
15294 phrase. If usrdata and the callback are NULL then the pass phrase
15295 is prompted for as usual.
15299 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15300 the support is automatically enabled. The resulting binaries will
15301 autodetect the card and use it if present.
15303 *Ben Laurie and Compaq Inc.*
15305 * Work around for Netscape hang bug. This sends certificate request
15306 and server done in one record. Since this is perfectly legal in the
15307 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15308 the bugs/SSLv3 entry for more info.
15312 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15316 * Add -rand argument to smime and pkcs12 applications and read/write
15321 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15325 * Add command line password options to the remaining applications.
15329 * Bug fix for BN_div_recp() for numerators with an even number of
15334 * More tests in bntest.c, and changed test_bn output.
15338 * ./config recognizes MacOS X now.
15342 * Bug fix for BN_div() when the first words of num and divisor are
15343 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15347 * Add support for various broken PKCS#8 formats, and command line
15348 options to produce them.
15352 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15353 get temporary BIGNUMs from a BN_CTX.
15357 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15362 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15363 include a #define from the old name to the new. The original intent
15364 was that statically linked binaries could for example just call
15365 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15366 link with digests. This never worked because SSLeay_add_all_digests()
15367 and SSLeay_add_all_ciphers() were in the same source file so calling
15368 one would link with the other. They are now in separate source files.
15372 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15376 * Use a less unusual form of the Miller-Rabin primality test (it used
15377 a binary algorithm for exponentiation integrated into the Miller-Rabin
15378 loop, our standard modexp algorithms are faster).
15382 * Support for the EBCDIC character set completed.
15384 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15386 * Source code cleanups: use const where appropriate, eliminate casts,
15387 use `void *` instead of `char *` in lhash.
15391 * Bugfix: ssl3_send_server_key_exchange was not restartable
15392 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15393 this the server could overwrite ephemeral keys that the client
15398 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15399 using 50 iterations of the Rabin-Miller test.
15401 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15402 iterations of the Rabin-Miller test as required by the appendix
15403 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15404 As BN_is_prime_fasttest includes trial division, DSA parameter
15405 generation becomes much faster.
15407 This implies a change for the callback functions in DSA_is_prime
15408 and DSA_generate_parameters: The callback function is called once
15409 for each positive witness in the Rabin-Miller test, not just
15410 occasionally in the inner loop; and the parameters to the
15411 callback function now provide an iteration count for the outer
15412 loop rather than for the current invocation of the inner loop.
15413 DSA_generate_parameters additionally can call the callback
15414 function with an 'iteration count' of -1, meaning that a
15415 candidate has passed the trial division test (when q is generated
15416 from an application-provided seed, trial division is skipped).
15420 * New function BN_is_prime_fasttest that optionally does trial
15421 division before starting the Rabin-Miller test and has
15422 an additional BN_CTX * argument (whereas BN_is_prime always
15423 has to allocate at least one BN_CTX).
15424 'callback(1, -1, cb_arg)' is called when a number has passed the
15425 trial division stage.
15429 * Fix for bug in CRL encoding. The validity dates weren't being handled
15434 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15438 * New function BN_pseudo_rand().
15442 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15443 bignum version of BN_from_montgomery() with the working code from
15444 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15449 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15450 made it impossible to use the same SSL_SESSION data structure in
15451 SSL2 clients in multiple threads.
15455 * The return value of RAND_load_file() no longer counts bytes obtained
15456 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15457 to seed the PRNG (previously an explicit byte count was required).
15459 *Ulf Möller, Bodo Möller*
15461 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15462 used `char *` instead of `void *` and had casts all over the place.
15466 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15470 * Retain source code compatibility for BN_prime_checks macro:
15471 BN_is_prime(..., BN_prime_checks, ...) now uses
15472 BN_prime_checks_for_size to determine the appropriate number of
15473 Rabin-Miller iterations.
15477 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15478 DH_CHECK_P_NOT_SAFE_PRIME.
15479 (Check if this is true? OpenPGP calls them "strong".)
15483 * Merge the functionality of "dh" and "gendh" programs into a new program
15484 "dhparam". The old programs are retained for now but will handle DH keys
15485 (instead of parameters) in future.
15489 * Make the ciphers, s_server and s_client programs check the return values
15490 when a new cipher list is set.
15494 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15495 ciphers. Before when the 56bit ciphers were enabled the sorting was
15498 The syntax for the cipher sorting has been extended to support sorting by
15499 cipher-strength (using the strength_bits hard coded in the tables).
15500 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15502 Fix a bug in the cipher-command parser: when supplying a cipher command
15503 string with an "undefined" symbol (neither command nor alphanumeric
15504 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15505 an error is flagged.
15507 Due to the strength-sorting extension, the code of the
15508 ssl_create_cipher_list() function was completely rearranged. I hope that
15509 the readability was also increased :-)
15511 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15513 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15514 for the first serial number and places 2 in the serial number file. This
15515 avoids problems when the root CA is created with serial number zero and
15516 the first user certificate has the same issuer name and serial number
15521 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15522 the new code. Add documentation for this stuff.
15526 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15527 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15528 structures and behave in an analogous way to the X509v3 functions:
15529 they shouldn't be called directly but wrapper functions should be used
15532 So we also now have some wrapper functions that call the X509at functions
15533 when passed certificate requests. (TO DO: similar things can be done with
15534 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15535 things. Some of these need some d2i or i2d and print functionality
15536 because they handle more complex structures.)
15540 * Add missing #ifndefs that caused missing symbols when building libssl
15541 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15542 NO_RSA in `ssl/s2*.c`.
15544 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15546 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15547 has a return value which indicates the quality of the random data
15548 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15549 error queue. New function RAND_pseudo_bytes() generates output that is
15550 guaranteed to be unique but not unpredictable. RAND_add is like
15551 RAND_seed, but takes an extra argument for an entropy estimate
15552 (RAND_seed always assumes full entropy).
15556 * Do more iterations of Rabin-Miller probable prime test (specifically,
15557 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15558 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15559 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15560 false-positive rate of at most 2^-80 for random input.
15564 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15568 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15569 in the 0.9.5 release), this returns the chain
15570 from an X509_CTX structure with a dup of the stack and all
15571 the X509 reference counts upped: so the stack will exist
15572 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15575 Also make SSL_SESSION_print() print out the verify return
15580 * Add manpage for the pkcs12 command. Also change the default
15581 behaviour so MAC iteration counts are used unless the new
15582 -nomaciter option is used. This improves file security and
15583 only older versions of MSIE (4.0 for example) need it.
15587 * Honor the no-xxx Configure options when creating .DEF files.
15591 * Add PKCS#10 attributes to field table: challengePassword,
15592 unstructuredName and unstructuredAddress. These are taken from
15593 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15594 international characters are used.
15596 More changes to X509_ATTRIBUTE code: allow the setting of types
15597 based on strings. Remove the 'loc' parameter when adding
15598 attributes because these will be a SET OF encoding which is sorted
15603 * Initial changes to the 'req' utility to allow request generation
15604 automation. This will allow an application to just generate a template
15605 file containing all the field values and have req construct the
15608 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15609 used all over the place including certificate requests and PKCS#7
15610 structures. They are currently handled manually where necessary with
15611 some primitive wrappers for PKCS#7. The new functions behave in a
15612 manner analogous to the X509 extension functions: they allow
15613 attributes to be looked up by NID and added.
15615 Later something similar to the X509V3 code would be desirable to
15616 automatically handle the encoding, decoding and printing of the
15617 more complex types. The string types like challengePassword can
15618 be handled by the string table functions.
15620 Also modified the multi byte string table handling. Now there is
15621 a 'global mask' which masks out certain types. The table itself
15622 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15623 is useful when for example there is only one permissible type
15624 (as in countryName) and using the mask might result in no valid
15629 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15630 SSL_get_peer_finished to allow applications to obtain the latest
15631 Finished messages sent to the peer or expected from the peer,
15632 respectively. (SSL_get_peer_finished is usually the Finished message
15633 actually received from the peer, otherwise the protocol will be aborted.)
15635 As the Finished message are message digests of the complete handshake
15636 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15637 be used for external authentication procedures when the authentication
15638 provided by SSL/TLS is not desired or is not enough.
15642 * Enhanced support for Alpha Linux is added. Now ./config checks if
15643 the host supports BWX extension and if Compaq C is present on the
15644 $PATH. Just exploiting of the BWX extension results in 20-30%
15645 performance kick for some algorithms, e.g. DES and RC4 to mention
15646 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15651 * Add support for MS "fast SGC". This is arguably a violation of the
15652 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15653 weak crypto and after checking the certificate is SGC a second one
15654 with strong crypto. MS SGC stops the first handshake after receiving
15655 the server certificate message and sends a second client hello. Since
15656 a server will typically do all the time consuming operations before
15657 expecting any further messages from the client (server key exchange
15658 is the most expensive) there is little difference between the two.
15660 To get OpenSSL to support MS SGC we have to permit a second client
15661 hello message after we have sent server done. In addition we have to
15662 reset the MAC if we do get this second client hello.
15666 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15667 if a DER encoded private key is RSA or DSA traditional format. Changed
15668 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15669 format DER encoded private key. Newer code should use PKCS#8 format which
15670 has the key type encoded in the ASN1 structure. Added DER private key
15671 support to pkcs8 application.
