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 alpha 6 [6 Aug 2020]
26 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
27 as well as actual hostnames.
31 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
32 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
33 conversely, silently ignore DTLS protocol version bounds when configuring
34 TLS-based contexts. The commands can be repeated to set bounds of both
35 types. The same applies with the corresponding "min_protocol" and
36 "max_protocol" command-line switches, in case some application uses both TLS
39 SSL_CTX instances that are created for a fixed protocol version (e.g.
40 TLSv1_server_method()) also silently ignore version bounds. Previously
41 attempts to apply bounds to these protocol versions would result in an
42 error. Now only the "version-flexible" SSL_CTX instances are subject to
43 limits in configuration files in command-line options.
47 * Deprecated the `ENGINE` API. Engines should be replaced with providers
52 * Reworked the recorded ERR codes to make better space for system errors.
53 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
54 given code is a system error (true) or an OpenSSL error (false).
58 * Reworked the test perl framework to better allow parallel testing.
60 *Nicola Tuveri and David von Oheimb*
62 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
63 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
67 * 'Configure' has been changed to figure out the configuration target if
68 none is given on the command line. Consequently, the 'config' script is
69 now only a mere wrapper. All documentation is changed to only mention
72 *Rich Salz and Richard Levitte*
74 * Added a library context that applications as well as other
75 libraries can use to form a separate context within which libcrypto
76 operations are performed.
78 There are two ways this can be used:
80 - Directly, by passing a library context to functions that take
81 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
83 - Indirectly, by creating a new library context and then assigning
84 it as the new default, with `OPENSSL_CTX_set0_default`.
86 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
87 apart from the functions directly related to `OPENSSL_CTX`, accept
88 NULL to indicate that the default library context should be used.
90 Library code that changes the default library context using
91 `OPENSSL_CTX_set0_default` should take care to restore it with a
92 second call before returning to the caller.
96 * Handshake now fails if Extended Master Secret extension is dropped
101 * Dropped interactive mode from the 'openssl' program. From now on,
102 the `openssl` command without arguments is equivalent to `openssl
107 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
108 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
109 While the old function names have been retained for backward compatibility
110 they should not be used in new developments
111 because their return values are confusing: Unlike other `_cmp()` functions
112 they do not return 0 in case their arguments are equal.
116 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
117 EC_GROUP_get_field_type().
121 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
122 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
123 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
124 Applications should rely on the library automatically assigning a suitable
125 EC_METHOD internally upon EC_GROUP construction.
129 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
130 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
131 assigned internally without application intervention.
132 Users of EC_GROUP_new() should switch to a different suitable constructor.
136 * Add CAdES-BES signature verification support, mostly derived
137 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
139 *Filipe Raimundo da Silva*
141 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
145 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
146 functions are not widely used and now OpenSSL automatically perform this
147 conversion when needed.
151 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
152 EC_KEY_precompute_mult(). These functions are not widely used and
153 applications should instead switch to named curves which OpenSSL has
154 hardcoded lookup tables for.
158 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
159 should instead use the L<EC_POINT_mul(3)> function.
163 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
164 that are not applicable to the new provider model. Applications should
165 instead use EVP_default_properties_is_fips_enabled() and
166 EVP_default_properties_enable_fips().
170 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
171 is set, an unexpected EOF is ignored, it pretends a close notify was received
172 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
176 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
177 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
178 used and applications should instead use the
179 L<EC_POINT_set_affine_coordinates(3)> and
180 L<EC_POINT_get_affine_coordinates(3)> functions.
184 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
185 arrays to be more easily constructed via a series of utility functions.
186 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
187 the various push functions and finally convert to a passable OSSL_PARAM
188 array using OSSL_PARAM_BLD_to_param().
192 * The security strength of SHA1 and MD5 based signatures in TLS has been
193 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
194 working at the default security level of 1 and instead requires security
195 level 0. The security level can be changed either using the cipher string
196 with @SECLEVEL, or calling SSL_CTX_set_security_level().