15675 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15676 ciphersuites has been selected (as required by the SSL 3/TLS 1
15677 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15678 is set, we interpret this as a request to violate the specification
15679 (the worst that can happen is a handshake failure, and 'correct'
15680 behaviour would result in a handshake failure anyway).
15684 * In SSL_CTX_add_session, take into account that there might be multiple
15685 SSL_SESSION structures with the same session ID (e.g. when two threads
15686 concurrently obtain them from an external cache).
15687 The internal cache can handle only one SSL_SESSION with a given ID,
15688 so if there's a conflict, we now throw out the old one to achieve
15693 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15694 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15695 some routines that use cipher OIDs: some ciphers do not have OIDs
15696 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15701 * Simplify the trust setting structure and code. Now we just have
15702 two sequences of OIDs for trusted and rejected settings. These will
15703 typically have values the same as the extended key usage extension
15704 and any application specific purposes.
15706 The trust checking code now has a default behaviour: it will just
15707 check for an object with the same NID as the passed id. Functions can
15708 be provided to override either the default behaviour or the behaviour
15709 for a given id. SSL client, server and email already have functions
15710 in place for compatibility: they check the NID and also return "trusted"
15711 if the certificate is self signed.
15715 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15716 traditional format into an EVP_PKEY structure.
15720 * Add a password callback function PEM_cb() which either prompts for
15721 a password if usr_data is NULL or otherwise assumes it is a null
15722 terminated password. Allow passwords to be passed on command line
15723 environment or config files in a few more utilities.
15727 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15728 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15729 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15730 Update documentation.
15734 * Support for ASN1 "NULL" type. This could be handled before by using
15735 ASN1_TYPE but there wasn't any function that would try to read a NULL
15736 and produce an error if it couldn't. For compatibility we also have
15737 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15738 don't allocate anything because they don't need to.
15742 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15745 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15747 * Rebuild of the memory allocation routines used by OpenSSL code and
15748 possibly others as well. The purpose is to make an interface that
15749 provide hooks so anyone can build a separate set of allocation and
15750 deallocation routines to be used by OpenSSL, for example memory
15751 pool implementations, or something else, which was previously hard
15752 since Malloc(), Realloc() and Free() were defined as macros having
15753 the values malloc, realloc and free, respectively (except for Win32
15754 compilations). The same is provided for memory debugging code.
15755 OpenSSL already comes with functionality to find memory leaks, but
15756 this gives people a chance to debug other memory problems.
15758 With these changes, a new set of functions and macros have appeared:
15760 CRYPTO_set_mem_debug_functions() [F]
15761 CRYPTO_get_mem_debug_functions() [F]
15762 CRYPTO_dbg_set_options() [F]
15763 CRYPTO_dbg_get_options() [F]
15764 CRYPTO_malloc_debug_init() [M]
15766 The memory debug functions are NULL by default, unless the library
15767 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15768 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15769 gives the standard debugging functions that come with OpenSSL) or
15770 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15771 provided by the library user) must be used. When the standard
15772 debugging functions are used, CRYPTO_dbg_set_options can be used to
15773 request additional information:
15774 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15775 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15777 Also, things like CRYPTO_set_mem_functions will always give the
15778 expected result (the new set of functions is used for allocation
15779 and deallocation) at all times, regardless of platform and compiler
15782 To finish it up, some functions that were never use in any other
15783 way than through macros have a new API and new semantic:
15785 CRYPTO_dbg_malloc()
15786 CRYPTO_dbg_realloc()
15789 All macros of value have retained their old syntax.
15791 *Richard Levitte and Bodo Moeller*
15793 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15794 ordering of SMIMECapabilities wasn't in "strength order" and there
15795 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15800 * Some ASN1 types with illegal zero length encoding (INTEGER,
15801 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15803 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15805 * Merge in my S/MIME library for OpenSSL. This provides a simple
15806 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15807 functionality to handle multipart/signed properly) and a utility
15808 called 'smime' to call all this stuff. This is based on code I
15809 originally wrote for Celo who have kindly allowed it to be
15810 included in OpenSSL.
15814 * Add variants des_set_key_checked and des_set_key_unchecked of
15815 des_set_key (aka des_key_sched). Global variable des_check_key
15816 decides which of these is called by des_set_key; this way
15817 des_check_key behaves as it always did, but applications and
15818 the library itself, which was buggy for des_check_key == 1,
15819 have a cleaner way to pick the version they need.
15823 * New function PKCS12_newpass() which changes the password of a
15828 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15829 dynamic mix. In both cases the ids can be used as an index into the
15830 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15831 functions so they accept a list of the field values and the
15832 application doesn't need to directly manipulate the X509_TRUST
15837 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15842 * Modify the way the V3 extension code looks up extensions. This now
15843 works in a similar way to the object code: we have some "standard"
15844 extensions in a static table which is searched with OBJ_bsearch()
15845 and the application can add dynamic ones if needed. The file
15846 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15847 updated whenever a new extension is added to the core code and kept
15848 in ext_nid order. There is a simple program 'tabtest.c' which checks
15849 this. New extensions are not added too often so this file can readily
15850 be maintained manually.
15852 There are two big advantages in doing things this way. The extensions
15853 can be looked up immediately and no longer need to be "added" using
15854 X509V3_add_standard_extensions(): this function now does nothing.
15855 Side note: I get *lots* of email saying the extension code doesn't
15856 work because people forget to call this function.
15857 Also no dynamic allocation is done unless new extensions are added:
15858 so if we don't add custom extensions there is no need to call
15859 X509V3_EXT_cleanup().
15863 * Modify enc utility's salting as follows: make salting the default. Add a
15864 magic header, so unsalted files fail gracefully instead of just decrypting
15865 to garbage. This is because not salting is a big security hole, so people
15866 should be discouraged from doing it.
15870 * Fixes and enhancements to the 'x509' utility. It allowed a message
15871 digest to be passed on the command line but it only used this
15872 parameter when signing a certificate. Modified so all relevant
15873 operations are affected by the digest parameter including the
15874 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15875 DSA key was used because it didn't fix the digest.
15879 * Initial certificate chain verify code. Currently tests the untrusted
15880 certificates for consistency with the verify purpose (which is set
15881 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15883 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15884 this is because it will reject chains with invalid extensions whereas
15885 every previous version of OpenSSL and SSLeay made no checks at all.
15887 Trust code: checks the root CA for the relevant trust settings. Trust
15888 settings have an initial value consistent with the verify purpose: e.g.
15889 if the verify purpose is for SSL client use it expects the CA to be
15890 trusted for SSL client use. However the default value can be changed to
15891 permit custom trust settings: one example of this would be to only trust
15892 certificates from a specific "secure" set of CAs.
15894 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15895 which should be used for version portability: especially since the
15896 verify structure is likely to change more often now.
15898 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15899 to set them. If not set then assume SSL clients will verify SSL servers
15902 Two new options to the verify program: -untrusted allows a set of
15903 untrusted certificates to be passed in and -purpose which sets the
15904 intended purpose of the certificate. If a purpose is set then the
15905 new chain verify code is used to check extension consistency.
15909 * Support for the authority information access extension.
15913 * Modify RSA and DSA PEM read routines to transparently handle
15914 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15915 public keys in a format compatible with certificate
15916 SubjectPublicKeyInfo structures. Unfortunately there were already
15917 functions called *_PublicKey_* which used various odd formats so
15918 these are retained for compatibility: however the DSA variants were
15919 never in a public release so they have been deleted. Changed dsa/rsa
15920 utilities to handle the new format: note no releases ever handled public
15921 keys so we should be OK.
15923 The primary motivation for this change is to avoid the same fiasco
15924 that dogs private keys: there are several incompatible private key
15925 formats some of which are standard and some OpenSSL specific and
15926 require various evil hacks to allow partial transparent handling and
15927 even then it doesn't work with DER formats. Given the option anything
15928 other than PKCS#8 should be dumped: but the other formats have to
15929 stay in the name of compatibility.
15931 With public keys and the benefit of hindsight one standard format
15932 is used which works with EVP_PKEY, RSA or DSA structures: though
15933 it clearly returns an error if you try to read the wrong kind of key.
15935 Added a -pubkey option to the 'x509' utility to output the public key.
15936 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15937 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15938 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15939 that do the same as the `EVP_PKEY_assign_*()` except they up the
15940 reference count of the added key (they don't "swallow" the
15945 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15946 CRLs would fail if the file contained no certificates or no CRLs:
15947 added a new function to read in both types and return the number
15948 read: this means that if none are read it will be an error. The
15949 DER versions of the certificate and CRL reader would always fail
15950 because it isn't possible to mix certificates and CRLs in DER format
15951 without choking one or the other routine. Changed this to just read
15952 a certificate: this is the best we can do. Also modified the code
15953 in apps/verify.c to take notice of return codes: it was previously
15954 attempting to read in certificates from NULL pointers and ignoring
15955 any errors: this is one reason why the cert and CRL reader seemed
15956 to work. It doesn't check return codes from the default certificate
15957 routines: these may well fail if the certificates aren't installed.
15961 * Code to support otherName option in GeneralName.
15965 * First update to verify code. Change the verify utility
15966 so it warns if it is passed a self signed certificate:
15967 for consistency with the normal behaviour. X509_verify
15968 has been modified to it will now verify a self signed
15969 certificate if *exactly* the same certificate appears
15970 in the store: it was previously impossible to trust a
15971 single self signed certificate. This means that:
15972 openssl verify ss.pem
15973 now gives a warning about a self signed certificate but
15974 openssl verify -CAfile ss.pem ss.pem
15979 * For servers, store verify_result in SSL_SESSION data structure
15980 (and add it to external session representation).
15981 This is needed when client certificate verifications fails,
15982 but an application-provided verification callback (set by
15983 SSL_CTX_set_cert_verify_callback) allows accepting the session
15984 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15985 but returns 1): When the session is reused, we have to set
15986 ssl->verify_result to the appropriate error code to avoid
15989 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15991 * Fix a bug in the new PKCS#7 code: it didn't consider the
15992 case in PKCS7_dataInit() where the signed PKCS7 structure
15993 didn't contain any existing data because it was being created.
15995 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15997 * Add a salt to the key derivation routines in enc.c. This
15998 forms the first 8 bytes of the encrypted file. Also add a
15999 -S option to allow a salt to be input on the command line.
16003 * New function X509_cmp(). Oddly enough there wasn't a function
16004 to compare two certificates. We do this by working out the SHA1
16005 hash and comparing that. X509_cmp() will be needed by the trust
16010 * SSL_get1_session() is like SSL_get_session(), but increments
16011 the reference count in the SSL_SESSION returned.
16013 *Geoff Thorpe <geoff@eu.c2.net>*
16015 * Fix for 'req': it was adding a null to request attributes.
16016 Also change the X509_LOOKUP and X509_INFO code to handle
16017 certificate auxiliary information.
16021 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16026 * Add the possibility to add extra information to the memory leak
16027 detecting output, to form tracebacks, showing from where each
16028 allocation was originated: CRYPTO_push_info("constant string") adds
16029 the string plus current file name and line number to a per-thread
16030 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16031 is like calling CYRPTO_pop_info() until the stack is empty.
16032 Also updated memory leak detection code to be multi-thread-safe.
16036 * Add options -text and -noout to pkcs7 utility and delete the
16037 encryption options which never did anything. Update docs.
16041 * Add options to some of the utilities to allow the pass phrase
16042 to be included on either the command line (not recommended on
16043 OSes like Unix) or read from the environment. Update the
16044 manpages and fix a few bugs.
16048 * Add a few manpages for some of the openssl commands.
16052 * Fix the -revoke option in ca. It was freeing up memory twice,
16053 leaking and not finding already revoked certificates.
16057 * Extensive changes to support certificate auxiliary information.
16058 This involves the use of X509_CERT_AUX structure and X509_AUX
16059 functions. An X509_AUX function such as PEM_read_X509_AUX()
16060 can still read in a certificate file in the usual way but it
16061 will also read in any additional "auxiliary information". By
16062 doing things this way a fair degree of compatibility can be
16063 retained: existing certificates can have this information added
16064 using the new 'x509' options.
16066 Current auxiliary information includes an "alias" and some trust
16067 settings. The trust settings will ultimately be used in enhanced
16068 certificate chain verification routines: currently a certificate
16069 can only be trusted if it is self signed and then it is trusted
16074 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16075 The problem was that one of the replacement routines had not been working
16076 since SSLeay releases. For now the offending routine has been replaced
16077 with non-optimised assembler. Even so, this now gives around 95%
16078 performance improvement for 1024 bit RSA signs.
16082 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16083 handling. Most clients have the effective key size in bits equal to
16084 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16085 A few however don't do this and instead use the size of the decrypted key
16086 to determine the RC2 key length and the AlgorithmIdentifier to determine
16087 the effective key length. In this case the effective key length can still
16088 be 40 bits but the key length can be 168 bits for example. This is fixed
16089 by manually forcing an RC2 key into the EVP_PKEY structure because the
16090 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16091 the key length and effective key length are equal.
16095 * Add a bunch of functions that should simplify the creation of
16096 X509_NAME structures. Now you should be able to do:
16097 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16098 and have it automatically work out the correct field type and fill in
16099 the structures. The more adventurous can try:
16100 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16101 and it will (hopefully) work out the correct multibyte encoding.
16105 * Change the 'req' utility to use the new field handling and multibyte
16106 copy routines. Before the DN field creation was handled in an ad hoc
16107 way in req, ca, and x509 which was rather broken and didn't support
16108 BMPStrings or UTF8Strings. Since some software doesn't implement
16109 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16110 using the dirstring_type option. See the new comment in the default
16111 openssl.cnf for more info.
16115 * Make crypto/rand/md_rand.c more robust:
16116 - Assure unique random numbers after fork().
16117 - Make sure that concurrent threads access the global counter and
16118 md serializably so that we never lose entropy in them
16119 or use exactly the same state in multiple threads.
16120 Access to the large state is not always serializable because
16121 the additional locking could be a performance killer, and
16122 md should be large enough anyway.
16126 * New file apps/app_rand.c with commonly needed functionality
16127 for handling the random seed file.
16129 Use the random seed file in some applications that previously did not:
16131 dsaparam -genkey (which also ignored its '-rand' option),
16134 x509 (when signing).
16135 Except on systems with /dev/urandom, it is crucial to have a random
16136 seed file at least for key creation, DSA signing, and for DH exchanges;
16137 for RSA signatures we could do without one.
16139 gendh and gendsa (unlike genrsa) used to read only the first byte
16140 of each file listed in the '-rand' option. The function as previously
16141 found in genrsa is now in app_rand.c and is used by all programs
16142 that support '-rand'.
16146 * In RAND_write_file, use mode 0600 for creating files;
16147 don't just chmod when it may be too late.
16151 * Report an error from X509_STORE_load_locations
16152 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16156 * New function ASN1_mbstring_copy() this copies a string in either
16157 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16158 into an ASN1_STRING type. A mask of permissible types is passed
16159 and it chooses the "minimal" type to use or an error if not type
16164 * Add function equivalents to the various macros in asn1.h. The old
16165 macros are retained with an `M_` prefix. Code inside the library can
16166 use the `M_` macros. External code (including the openssl utility)
16167 should *NOT* in order to be "shared library friendly".
16171 * Add various functions that can check a certificate's extensions
16172 to see if it usable for various purposes such as SSL client,
16173 server or S/MIME and CAs of these types. This is currently
16174 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16175 verification. Also added a -purpose flag to x509 utility to
16176 print out all the purposes.
16180 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16185 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16186 for, obtain and decode and extension and obtain its critical flag.
16187 This allows all the necessary extension code to be handled in a
16188 single function call.
16192 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16193 platforms. See crypto/rc4/rc4_enc.c for further details.
16197 * New -noout option to asn1parse. This causes no output to be produced
16198 its main use is when combined with -strparse and -out to extract data
16199 from a file (which may not be in ASN.1 format).
16203 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16204 when producing the local key id.
16206 *Richard Levitte <levitte@stacken.kth.se>*
16208 * New option -dhparam in s_server. This allows a DH parameter file to be
16209 stated explicitly. If it is not stated then it tries the first server
16210 certificate file. The previous behaviour hard coded the filename
16215 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16216 a public key to be input or output. For example:
16217 openssl rsa -in key.pem -pubout -out pubkey.pem
16218 Also added necessary DSA public key functions to handle this.
16222 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16223 in the message. This was handled by allowing
16224 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16226 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16228 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16229 to the end of the strings whereas this didn't. This would cause problems
16230 if strings read with d2i_ASN1_bytes() were later modified.
16232 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16234 * Fix for base64 decode bug. When a base64 bio reads only one line of
16235 data and it contains EOF it will end up returning an error. This is
16236 caused by input 46 bytes long. The cause is due to the way base64
16237 BIOs find the start of base64 encoded data. They do this by trying a
16238 trial decode on each line until they find one that works. When they
16239 do a flag is set and it starts again knowing it can pass all the
16240 data directly through the decoder. Unfortunately it doesn't reset
16241 the context it uses. This means that if EOF is reached an attempt
16242 is made to pass two EOFs through the context and this causes the
16243 resulting error. This can also cause other problems as well. As is
16244 usual with these problems it takes *ages* to find and the fix is
16245 trivial: move one line.
16247 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16249 * Ugly workaround to get s_client and s_server working under Windows. The
16250 old code wouldn't work because it needed to select() on sockets and the
16251 tty (for keypresses and to see if data could be written). Win32 only
16252 supports select() on sockets so we select() with a 1s timeout on the
16253 sockets and then see if any characters are waiting to be read, if none
16254 are present then we retry, we also assume we can always write data to
16255 the tty. This isn't nice because the code then blocks until we've
16256 received a complete line of data and it is effectively polling the
16257 keyboard at 1s intervals: however it's quite a bit better than not
16258 working at all :-) A dedicated Windows application might handle this
16259 with an event loop for example.
16263 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16264 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16265 will be called when RSA_sign() and RSA_verify() are used. This is useful
16266 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16267 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16268 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16269 This necessitated the support of an extra signature type NID_md5_sha1
16270 for SSL signatures and modifications to the SSL library to use it instead
16271 of calling RSA_public_decrypt() and RSA_private_encrypt().
16275 * Add new -verify -CAfile and -CApath options to the crl program, these
16276 will lookup a CRL issuers certificate and verify the signature in a
16277 similar way to the verify program. Tidy up the crl program so it
16278 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16279 less strict. It will now permit CRL extensions even if it is not
16280 a V2 CRL: this will allow it to tolerate some broken CRLs.
16284 * Initialize all non-automatic variables each time one of the openssl
16285 sub-programs is started (this is necessary as they may be started
16286 multiple times from the "OpenSSL>" prompt).