200 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
201 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
202 internal keys, if they correspond to one of those built in types.
206 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
207 contain a provider side internal key.
211 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
212 They are old functions that we don't use, and that you could disable with
213 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
217 * Project text documents not yet having a proper file name extension
218 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
219 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
223 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
224 have been converted to Markdown with the goal to produce documents
225 which not only look pretty when viewed online in the browser, but
226 remain well readable inside a plain text editor.
228 To achieve this goal, a 'minimalistic' Markdown style has been applied
229 which avoids formatting elements that interfere too much with the
230 reading flow in the text file. For example, it
232 * avoids [ATX headings][] and uses [setext headings][] instead
233 (which works for `<h1>` and `<h2>` headings only).
234 * avoids [inline links][] and uses [reference links][] instead.
235 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
237 [ATX headings]: https://github.github.com/gfm/#atx-headings
238 [setext headings]: https://github.github.com/gfm/#setext-headings
239 [inline links]: https://github.github.com/gfm/#inline-link
240 [reference links]: https://github.github.com/gfm/#reference-link
241 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
242 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
244 *Matthias St. Pierre*
246 * The test suite is changed to preserve results of each test recipe.
247 A new directory test-runs/ with subdirectories named like the
248 test recipes are created in the build tree for this purpose.
252 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
253 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
254 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
256 *David von Oheimb, Martin Peylo*
258 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
259 The legacy OCSP-focused and only partly documented API is retained.
260 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
264 * Added `util/check-format.pl`, a tool for checking adherence to the
265 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
266 The checks performed are incomplete and yield some false positives.
267 Still the tool should be useful for detecting most typical glitches.
271 * BIO_do_connect and BIO_do_handshake have been extended:
272 If domain name resolution yields multiple IP addresses all of them are tried
273 after connect() failures.
277 * All of the low level RSA functions have been deprecated including:
279 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
280 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
281 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
282 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
283 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
284 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
285 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
286 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
287 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
288 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
289 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
290 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
291 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
292 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
293 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
294 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
295 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
296 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
297 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
298 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
299 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
300 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
301 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
302 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
303 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
304 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
305 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
306 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
308 Use of these low level functions has been informally discouraged for a long
309 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
310 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
311 L<EVP_PKEY_decrypt(3)>.
315 * X509 certificates signed using SHA1 are no longer allowed at security
317 In TLS/SSL the default security level is 1. It can be set either
318 using the cipher string with @SECLEVEL, or calling
319 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
320 a call to SSL_CTX_use_certificate() will fail if the security level is not
322 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
323 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
328 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
329 modified to use PKEY APIs. These commands are now in maintenance mode
330 and no new features will be added to them.
334 * The command line utility rsautl has been deprecated.
335 Instead use the pkeyutl program.
339 * The command line utilities genrsa and rsa have been modified to use PKEY
340 APIs These commands are now in maintenance mode and no new features will
345 * All of the low level DH functions have been deprecated including:
347 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
348 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
349 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
350 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
351 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
352 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
353 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
354 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
355 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
356 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
357 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
358 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
359 DH_meth_get_generate_params and DH_meth_set_generate_params.
361 Use of these low level functions has been informally discouraged for a long
362 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
363 and L<EVP_PKEY_derive(3)>.
367 * All of the low level DSA functions have been deprecated including:
369 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
370 DSA_get_default_method, DSA_set_method, DSA_get_method,
371 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
372 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
373 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
374 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
375 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
376 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
377 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
378 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
379 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
380 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
381 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
384 Use of these low level functions has been informally discouraged for a long
385 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
386 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
390 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
391 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
392 This means that applications don't have to look at the curve NID and
393 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
394 However, they still can, that EVP_PKEY_set_alias_type() call acts as
395 a no-op when the EVP_PKEY is already of the given type.
397 Parameter and key generation is also reworked to make it possible
398 to generate EVP_PKEY_SM2 parameters and keys without having to go
399 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
400 However, code that does the latter will still work as before.
404 * Deprecated low level ECDH and ECDSA functions. These include:
406 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
407 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
410 Use of these low level functions has been informally discouraged for a long
411 time. Instead applications should use the EVP_PKEY_derive(3),
412 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
416 * Deprecated the EC_KEY_METHOD functions. These include:
418 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
419 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
420 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
421 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
422 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
423 EC_KEY_METHOD_get_verify.
425 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
429 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
430 and EVP_PKEY_decrypt() instead.
431 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
432 and EVP_PKEY_encrypt() instead.
436 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
437 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
438 a new formulation to include all the things it can be used for,
439 as well as words of caution.
443 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
444 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
448 * All of the low level HMAC functions have been deprecated including:
450 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
451 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
454 Use of these low level functions has been informally discouraged for a long
455 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
456 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
457 and L<EVP_MAC_final(3)>.
461 * Over two thousand fixes were made to the documentation, including:
462 - Common options (such as -rand/-writerand, TLS version control, etc)
463 were refactored and point to newly-enhanced descriptions in openssl.pod.
464 - Added style conformance for all options (with help from Richard Levitte),
465 documented all reported missing options, added a CI build to check
466 that all options are documented and that no unimplemented options
468 - Documented some internals, such as all use of environment variables.
469 - Addressed all internal broken L<> references.
473 * All of the low level CMAC functions have been deprecated including:
475 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
476 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
478 Use of these low level functions has been informally discouraged for a long
479 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
480 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
481 and L<EVP_MAC_final(3)>.
485 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
486 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
489 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
490 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
491 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
492 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
493 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
494 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
495 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
496 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
497 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
498 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
500 Use of these low level functions has been informally discouraged
501 for a long time. Applications should use the EVP_DigestInit_ex(3),
502 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
506 * Corrected the documentation of the return values from the `EVP_DigestSign*`
507 set of functions. The documentation mentioned negative values for some
508 errors, but this was never the case, so the mention of negative values
511 Code that followed the documentation and thereby check with something
512 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
516 * All of the low level cipher functions have been deprecated including:
518 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
519 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
520 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
521 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
522 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
523 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
524 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
525 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
526 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
527 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
528 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
529 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
530 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
531 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
532 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
533 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
534 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
535 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
536 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
537 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
538 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
539 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
540 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
541 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
542 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
543 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
544 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
545 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
546 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
548 Use of these low level functions has been informally discouraged for
549 a long time. Applications should use the high level EVP APIs, e.g.
550 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
551 equivalently named decrypt functions instead.
553 *Matt Caswell and Paul Dale*
555 * Removed include/openssl/opensslconf.h.in and replaced it with
556 include/openssl/configuration.h.in, which differs in not including
557 <openssl/macros.h>. A short header include/openssl/opensslconf.h
558 was added to include both.
560 This allows internal hacks where one might need to modify the set
561 of configured macros, for example this if deprecated symbols are
562 still supposed to be available internally:
564 #include <openssl/configuration.h>
566 #undef OPENSSL_NO_DEPRECATED
567 #define OPENSSL_SUPPRESS_DEPRECATED
569 #include <openssl/macros.h>
571 This should not be used by applications that use the exported
572 symbols, as that will lead to linking errors.
576 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
577 used in exponentiation with 512-bit moduli. No EC algorithms are
578 affected. Analysis suggests that attacks against 2-prime RSA1024,
579 3-prime RSA1536, and DSA1024 as a result of this defect would be very
580 difficult to perform and are not believed likely. Attacks against DH512
581 are considered just feasible. However, for an attack the target would
582 have to re-use the DH512 private key, which is not recommended anyway.
583 Also applications directly using the low level API BN_mod_exp may be
584 affected if they use BN_FLG_CONSTTIME.
589 * Most memory-debug features have been deprecated, and the functionality
590 replaced with no-ops.
594 * Added documentation for the STACK API. OpenSSL only defines the STACK
595 functions where they are used.