16288 *Lennart Bang, Bodo Moeller*
16290 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16291 removing all other RSA functionality (this is what NO_RSA does). This
16292 is so (for example) those in the US can disable those operations covered
16293 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16298 * Non-copying interface to BIO pairs.
16299 (still largely untested)
16303 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16304 ASCII string. This was handled independently in various places before.
16308 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16309 UTF8 strings a character at a time.
16313 * Use client_version from client hello to select the protocol
16314 (s23_srvr.c) and for RSA client key exchange verification
16315 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16319 * Add various utility functions to handle SPKACs, these were previously
16320 handled by poking round in the structure internals. Added new function
16321 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16322 print, verify and generate SPKACs. Based on an original idea from
16323 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16327 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16331 * Allow the config file extension section to be overwritten on the
16332 command line. Based on an original idea from Massimiliano Pala
16333 <madwolf@comune.modena.it>. The new option is called -extensions
16334 and can be applied to ca, req and x509. Also -reqexts to override
16335 the request extensions in req and -crlexts to override the crl extensions
16340 * Add new feature to the SPKAC handling in ca. Now you can include
16341 the same field multiple times by preceding it by "XXXX." for example:
16344 this is the same syntax as used in the req config file.
16348 * Allow certificate extensions to be added to certificate requests. These
16349 are specified in a 'req_extensions' option of the req section of the
16350 config file. They can be printed out with the -text option to req but
16351 are otherwise ignored at present.
16355 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16356 data read consists of only the final block it would not decrypted because
16357 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16358 A misplaced 'break' also meant the decrypted final block might not be
16359 copied until the next read.
16363 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16364 a few extra parameters to the DH structure: these will be useful if
16365 for example we want the value of 'q' or implement X9.42 DH.
16369 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16370 provides hooks that allow the default DSA functions or functions on a
16371 "per key" basis to be replaced. This allows hardware acceleration and
16372 hardware key storage to be handled without major modification to the
16373 library. Also added low level modexp hooks and CRYPTO_EX structure and
16374 associated functions.
16378 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16379 as "read only": it can't be written to and the buffer it points to will
16380 not be freed. Reading from a read only BIO is much more efficient than
16381 a normal memory BIO. This was added because there are several times when
16382 an area of memory needs to be read from a BIO. The previous method was
16383 to create a memory BIO and write the data to it, this results in two
16384 copies of the data and an O(n^2) reading algorithm. There is a new
16385 function BIO_new_mem_buf() which creates a read only memory BIO from
16386 an area of memory. Also modified the PKCS#7 routines to use read only
16391 * Bugfix: ssl23_get_client_hello did not work properly when called in
16392 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16393 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16394 but a retry condition occurred while trying to read the rest.
16398 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16399 NID_pkcs7_encrypted by default: this was wrong since this should almost
16400 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16401 the encrypted data type: this is a more sensible place to put it and it
16402 allows the PKCS#12 code to be tidied up that duplicated this
16407 * Changed obj_dat.pl script so it takes its input and output files on
16408 the command line. This should avoid shell escape redirection problems
16413 * Initial support for certificate extension requests, these are included
16414 in things like Xenroll certificate requests. Included functions to allow
16415 extensions to be obtained and added.
16419 * -crlf option to s_client and s_server for sending newlines as
16420 CRLF (as required by many protocols).
16424 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16426 * Install libRSAglue.a when OpenSSL is built with RSAref.
16428 *Ralf S. Engelschall*
16430 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16432 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16434 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16439 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16440 DH parameters/keys (q is lost during that conversion, but the resulting
16441 DH parameters contain its length).
16443 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16444 much faster than DH_generate_parameters (which creates parameters
16445 where `p = 2*q + 1`), and also the smaller q makes DH computations
16446 much more efficient (160-bit exponentiation instead of 1024-bit
16447 exponentiation); so this provides a convenient way to support DHE
16448 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16449 utter importance to use
16450 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16452 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16453 when such DH parameters are used, because otherwise small subgroup
16454 attacks may become possible!
16458 * Avoid memory leak in i2d_DHparams.
16462 * Allow the -k option to be used more than once in the enc program:
16463 this allows the same encrypted message to be read by multiple recipients.
16467 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16468 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16469 it will always use the numerical form of the OID, even if it has a short
16474 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16475 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16476 otherwise bn_mod_exp was called. In the case of hardware keys for example
16477 no private key components need be present and it might store extra data
16478 in the RSA structure, which cannot be accessed from bn_mod_exp.
16479 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16480 private key operations.
16484 * Added support for SPARC Linux.
16488 * pem_password_cb function type incompatibly changed from
16489 typedef int pem_password_cb(char *buf, int size, int rwflag);
16491 ....(char *buf, int size, int rwflag, void *userdata);
16492 so that applications can pass data to their callbacks:
16493 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16494 additional void * argument, which is just handed through whenever
16495 the password callback is called.
16497 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16499 New function SSL_CTX_set_default_passwd_cb_userdata.
16501 Compatibility note: As many C implementations push function arguments
16502 onto the stack in reverse order, the new library version is likely to
16503 interoperate with programs that have been compiled with the old
16504 pem_password_cb definition (PEM_whatever takes some data that
16505 happens to be on the stack as its last argument, and the callback
16506 just ignores this garbage); but there is no guarantee whatsoever that
16509 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16510 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16511 problems not only on Windows, but also on some Unix platforms.
16512 To avoid problematic command lines, these definitions are now in an
16513 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16514 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16518 * MIPS III/IV assembler module is reimplemented.
16522 * More DES library cleanups: remove references to srand/rand and
16523 delete an unused file.
16527 * Add support for the free Netwide assembler (NASM) under Win32,
16528 since not many people have MASM (ml) and it can be hard to obtain.
16529 This is currently experimental but it seems to work OK and pass all
16530 the tests. Check out INSTALL.W32 for info.
16534 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16535 without temporary keys kept an extra copy of the server key,
16536 and connections with temporary keys did not free everything in case
16541 * New function RSA_check_key and new openssl rsa option -check
16542 for verifying the consistency of RSA keys.
16544 *Ulf Moeller, Bodo Moeller*
16546 * Various changes to make Win32 compile work:
16547 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16548 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16549 comparison" warnings.
16550 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16554 * Add a debugging option to PKCS#5 v2 key generation function: when
16555 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16556 derived keys are printed to stderr.
16560 * Copy the flags in ASN1_STRING_dup().
16562 *Roman E. Pavlov <pre@mo.msk.ru>*
16564 * The x509 application mishandled signing requests containing DSA
16565 keys when the signing key was also DSA and the parameters didn't match.
16567 It was supposed to omit the parameters when they matched the signing key:
16568 the verifying software was then supposed to automatically use the CA's
16569 parameters if they were absent from the end user certificate.
16571 Omitting parameters is no longer recommended. The test was also
16572 the wrong way round! This was probably due to unusual behaviour in
16573 EVP_cmp_parameters() which returns 1 if the parameters match.
16574 This meant that parameters were omitted when they *didn't* match and
16575 the certificate was useless. Certificates signed with 'ca' didn't have
16578 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16580 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16581 The interface is as follows:
16582 Applications can use
16583 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16584 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16585 "off" is now the default.
16586 The library internally uses
16587 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16588 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16589 to disable memory-checking temporarily.
16591 Some inconsistent states that previously were possible (and were
16592 even the default) are now avoided.
16594 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16595 with each memory chunk allocated; this is occasionally more helpful
16596 than just having a counter.
16598 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16600 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16605 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16606 which largely parallels "options", but is for changing API behaviour,
16607 whereas "options" are about protocol behaviour.
16608 Initial "mode" flags are:
16610 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16611 a single record has been written.
16612 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16613 retries use the same buffer location.
16614 (But all of the contents must be
16619 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16622 * Fix problems with no-hmac etc.
16624 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16626 * New functions RSA_get_default_method(), RSA_set_method() and
16627 RSA_get_method(). These allows replacement of RSA_METHODs without having
16628 to mess around with the internals of an RSA structure.
16632 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16633 Also really enable memory leak checks in openssl.c and in some
16636 *Chad C. Mulligan, Bodo Moeller*
16638 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16639 up the length of negative integers. This has now been simplified to just
16640 store the length when it is first determined and use it later, rather
16641 than trying to keep track of where data is copied and updating it to
16643 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16645 * Add a new function PKCS7_signatureVerify. This allows the verification
16646 of a PKCS#7 signature but with the signing certificate passed to the
16647 function itself. This contrasts with PKCS7_dataVerify which assumes the
16648 certificate is present in the PKCS#7 structure. This isn't always the
16649 case: certificates can be omitted from a PKCS#7 structure and be
16650 distributed by "out of band" means (such as a certificate database).
16654 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16655 function prototypes in pem.h, also change util/mkdef.pl to add the
16656 necessary function names.
16660 * mk1mf.pl (used by Windows builds) did not properly read the
16661 options set by Configure in the top level Makefile, and Configure
16662 was not even able to write more than one option correctly.
16663 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16667 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16668 file to be loaded from a BIO or FILE pointer. The BIO version will
16669 for example allow memory BIOs to contain config info.
16673 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16674 Whoever hopes to achieve shared-library compatibility across versions
16675 must use this, not the compile-time macro.
16676 (Exercise 0.9.4: Which is the minimum library version required by
16678 Note: All this applies only to multi-threaded programs, others don't
16683 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16684 through a BIO pair triggered the default case, i.e.