599 * Introduced a new method type and API, OSSL_SERIALIZER, to
600 represent generic serializers. An implementation is expected to
601 be able to serialize an object associated with a given name (such
602 as an algorithm name for an asymmetric key) into forms given by
603 implementation properties.
605 Serializers are primarily used from inside libcrypto, through
606 calls to functions like EVP_PKEY_print_private(),
607 PEM_write_bio_PrivateKey() and similar.
609 Serializers are specified in such a way that they can be made to
610 directly handle the provider side portion of an object, if this
611 provider side part comes from the same provider as the serializer
612 itself, but can also be made to handle objects in parametrized
613 form (as an OSSL_PARAM array of data). This allows a provider to
614 offer generic serializers as a service for any other provider.
618 * Added a .pragma directive to the syntax of configuration files, to
619 allow varying behavior in a supported and predictable manner.
620 Currently added pragma:
624 This allows dollar signs to be a keyword character unless it's
625 followed by a opening brace or parenthesis. This is useful for
626 platforms where dollar signs are commonly used in names, such as
627 volume names and system directory names on VMS.
631 * Added functionality to create an EVP_PKEY from user data. This
632 is effectively the same as creating a RSA, DH or DSA object and
633 then assigning them to an EVP_PKEY, but directly using algorithm
634 agnostic EVP functions. A benefit is that this should be future
635 proof for public key algorithms to come.
639 * Change the interpretation of the '--api' configuration option to
640 mean that this is a desired API compatibility level with no
641 further meaning. The previous interpretation, that this would
642 also mean to remove all deprecated symbols up to and including
643 the given version, no requires that 'no-deprecated' is also used
644 in the configuration.
646 When building applications, the desired API compatibility level
647 can be set with the OPENSSL_API_COMPAT macro like before. For
648 API compatibility version below 3.0, the old style numerical
649 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
650 For version 3.0 and on, the value is expected to be the decimal
651 value calculated from the major and minor version like this:
653 MAJOR * 10000 + MINOR * 100
657 -DOPENSSL_API_COMPAT=30000 For 3.0
658 -DOPENSSL_API_COMPAT=30200 For 3.2
660 To hide declarations that are deprecated up to and including the
661 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
662 given when building the application as well.
666 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
667 access to certificate and CRL stores via URIs and OSSL_STORE
670 This adds the following functions:
672 - X509_LOOKUP_store()
673 - X509_STORE_load_file()
674 - X509_STORE_load_path()
675 - X509_STORE_load_store()
676 - SSL_add_store_cert_subjects_to_stack()
677 - SSL_CTX_set_default_verify_store()
678 - SSL_CTX_load_verify_file()
679 - SSL_CTX_load_verify_dir()
680 - SSL_CTX_load_verify_store()
684 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
685 The presence of this system service is determined at run-time.
689 * Added functionality to create an EVP_PKEY context based on data
690 for methods from providers. This takes an algorithm name and a
691 property query string and simply stores them, with the intent
692 that any operation that uses this context will use those strings
693 to fetch the needed methods implicitly, thereby making the port
694 of application written for pre-3.0 OpenSSL easier.
698 * The undocumented function NCONF_WIN32() has been deprecated; for
699 conversion details see the HISTORY section of doc/man5/config.pod
703 * Introduced the new functions EVP_DigestSignInit_ex() and
704 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
705 EVP_DigestVerifyUpdate() have been converted to functions. See the man
706 pages for further details.
710 * Over two thousand fixes were made to the documentation, including:
711 adding missing command flags, better style conformance, documentation
714 *Rich Salz, Richard Levitte*
716 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
717 X25519, X448, Ed25519 and Ed448.
721 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
726 * Deprecated the public definition of ERR_STATE as well as the function
727 ERR_get_state(). This is done in preparation of making ERR_STATE an
732 * Added ERR functionality to give callers access to the stored function
733 names that have replaced the older function code based functions.
735 New functions are ERR_get_error_func(), ERR_peek_error_func(),
736 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
737 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
738 and ERR_peek_last_error_all().
740 These functions have become deprecated: ERR_get_error_line_data(),
741 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
742 ERR_func_error_string().
746 * Extended testing to be verbose for failing tests only. The make variables
747 VERBOSE_FAILURE or VF can be used to enable this:
749 $ make VF=1 test # Unix
750 $ mms /macro=(VF=1) test ! OpenVMS
751 $ nmake VF=1 test # Windows
755 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
756 used even when parsing explicit parameters, when loading a serialized key
757 or calling `EC_GROUP_new_from_ecpkparameters()`/
758 `EC_GROUP_new_from_ecparameters()`.
759 This prevents bypass of security hardening and performance gains,
760 especially for curves with specialized EC_METHODs.
761 By default, if a key encoded with explicit parameters is loaded and later
762 serialized, the output is still encoded with explicit parameters, even if
763 internally a "named" EC_GROUP is used for computation.
767 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
768 this change, EC_GROUP_set_generator would accept order and/or cofactor as
769 NULL. After this change, only the cofactor parameter can be NULL. It also
770 does some minimal sanity checks on the passed order.
775 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
776 An attack is simple, if the first CMS_recipientInfo is valid but the
777 second CMS_recipientInfo is chosen ciphertext. If the second
778 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
779 encryption key will be replaced by garbage, and the message cannot be
780 decoded, but if the RSA decryption fails, the correct encryption key is
781 used and the recipient will not notice the attack.
782 As a work around for this potential attack the length of the decrypted
783 key must be equal to the cipher default key length, in case the
784 certifiate is not given and all recipientInfo are tried out.
785 The old behaviour can be re-enabled in the CMS code by setting the
786 CMS_DEBUG_DECRYPT flag.
790 * Early start up entropy quality from the DEVRANDOM seed source has been
791 improved for older Linux systems. The RAND subsystem will wait for
792 /dev/random to be producing output before seeding from /dev/urandom.
793 The seeded state is stored for future library initialisations using
794 a system global shared memory segment. The shared memory identifier
795 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
796 the desired value. The default identifier is 114.
800 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
801 when primes for RSA keys are computed.
802 Since we previously always generated primes == 2 (mod 3) for RSA keys,
803 the 2-prime and 3-prime RSA modules were easy to distinguish, since
804 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
805 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
806 This avoids possible fingerprinting of newly generated RSA modules.
810 * Correct the extended master secret constant on EBCDIC systems. Without this
811 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
812 negotiate EMS will fail. Unfortunately this also means that TLS connections
813 between EBCDIC systems with this fix, and EBCDIC systems without this
814 fix will fail if they negotiate EMS.
818 * Changed the library initialisation so that the config file is now loaded
819 by default. This was already the case for libssl. It now occurs for both
820 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
821 OPENSSL_init_crypto() to suppress automatic loading of a config file.
825 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
826 where the former acts as a replacement for ERR_put_error(), and the
827 latter replaces the combination ERR_put_error()+ERR_add_error_data().
828 ERR_raise_data() adds more flexibility by taking a format string and
829 an arbitrary number of arguments following it, to be processed with
834 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
835 to check if a named provider is loaded and available. When called, it
836 will also activate all fallback providers if such are still present.
840 * Enforce a minimum DH modulus size of 512 bits.
844 * Changed DH parameters to generate the order q subgroup instead of 2q.
845 Previously generated DH parameters are still accepted by DH_check
846 but DH_generate_key works around that by clearing bit 0 of the
847 private key for those. This avoids leaking bit 0 of the private key.
851 * Significantly reduce secure memory usage by the randomness pools.
855 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
860 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
861 algorithms. An implementation of a key exchange algorithm can be obtained
862 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
863 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
864 the older EVP_PKEY_derive_init() function. See the man pages for the new
865 functions for further details.
869 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
873 * Removed the function names from error messages and deprecated the
876 * Removed NextStep support and the macro OPENSSL_UNISTD
880 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
881 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
882 Also removed "export var as function" capability; we do not export
883 variables, only functions.