16685 SSLerr(...,SSL_R_UNKNOWN_STATE).
16689 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16690 can use the SSL library even if none of the specific BIOs is
16695 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16696 for the encoded length.
16698 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16700 * Add initial documentation of the X509V3 functions.
16704 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16705 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16706 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16707 secure PKCS#8 private key format with a high iteration count.
16711 * Fix determination of Perl interpreter: A perl or perl5
16712 *directory* in $PATH was also accepted as the interpreter.
16714 *Ralf S. Engelschall*
16716 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16717 wrong with it but it was very old and did things like calling
16718 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16719 unusual formatting.
16723 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16724 to use the new extension code.
16728 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16729 with macros. This should make it easier to change their form, add extra
16730 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16735 * Add to configuration table a new entry that can specify an alternative
16736 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16737 according to Mark Crispin <MRC@Panda.COM>.
16741 * DES CBC did not update the IV. Weird.
16745 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16746 Changing the behaviour of the former might break existing programs --
16747 where IV updating is needed, des_ncbc_encrypt can be used.
16750 * When bntest is run from "make test" it drives bc to check its
16751 calculations, as well as internally checking them. If an internal check
16752 fails, it needs to cause bc to give a non-zero result or make test carries
16753 on without noticing the failure. Fixed.
16757 * DES library cleanups.
16761 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16762 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16763 ciphers. NOTE: although the key derivation function has been verified
16764 against some published test vectors it has not been extensively tested
16765 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16770 * Instead of "mkdir -p", which is not fully portable, use new
16771 Perl script "util/mkdir-p.pl".
16775 * Rewrite the way password based encryption (PBE) is handled. It used to
16776 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16777 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16778 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16779 the 'parameter' field of the AlgorithmIdentifier is passed to the
16780 underlying key generation function so it must do its own ASN1 parsing.
16781 This has also changed the EVP_PBE_CipherInit() function which now has a
16782 'parameter' argument instead of literal salt and iteration count values
16783 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16787 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16788 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16789 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16790 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16791 value was just used as a "magic string" and not used directly its
16792 value doesn't matter.
16796 * Introduce some semblance of const correctness to BN. Shame C doesn't
16801 * "linux-sparc64" configuration (ultrapenguin).
16803 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16804 "linux-sparc" configuration.
16806 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16808 * config now generates no-xxx options for missing ciphers.
16812 * Support the EBCDIC character set (work in progress).
16813 File ebcdic.c not yet included because it has a different license.
16815 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16817 * Support BS2000/OSD-POSIX.
16819 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16821 * Make callbacks for key generation use `void *` instead of `char *`.
16825 * Make S/MIME samples compile (not yet tested).
16829 * Additional typesafe stacks.
16833 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16837 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16839 * New configuration variant "sco5-gcc".
16841 * Updated some demos.
16843 *Sean O Riordain, Wade Scholine*
16845 * Add missing BIO_free at exit of pkcs12 application.
16849 * Fix memory leak in conf.c.
16853 * Updates for Win32 to assembler version of MD5.
16857 * Set #! path to perl in apps/der_chop to where we found it
16858 instead of using a fixed path.
16862 * SHA library changes for irix64-mips4-cc.
16866 * Improvements for VMS support.
16870 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16872 * Bignum library bug fix. IRIX 6 passes "make test" now!
16873 This also avoids the problems with SC4.2 and unpatched SC5.
16875 *Andy Polyakov <appro@fy.chalmers.se>*
16877 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16878 These are required because of the typesafe stack would otherwise break
16879 existing code. If old code used a structure member which used to be STACK
16880 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16881 sk_num or sk_value it would produce an error because the num, data members
16882 are not present in STACK_OF. Now it just produces a warning. sk_set
16883 replaces the old method of assigning a value to sk_value
16884 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16885 that does this will no longer work (and should use sk_set instead) but
16886 this could be regarded as a "questionable" behaviour anyway.
16890 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16891 correctly handle encrypted S/MIME data.
16895 * Change type of various DES function arguments from des_cblock
16896 (which means, in function argument declarations, pointer to char)
16897 to des_cblock * (meaning pointer to array with 8 char elements),
16898 which allows the compiler to do more typechecking; it was like
16899 that back in SSLeay, but with lots of ugly casts.
16901 Introduce new type const_des_cblock.
16905 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16906 problems: find RecipientInfo structure that matches recipient certificate
16907 and initialise the ASN1 structures properly based on passed cipher.
16911 * Belatedly make the BN tests actually check the results.
16915 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16916 to and from BNs: it was completely broken. New compilation option
16917 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16918 key elements as negative integers.
16922 * Reorganize and speed up MD5.
16924 *Andy Polyakov <appro@fy.chalmers.se>*
16928 *Richard Levitte <richard@levitte.org>*
16930 * New option -out to asn1parse to allow the parsed structure to be
16931 output to a file. This is most useful when combined with the -strparse
16932 option to examine the output of things like OCTET STRINGS.
16936 * Make SSL library a little more fool-proof by not requiring any longer
16937 that `SSL_set_{accept,connect}_state` be called before
16938 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16939 in many applications because usually everything *appeared* to work as
16940 intended anyway -- now it really works as intended).
16944 * Move openssl.cnf out of lib/.
16948 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16949 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16950 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16952 *Ralf S. Engelschall*
16954 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16955 handle PKCS#7 enveloped data properly.
16957 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16959 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16960 copying pointers. The cert_st handling is changed by this in
16961 various ways (and thus what used to be known as ctx->default_cert
16962 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16963 any longer when s->cert does not give us what we need).
16964 ssl_cert_instantiate becomes obsolete by this change.
16965 As soon as we've got the new code right (possibly it already is?),
16966 we have solved a couple of bugs of the earlier code where s->cert
16967 was used as if it could not have been shared with other SSL structures.
16969 Note that using the SSL API in certain dirty ways now will result
16970 in different behaviour than observed with earlier library versions:
16971 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16972 does not influence s as it used to.
16974 In order to clean up things more thoroughly, inside SSL_SESSION
16975 we don't use CERT any longer, but a new structure SESS_CERT
16976 that holds per-session data (if available); currently, this is
16977 the peer's certificate chain and, for clients, the server's certificate
16978 and temporary key. CERT holds only those values that can have
16979 meaningful defaults in an SSL_CTX.
16983 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16984 from the internal representation. Various PKCS#7 fixes: remove some
16985 evil casts and set the enc_dig_alg field properly based on the signing
16990 * Allow PKCS#12 password to be set from the command line or the
16991 environment. Let 'ca' get its config file name from the environment
16992 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16997 * Allow certificate policies extension to use an IA5STRING for the
16998 organization field. This is contrary to the PKIX definition but
16999 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17004 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17005 without disallowing inline assembler and the like for non-pedantic builds.
17009 * Support Borland C++ builder.
17011 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17017 * SHA-1 cleanups and performance enhancements.
17019 *Andy Polyakov <appro@fy.chalmers.se>*
17021 * Sparc v8plus assembler for the bignum library.
17023 *Andy Polyakov <appro@fy.chalmers.se>*
17025 * Accept any -xxx and +xxx compiler options in Configure.
17029 * Update HPUX configuration.
17033 * Add missing `sk_<type>_unshift()` function to safestack.h
17035 *Ralf S. Engelschall*
17037 * New function SSL_CTX_use_certificate_chain_file that sets the
17038 "extra_cert"s in addition to the certificate. (This makes sense
17039 only for "PEM" format files, as chains as a whole are not
17044 * Support verify_depth from the SSL API.
17045 x509_vfy.c had what can be considered an off-by-one-error:
17046 Its depth (which was not part of the external interface)
17047 was actually counting the number of certificates in a chain;
17048 now it really counts the depth.
17052 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17053 instead of X509err, which often resulted in confusing error
17054 messages since the error codes are not globally unique
17055 (e.g. an alleged error in ssl3_accept when a certificate
17056 didn't match the private key).
17058 * New function SSL_CTX_set_session_id_context that allows to set a default
17059 value (so that you don't need SSL_set_session_id_context for each
17060 connection using the SSL_CTX).
17064 * OAEP decoding bug fix.
17068 * Support INSTALL_PREFIX for package builders, as proposed by
17073 * New Configure options "threads" and "no-threads". For systems
17074 where the proper compiler options are known (currently Solaris
17075 and Linux), "threads" is the default.
17079 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17083 * Install various scripts to $(OPENSSLDIR)/misc, not to
17084 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17085 such as /usr/local/bin.
17089 * "make linux-shared" to build shared libraries.
17091 *Niels Poppe <niels@netbox.org>*
17093 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17097 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17098 extension adding in x509 utility.
17102 * Remove NOPROTO sections and error code comments.
17106 * Partial rewrite of the DEF file generator to now parse the ANSI
17111 * New Configure options --prefix=DIR and --openssldir=DIR.
17115 * Complete rewrite of the error code script(s). It is all now handled
17116 by one script at the top level which handles error code gathering,
17117 header rewriting and C source file generation. It should be much better
17118 than the old method: it now uses a modified version of Ulf's parser to
17119 read the ANSI prototypes in all header files (thus the old K&R definitions
17120 aren't needed for error creation any more) and do a better job of
17121 translating function codes into names. The old 'ASN1 error code embedded
17122 in a comment' is no longer necessary and it doesn't use .err files which
17123 have now been deleted. Also the error code call doesn't have to appear all
17124 on one line (which resulted in some large lines...).