887 * RC5_32_set_key has been changed to return an int type, with 0 indicating
888 an error and 1 indicating success. In previous versions of OpenSSL this
889 was a void type. If a key was set longer than the maximum possible this
894 * Support SM2 signing and verification schemes with X509 certificate.
898 * Use SHA256 as the default digest for TS query in the ts app.
902 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
903 This checks that the salt length is at least 128 bits, the derived key
904 length is at least 112 bits, and that the iteration count is at least 1000.
905 For backwards compatibility these checks are disabled by default in the
906 default provider, but are enabled by default in the fips provider.
907 To enable or disable these checks use the control
908 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
912 * Default cipher lists/suites are now available via a function, the
913 #defines are deprecated.
917 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
918 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
919 for Windows Store apps easier. Also, the "no-uplink" option has been added.
923 * Join the directories crypto/x509 and crypto/x509v3
927 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
928 This changes the size when using the genpkey app when no size is given. It
929 fixes an omission in earlier changes that changed all RSA, DSA and DH
930 generation apps to use 2048 bits by default.
934 * Added command 'openssl kdf' that uses the EVP_KDF API.
938 * Added command 'openssl mac' that uses the EVP_MAC API.
942 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
943 as default directories. Also added the command 'openssl info'
944 for scripting purposes.
948 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
949 deprecated. These undocumented functions were never integrated into the EVP
950 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
951 Bi-directional IGE mode. These modes were never formally standardised and
952 usage of these functions is believed to be very small. In particular
953 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
954 is ever used. The security implications are believed to be minimal, but
955 this issue was never fixed for backwards compatibility reasons. New code
956 should not use these modes.
960 * Add prediction resistance to the DRBG reseeding process.
964 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
965 mandated by IEEE Std 1619-2018.
969 * Added newline escaping functionality to a filename when using openssl dgst.
970 This output format is to replicate the output format found in the `*sum`
971 checksum programs. This aims to preserve backward compatibility.
973 *Matt Eaton, Richard Levitte, and Paul Dale*
975 * Removed the heartbeat message in DTLS feature, as it has very
976 little usage and doesn't seem to fulfill a valuable purpose.
977 The configuration option is now deprecated.
981 * Changed the output of 'openssl {digestname} < file' to display the
982 digest name in its output.
986 * Added a new generic trace API which provides support for enabling
987 instrumentation through trace output. This feature is mainly intended
988 as an aid for developers and is disabled by default. To utilize it,
989 OpenSSL needs to be configured with the `enable-trace` option.
991 If the tracing API is enabled, the application can activate trace output
992 by registering BIOs as trace channels for a number of tracing and debugging
995 The 'openssl' application has been expanded to enable any of the types
996 available via environment variables defined by the user, and serves as
997 one possible example on how to use this functionality.
999 *Richard Levitte & Matthias St. Pierre*
1001 * Added build tests for C++. These are generated files that only do one
1002 thing, to include one public OpenSSL head file each. This tests that
1003 the public header files can be usefully included in a C++ application.
1005 This test isn't enabled by default. It can be enabled with the option
1006 'enable-buildtest-c++'.
1010 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1014 * Add KMAC to EVP_MAC.
1018 * Added property based algorithm implementation selection framework to
1023 * Added SCA hardening for modular field inversion in EC_GROUP through
1024 a new dedicated field_inv() pointer in EC_METHOD.
1025 This also addresses a leakage affecting conversions from projective
1026 to affine coordinates.
1028 *Billy Bob Brumley, Nicola Tuveri*
1030 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1031 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1032 those algorithms that were already supported through the EVP_PKEY API
1033 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1034 and scrypt are now wrappers that call EVP_KDF.
1038 * Build devcrypto engine as a dynamic engine.
1040 *Eneas U de Queiroz*
1042 * Add keyed BLAKE2 to EVP_MAC.