17128 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17132 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17133 0 (which usually indicates a closed connection), but continue reading.
17137 * Fix some race conditions.
17141 * Add support for CRL distribution points extension. Add Certificate
17142 Policies and CRL distribution points documentation.
17146 * Move the autogenerated header file parts to crypto/opensslconf.h.
17150 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17151 8 of keying material. Merlin has also confirmed interop with this fix
17152 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17154 *Merlin Hughes <merlin@baltimore.ie>*
17156 * Fix lots of warnings.
17158 *Richard Levitte <levitte@stacken.kth.se>*
17160 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17161 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17163 *Richard Levitte <levitte@stacken.kth.se>*
17165 * Fix problems with sizeof(long) == 8.
17167 *Andy Polyakov <appro@fy.chalmers.se>*
17169 * Change functions to ANSI C.
17173 * Fix typos in error codes.
17175 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17177 * Remove defunct assembler files from Configure.
17181 * SPARC v8 assembler BIGNUM implementation.
17183 *Andy Polyakov <appro@fy.chalmers.se>*
17185 * Support for Certificate Policies extension: both print and set.
17186 Various additions to support the r2i method this uses.
17190 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17191 return a const string when you are expecting an allocated buffer.
17195 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17196 types DirectoryString and DisplayText.
17200 * Add code to allow r2i extensions to access the configuration database,
17201 add an LHASH database driver and add several ctx helper functions.
17205 * Fix an evil bug in bn_expand2() which caused various BN functions to
17206 fail when they extended the size of a BIGNUM.
17210 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17211 support typesafe stack.
17215 * Fix typo in SSL_[gs]et_options().
17217 *Nils Frostberg <nils@medcom.se>*
17219 * Delete various functions and files that belonged to the (now obsolete)
17220 old X509V3 handling code.
17224 * New Configure option "rsaref".
17228 * Don't auto-generate pem.h.
17232 * Introduce type-safe ASN.1 SETs.
17236 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17238 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17240 * Introduce type-safe STACKs. This will almost certainly break lots of code
17241 that links with OpenSSL (well at least cause lots of warnings), but fear
17242 not: the conversion is trivial, and it eliminates loads of evil casts. A
17243 few STACKed things have been converted already. Feel free to convert more.
17244 In the fullness of time, I'll do away with the STACK type altogether.
17248 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17249 specified in `<certfile>` by updating the entry in the index.txt file.
17250 This way one no longer has to edit the index.txt file manually for
17251 revoking a certificate. The -revoke option does the gory details now.
17253 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17255 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17256 `-text` option at all and this way the `-noout -text` combination was
17257 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17259 *Ralf S. Engelschall*
17261 * Make sure a corresponding plain text error message exists for the
17262 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17263 verify callback function determined that a certificate was revoked.
17265 *Ralf S. Engelschall*
17267 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17268 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17269 all available ciphers including rc5, which was forgotten until now.
17270 In order to let the testing shell script know which algorithms
17271 are available, a new (up to now undocumented) command
17272 `openssl list-cipher-commands` is used.
17276 * Bugfix: s_client occasionally would sleep in select() when
17277 it should have checked SSL_pending() first.
17281 * New functions DSA_do_sign and DSA_do_verify to provide access to
17282 the raw DSA values prior to ASN.1 encoding.
17286 * Tweaks to Configure
17288 *Niels Poppe <niels@netbox.org>*
17290 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17295 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17299 * New config option to avoid instructions that are illegal on the 80386.
17300 The default code is faster, but requires at least a 486.
17304 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17305 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17306 same as SSL2_VERSION anyway.
17310 * New "-showcerts" option for s_client.
17314 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17315 application. Various cleanups and fixes.
17319 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17320 modify error routines to work internally. Add error codes and PBE init
17321 to library startup routines.
17325 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17326 packing functions to asn1 and evp. Changed function names and error
17327 codes along the way.
17331 * PKCS12 integration: and so it begins... First of several patches to
17332 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17333 objects to objects.h
17337 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17338 and display support for Thawte strong extranet extension.
17342 * Add LinuxPPC support.
17344 *Jeff Dubrule <igor@pobox.org>*
17346 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17347 bn_div_words in alpha.s.
17349 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17351 * Make sure the RSA OAEP test is skipped under -DRSAref because
17352 OAEP isn't supported when OpenSSL is built with RSAref.
17354 *Ulf Moeller <ulf@fitug.de>*
17356 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17357 so they no longer are missing under -DNOPROTO.
17359 *Soren S. Jorvang <soren@t.dk>*
17361 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17363 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17364 doesn't work when the session is reused. Coming soon!
17368 * Fix a security hole, that allows sessions to be reused in the wrong
17369 context thus bypassing client cert protection! All software that uses
17370 client certs and session caches in multiple contexts NEEDS PATCHING to
17371 allow session reuse! A fuller solution is in the works.
17373 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17375 * Some more source tree cleanups (removed obsolete files
17376 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17377 permission on "config" script to be executable) and a fix for the INSTALL
17380 *Ulf Moeller <ulf@fitug.de>*
17382 * Remove some legacy and erroneous uses of malloc, free instead of
17385 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17387 * Make rsa_oaep_test return non-zero on error.
17389 *Ulf Moeller <ulf@fitug.de>*
17391 * Add support for native Solaris shared libraries. Configure
17392 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17393 if someone would make that last step automatic.
17395 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17397 * ctx_size was not built with the right compiler during "make links". Fixed.
17401 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17402 except NULL ciphers". This means the default cipher list will no longer
17403 enable NULL ciphers. They need to be specifically enabled e.g. with
17404 the string "DEFAULT:eNULL".
17408 * Fix to RSA private encryption routines: if p < q then it would
17409 occasionally produce an invalid result. This will only happen with
17410 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17414 * Be less restrictive and allow also `perl util/perlpath.pl
17415 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17416 because this way one can also use an interpreter named `perl5` (which is
17417 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17418 installed as `perl`).
17420 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17422 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17424 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17426 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17427 advapi32.lib to Win32 build and change the pem test comparison
17428 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17429 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17430 and crypto/des/ede_cbcm_enc.c.
17434 * DES quad checksum was broken on big-endian architectures. Fixed.
17438 * Comment out two functions in bio.h that aren't implemented. Fix up the
17439 Win32 test batch file so it (might) work again. The Win32 test batch file
17440 is horrible: I feel ill....
17444 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17445 in e_os.h. Audit of header files to check ANSI and non ANSI
17446 sections: 10 functions were absent from non ANSI section and not exported
17447 from Windows DLLs. Fixed up libeay.num for new functions.
17451 * Make `openssl version` output lines consistent.
17453 *Ralf S. Engelschall*
17455 * Fix Win32 symbol export lists for BIO functions: Added
17456 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17457 to ms/libeay{16,32}.def.
17459 *Ralf S. Engelschall*
17461 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17462 fine under Unix and passes some trivial tests I've now added. But the
17463 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17464 added to make sure no one expects that this stuff really works in the
17465 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17466 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17469 *Ralf S. Engelschall*
17471 * Fix the generation of two part addresses in perl.
17473 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17475 * Add config entry for Linux on MIPS.
17477 *John Tobey <jtobey@channel1.com>*
17479 * Make links whenever Configure is run, unless we are on Windoze.
17483 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17484 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17489 * Add a useful kludge to allow package maintainers to specify compiler and
17490 other platforms details on the command line without having to patch the
17491 Configure script every time: One now can use
17492 `perl Configure <id>:<details>`,
17493 i.e. platform ids are allowed to have details appended
17494 to them (separated by colons). This is treated as there would be a static
17495 pre-configured entry in Configure's %table under key `<id>` with value
17496 `<details>` and `perl Configure <id>` is called. So, when you want to
17497 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17498 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17499 now, which overrides the FreeBSD-elf entry on-the-fly.
17501 *Ralf S. Engelschall*
17503 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17507 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17508 on the `perl Configure ...` command line. This way one can compile
17509 OpenSSL libraries with Position Independent Code (PIC) which is needed
17510 for linking it into DSOs.
17512 *Ralf S. Engelschall*
17514 * Remarkably, export ciphers were totally broken and no-one had noticed!
17519 * Cleaned up the LICENSE document: The official contact for any license
17520 questions now is the OpenSSL core team under openssl-core@openssl.org.
17521 And add a paragraph about the dual-license situation to make sure people
17522 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17523 to the OpenSSL toolkit.
17525 *Ralf S. Engelschall*
17527 * General source tree makefile cleanups: Made `making xxx in yyy...`
17528 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17529 Additionally cleaned up the `make links` target: Remove unnecessary
17530 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17531 to speed processing and no longer clutter the display with confusing
17532 stuff. Instead only the actually done links are displayed.
17534 *Ralf S. Engelschall*
17536 * Permit null encryption ciphersuites, used for authentication only. It used
17537 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17538 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17543 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17544 signed attributes when verifying signatures (this would break them),
17545 the detached data encoding was wrong and public keys obtained using
17546 X509_get_pubkey() weren't freed.
17550 * Add text documentation for the BUFFER functions. Also added a work around
17551 to a Win95 console bug. This was triggered by the password read stuff: the
17552 last character typed gets carried over to the next fread(). If you were
17553 generating a new cert request using 'req' for example then the last
17554 character of the passphrase would be CR which would then enter the first
17559 * Added the new 'Includes OpenSSL Cryptography Software' button as
17560 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17561 button and can be used by applications based on OpenSSL to show the
17562 relationship to the OpenSSL project.
17564 *Ralf S. Engelschall*
17566 * Remove confusing variables in function signatures in files
17567 ssl/ssl_lib.c and ssl/ssl.h.
17569 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17571 * Don't install bss_file.c under PREFIX/include/
17573 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17575 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17576 functions that return function pointers and has support for NT specific
17577 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17578 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17579 unsigned to signed types: this was killing the Win32 compile.
17583 * Add new certificate file to stack functions,
17584 SSL_add_dir_cert_subjects_to_stack() and
17585 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17586 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17587 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17588 This means that Apache-SSL and similar packages don't have to mess around
17589 to add as many CAs as they want to the preferred list.
17593 * Experiment with doxygen documentation. Currently only partially applied to
17595 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17596 openssl.doxy as the configuration file.
17600 * Get rid of remaining C++-style comments which strict C compilers hate.
17602 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17604 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17605 compiled in by default: it has problems with large keys.
17609 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17610 DH private keys and/or callback functions which directly correspond to
17611 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17612 is needed for applications which have to configure certificates on a
17613 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17615 For the RSA certificate situation is makes no difference, but
17616 for the DSA certificate situation this fixes the "no shared cipher"
17617 problem where the OpenSSL cipher selection procedure failed because the
17618 temporary keys were not overtaken from the context and the API provided
17619 no way to reconfigure them.
17620 The new functions now let applications reconfigure the stuff and they
17621 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17622 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17623 non-public-API function ssl_cert_instantiate() is used as a helper
17624 function and also to reduce code redundancy inside ssl_rsa.c.
17626 *Ralf S. Engelschall*
17628 * Move s_server -dcert and -dkey options out of the undocumented feature
17629 area because they are useful for the DSA situation and should be
17630 recognized by the users.
17632 *Ralf S. Engelschall*
17634 * Fix the cipher decision scheme for export ciphers: the export bits are
17635 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17636 SSL_EXP_MASK. So, the original variable has to be used instead of the
17637 already masked variable.
17639 *Richard Levitte <levitte@stacken.kth.se>*
17641 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17643 *Richard Levitte <levitte@stacken.kth.se>*
17645 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17646 from `int` to `unsigned int` because it is a length and initialized by
17647 EVP_DigestFinal() which expects an `unsigned int *`.
17649 *Richard Levitte <levitte@stacken.kth.se>*
17651 * Don't hard-code path to Perl interpreter on shebang line of Configure
17652 script. Instead use the usual Shell->Perl transition trick.
17654 *Ralf S. Engelschall*
17656 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17657 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17658 -noout -modulus` as it's already the case for `openssl rsa -noout
17659 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17660 currently the public key is printed (a decision which was already done by
17661 `openssl dsa -modulus` in the past) which serves a similar purpose.
17662 Additionally the NO_RSA no longer completely removes the whole -modulus
17663 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17666 *Ralf S. Engelschall*
17668 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17669 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17671 *Arne Ansper <arne@ats.cyber.ee>*
17673 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17674 to be added. Now both 'req' and 'ca' can use new objects defined in the
17679 * Add cool BIO that does syslog (or event log on NT).
17681 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17683 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17684 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17685 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17686 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17690 * Add preliminary config info for new extension code.
17694 * Make RSA_NO_PADDING really use no padding.
17696 *Ulf Moeller <ulf@fitug.de>*
17698 * Generate errors when private/public key check is done.
17702 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17703 for some CRL extensions and new objects added.
17707 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17708 key usage extension and fuller support for authority key id.
17712 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17713 padding method for RSA, which is recommended for new applications in PKCS
17714 #1 v2.0 (RFC 2437, October 1998).
17715 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17716 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17717 against Bleichbacher's attack on RSA.
17718 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17721 * Updates to the new SSL compression code
17723 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17725 * Fix so that the version number in the master secret, when passed
17726 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17727 (because the server will not accept higher), that the version number
17728 is 0x03,0x01, not 0x03,0x00
17730 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17732 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17733 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17734 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17738 * Support for RAW extensions where an arbitrary extension can be
17739 created by including its DER encoding. See apps/openssl.cnf for
17744 * Make sure latest Perl versions don't interpret some generated C array
17745 code as Perl array code in the crypto/err/err_genc.pl script.
17747 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17749 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17750 not many people have the assembler. Various Win32 compilation fixes and
17751 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17752 build instructions.
17756 * Modify configure script 'Configure' to automatically create crypto/date.h
17757 file under Win32 and also build pem.h from pem.org. New script
17758 util/mkfiles.pl to create the MINFO file on environments that can't do a
17759 'make files': perl util/mkfiles.pl >MINFO should work.
17763 * Major rework of DES function declarations, in the pursuit of correctness
17764 and purity. As a result, many evil casts evaporated, and some weirdness,
17765 too. You may find this causes warnings in your code. Zapping your evil
17766 casts will probably fix them. Mostly.
17770 * Fix for a typo in asn1.h. Bug fix to object creation script
17771 obj_dat.pl. It considered a zero in an object definition to mean
17772 "end of object": none of the objects in objects.h have any zeros
17773 so it wasn't spotted.
17775 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17777 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17778 Masking (CBCM). In the absence of test vectors, the best I have been able
17779 to do is check that the decrypt undoes the encrypt, so far. Send me test
17780 vectors if you have them.
17784 * Correct calculation of key length for export ciphers (too much space was
17785 allocated for null ciphers). This has not been tested!
17789 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17790 message is now correct (it understands "crypto" and "ssl" on its
17791 command line). There is also now an "update" option. This will update
17792 the util/ssleay.num and util/libeay.num files with any new functions.
17794 perl util/mkdef.pl crypto ssl update
17795 it will update them.
17799 * Overhauled the Perl interface:
17800 - ported BN stuff to OpenSSL's different BN library
17801 - made the perl/ source tree CVS-aware
17802 - renamed the package from SSLeay to OpenSSL (the files still contain
17803 their history because I've copied them in the repository)
17804 - removed obsolete files (the test scripts will be replaced
17805 by better Test::Harness variants in the future)
17807 *Ralf S. Engelschall*
17809 * First cut for a very conservative source tree cleanup:
17810 1. merge various obsolete readme texts into doc/ssleay.txt
17811 where we collect the old documents and readme texts.
17812 2. remove the first part of files where I'm already sure that we no
17813 longer need them because of three reasons: either they are just temporary
17814 files which were left by Eric or they are preserved original files where
17815 I've verified that the diff is also available in the CVS via "cvs diff
17816 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17817 the crypto/md/ stuff).
17819 *Ralf S. Engelschall*
17821 * More extension code. Incomplete support for subject and issuer alt
17822 name, issuer and authority key id. Change the i2v function parameters
17823 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17824 what that's for :-) Fix to ASN1 macro which messed up
17825 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17829 * Preliminary support for ENUMERATED type. This is largely copied from the
17834 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17836 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17838 * Make sure `make rehash` target really finds the `openssl` program.
17840 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17842 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17843 like to hear about it if this slows down other processors.
17847 * Add CygWin32 platform information to Configure script.
17849 *Alan Batie <batie@aahz.jf.intel.com>*
17851 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17853 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17855 * New program nseq to manipulate netscape certificate sequences
17859 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17864 * Fixes to BN code. Previously the default was to define BN_RECURSION
17865 but the BN code had some problems that would cause failures when
17866 doing certificate verification and some other functions.
17868 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17870 * Add ASN1 and PEM code to support netscape certificate sequences.
17874 * Add ASN1 and PEM code to support netscape certificate sequences.
17878 * Add several PKIX and private extended key usage OIDs.
17882 * Modify the 'ca' program to handle the new extension code. Modify
17883 openssl.cnf for new extension format, add comments.
17887 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17888 and add a sample to openssl.cnf so req -x509 now adds appropriate
17893 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17894 error code, add initial support to X509_print() and x509 application.
17898 * Takes a deep breath and start adding X509 V3 extension support code. Add
17899 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17900 stuff is currently isolated and isn't even compiled yet.
17904 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17905 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17906 Removed the versions check from X509 routines when loading extensions:
17907 this allows certain broken certificates that don't set the version
17908 properly to be processed.
17912 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17913 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17914 can still be regenerated with "make depend".
17918 * Spelling mistake in C version of CAST-128.
17920 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17922 * Changes to the error generation code. The perl script err-code.pl
17923 now reads in the old error codes and retains the old numbers, only
17924 adding new ones if necessary. It also only changes the .err files if new
17925 codes are added. The makefiles have been modified to only insert errors
17926 when needed (to avoid needlessly modifying header files). This is done
17927 by only inserting errors if the .err file is newer than the auto generated
17928 C file. To rebuild all the error codes from scratch (the old behaviour)
17929 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17930 or delete all the .err files.
17934 * CAST-128 was incorrectly implemented for short keys. The C version has
17935 been fixed, but is untested. The assembler versions are also fixed, but
17936 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17937 to regenerate it if needed.
17938 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17939 Hagino <itojun@kame.net>*
17941 * File was opened incorrectly in randfile.c.
17943 *Ulf Möller <ulf@fitug.de>*
17945 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17946 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17947 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17948 al: it's just almost always a UTCTime. Note this patch adds new error
17949 codes so do a "make errors" if there are problems.
17953 * Correct Linux 1 recognition in config.
17955 *Ulf Möller <ulf@fitug.de>*
17957 * Remove pointless MD5 hash when using DSA keys in ca.
17959 *Anonymous <nobody@replay.com>*
17961 * Generate an error if given an empty string as a cert directory. Also
17962 generate an error if handed NULL (previously returned 0 to indicate an
17963 error, but didn't set one).
17965 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17967 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17971 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17972 parameters. This was causing a warning which killed off the Win32 compile.
17976 * Remove C++ style comments from crypto/bn/bn_local.h.
17978 *Neil Costigan <neil.costigan@celocom.com>*
17980 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17981 based on a text string, looking up short and long names and finally
17982 "dot" format. The "dot" format stuff didn't work. Added new function
17983 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17984 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17985 OID is not part of the table.
17989 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17990 X509_LOOKUP_by_alias().
17994 * Sort openssl functions by name.
17998 * Get the gendsa program working (hopefully) and add it to app list. Remove
17999 encryption from sample DSA keys (in case anyone is interested the password
18004 * Make *all* `*_free` functions accept a NULL pointer.
18006 *Frans Heymans <fheymans@isaserver.be>*
18008 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18011 *Anonymous <nobody@replay.com>*
18013 * s_server should send the CAfile as acceptable CAs, not its own cert.
18015 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18017 * Don't blow it for numeric -newkey arguments to apps/req.
18019 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18021 * Temp key "for export" tests were wrong in s3_srvr.c.
18023 *Anonymous <nobody@replay.com>*
18025 * Add prototype for temp key callback functions
18026 SSL_CTX_set_tmp_{rsa,dh}_callback().
18030 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18031 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18035 * X509_name_add_entry() freed the wrong thing after an error.
18037 *Arne Ansper <arne@ats.cyber.ee>*
18039 * rsa_eay.c would attempt to free a NULL context.
18041 *Arne Ansper <arne@ats.cyber.ee>*
18043 * BIO_s_socket() had a broken should_retry() on Windoze.
18045 *Arne Ansper <arne@ats.cyber.ee>*
18047 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18049 *Arne Ansper <arne@ats.cyber.ee>*
18051 * Make sure the already existing X509_STORE->depth variable is initialized
18052 in X509_STORE_new(), but document the fact that this variable is still
18053 unused in the certificate verification process.
18055 *Ralf S. Engelschall*
18057 * Fix the various library and apps files to free up pkeys obtained from
18058 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18062 * Fix reference counting in X509_PUBKEY_get(). This makes
18063 demos/maurice/example2.c work, amongst others, probably.
18065 *Steve Henson and Ben Laurie*
18067 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18068 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18069 are no longer created. This way we have a single and consistent command
18070 line interface `openssl <command>`, similar to `cvs <command>`.
18072 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18074 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18075 BIT STRING wrapper always have zero unused bits.
18079 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18083 * Make the top-level INSTALL documentation easier to understand.
18087 * Makefiles updated to exit if an error occurs in a sub-directory
18088 make (including if user presses ^C) [Paul Sutton]
18090 * Make Montgomery context stuff explicit in RSA data structure.
18094 * Fix build order of pem and err to allow for generated pem.h.
18098 * Fix renumbering bug in X509_NAME_delete_entry().
18102 * Enhanced the err-ins.pl script so it makes the error library number
18103 global and can add a library name. This is needed for external ASN1 and
18104 other error libraries.
18108 * Fixed sk_insert which never worked properly.
18112 * Fix ASN1 macros so they can handle indefinite length constructed
18113 EXPLICIT tags. Some non standard certificates use these: they can now
18118 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18119 into a single doc/ssleay.txt bundle. This way the information is still
18120 preserved but no longer messes up this directory. Now it's new room for
18121 the new set of documentation files.
18123 *Ralf S. Engelschall*
18125 * SETs were incorrectly DER encoded. This was a major pain, because they
18126 shared code with SEQUENCEs, which aren't coded the same. This means that
18127 almost everything to do with SETs or SEQUENCEs has either changed name or
18128 number of arguments.
18130 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18132 * Fix test data to work with the above.
18136 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18137 was already fixed by Eric for 0.9.1 it seems.
18139 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18141 * Autodetect FreeBSD3.
18145 * Fix various bugs in Configure. This affects the following platforms:
18149 unixware-2.0-pentium
18154 * Eliminate generated files from CVS. Reorder tests to regenerate files
18155 before they are needed.
18159 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18163 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18165 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18166 changed SSLeay to OpenSSL in version strings.
18168 *Ralf S. Engelschall*
18170 * Some fixups to the top-level documents.
18174 * Fixed the nasty bug where rsaref.h was not found under compile-time
18175 because the symlink to include/ was missing.
18177 *Ralf S. Engelschall*
18179 * Incorporated the popular no-RSA/DSA-only patches
18180 which allow to compile a RSA-free SSLeay.
18182 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18184 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18185 when "ssleay" is still not found.
18187 *Ralf S. Engelschall*
18189 * Added more platforms to Configure: Cray T3E, HPUX 11,
18191 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18193 * Updated the README file.
18195 *Ralf S. Engelschall*
18197 * Added various .cvsignore files in the CVS repository subdirs
18198 to make a "cvs update" really silent.
18200 *Ralf S. Engelschall*
18202 * Recompiled the error-definition header files and added
18203 missing symbols to the Win32 linker tables.
18205 *Ralf S. Engelschall*
18207 * Cleaned up the top-level documents;
18208 o new files: CHANGES and LICENSE
18209 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18210 o merged COPYRIGHT into LICENSE
18211 o removed obsolete TODO file
18212 o renamed MICROSOFT to INSTALL.W32
18214 *Ralf S. Engelschall*
18216 * Removed dummy files from the 0.9.1b source tree:
18217 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18218 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18219 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18220 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18221 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18223 *Ralf S. Engelschall*
18225 * Added various platform portability fixes.
18229 * The Genesis of the OpenSSL rpject:
18230 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18231 Young and Tim J. Hudson created while they were working for C2Net until
18234 *The OpenSSL Project*
18236 ### Changes between 0.9.0b and 0.9.1b [not released]
18238 * Updated a few CA certificates under certs/
18242 * Changed some BIGNUM api stuff.
18246 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18247 DGUX x86, Linux Alpha, etc.
18251 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18252 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18257 * Add -strparse option to asn1pars program which parses nested
18260 *Dr Stephen Henson <shenson@bigfoot.com>*
18262 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18266 * DSA fix for "ca" program.
18270 * Added "-genkey" option to "dsaparam" program.
18274 * Added RIPE MD160 (rmd160) message digest.
18278 * Added -a (all) option to "ssleay version" command.
18282 * Added PLATFORM define which is the id given to Configure.
18286 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18290 * Extended the ASN.1 parser routines.
18294 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18298 * Added a BN_CTX to the BN library.
18302 * Fixed the weak key values in DES library
18306 * Changed API in EVP library for cipher aliases.
18310 * Added support for RC2/64bit cipher.
18314 * Converted the lhash library to the crypto/mem.c functions.
18318 * Added more recognized ASN.1 object ids.
18322 * Added more RSA padding checks for SSL/TLS.
18326 * Added BIO proxy/filter functionality.
18330 * Added extra_certs to SSL_CTX which can be used
18331 send extra CA certificates to the client in the CA cert chain sending
18332 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18336 * Now Fortezza is denied in the authentication phase because
18337 this is key exchange mechanism is not supported by SSLeay at all.
18341 * Additional PKCS1 checks.
18345 * Support the string "TLSv1" for all TLS v1 ciphers.
18349 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18350 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18354 * Fixed a few memory leaks.
18358 * Fixed various code and comment typos.
18362 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18363 bytes sent in the client random.
18365 *Edward Bishop <ebishop@spyglass.com>*
18369 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18370 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18371 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18372 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18373 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18374 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18375 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18376 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18377 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18378 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18379 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18380 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18381 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18382 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18383 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18384 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18385 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18386 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18387 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18388 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18389 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18390 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18391 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18392 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18393 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18394 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18395 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18396 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18397 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18398 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18399 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18400 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18401 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18402 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18403 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18404 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18405 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18406 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18407 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18408 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18409 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18410 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18411 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18412 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18413 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18414 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18415 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18416 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18417 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18418 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18419 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18420 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18421 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18422 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18423 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18424 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18425 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18426 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18427 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18428 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18429 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18430 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18431 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18432 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18433 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18434 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18435 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18436 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18437 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18438 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18439 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18440 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18441 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18442 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18443 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18444 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18445 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18446 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18447 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18448 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18449 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18450 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18451 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18452 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18453 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18454 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18455 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18456 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18457 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18458 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18459 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18460 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18461 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18462 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18463 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18464 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18465 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18466 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18467 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18468 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18469 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18470 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18471 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18472 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18473 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18474 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18475 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18476 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18477 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18478 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18479 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18480 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18481 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18482 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18483 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18484 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18485 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18486 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18487 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18488 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18489 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18490 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18491 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18492 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18493 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18494 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18495 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18496 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18497 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18498 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18499 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18500 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18501 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18502 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18503 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18504 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18505 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18506 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18507 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18508 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18509 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18510 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18511 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18512 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18513 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18514 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18515 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18516 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18517 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18518 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18519 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18520 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18521 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18522 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18523 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18524 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18525 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18526 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18527 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18528 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655