1046 * Fix a bug in the computation of the endpoint-pair shared secret used
1047 by DTLS over SCTP. This breaks interoperability with older versions
1048 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1049 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1050 interoperability with such broken implementations. However, enabling
1051 this switch breaks interoperability with correct implementations.
1053 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1054 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1058 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1062 * Change the license to the Apache License v2.0.
1066 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1068 - Major releases (indicated by incrementing the MAJOR release number)
1069 may introduce incompatible API/ABI changes.
1070 - Minor releases (indicated by incrementing the MINOR release number)
1071 may introduce new features but retain API/ABI compatibility.
1072 - Patch releases (indicated by incrementing the PATCH number)
1073 are intended for bug fixes and other improvements of existing
1074 features only (like improving performance or adding documentation)
1075 and retain API/ABI compatibility.
1079 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1083 * Remove the 'dist' target and add a tarball building script. The
1084 'dist' target has fallen out of use, and it shouldn't be
1085 necessary to configure just to create a source distribution.
1089 * Recreate the OS390-Unix config target. It no longer relies on a
1090 special script like it did for OpenSSL pre-1.1.0.
1094 * Instead of having the source directories listed in Configure, add
1095 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1100 * Add GMAC to EVP_MAC.
1104 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1108 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1109 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1110 to facilitate the continued use of MACs through raw private keys in
1111 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1115 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1116 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1120 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1121 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1122 are retained for backwards compatibility.
1126 * AES-XTS mode now enforces that its two keys are different to mitigate
1127 the attacked described in "Efficient Instantiations of Tweakable
1128 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1129 Details of this attack can be obtained from:
1130 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1134 * Rename the object files, i.e. give them other names than in previous
1135 versions. Their names now include the name of the final product, as
1136 well as its type mnemonic (bin, lib, shlib).
1140 * Added new option for 'openssl list', '-objects', which will display the
1141 list of built in objects, i.e. OIDs with names.
1145 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1146 improves application performance by removing data copies and providing
1147 applications with zero-copy system calls such as sendfile and splice.
1151 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1152 option is set, openssl cleanses (zeroize) plaintext bytes from
1153 internal buffers after delivering them to the application. Note,
1154 the application is still responsible for cleansing other copies
1155 (e.g.: data received by SSL_read(3)).
1162 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1164 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1166 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1167 while reading in libssl then we would report an error back to the
1168 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1169 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1170 therefore give a hint as to what went wrong.
1174 * Check that ed25519 and ed448 are allowed by the security level. Previously
1175 signature algorithms not using an MD were not being checked that they were
1176 allowed by the security level.
1180 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1181 was not quite right. The behaviour was not consistent between resumption
1182 and normal handshakes, and also not quite consistent with historical
1183 behaviour. The behaviour in various scenarios has been clarified and
1184 it has been updated to make it match historical behaviour as closely as
1189 * *[VMS only]* The header files that the VMS compilers include automatically,
1190 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1191 that the C++ compiler doesn't understand. This is a shortcoming in the
1192 compiler, but can be worked around with `__cplusplus` guards.
1194 C++ applications that use OpenSSL libraries must be compiled using the
1195 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1196 functions. Otherwise, only functions with symbols of less than 31
1197 characters can be used, as the linker will not be able to successfully
1198 resolve symbols with longer names.
1202 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1203 The presence of this system service is determined at run-time.
1207 * Added newline escaping functionality to a filename when using openssl dgst.
1208 This output format is to replicate the output format found in the `*sum`
1209 checksum programs. This aims to preserve backward compatibility.
1211 *Matt Eaton, Richard Levitte, and Paul Dale*
1213 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1218 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1220 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1221 number generator (RNG). This was intended to include protection in the
1222 event of a fork() system call in order to ensure that the parent and child
1223 processes did not share the same RNG state. However this protection was not
1224 being used in the default case.
1226 A partial mitigation for this issue is that the output from a high
1227 precision timer is mixed into the RNG state so the likelihood of a parent
1228 and child process sharing state is significantly reduced.
1230 If an application already calls OPENSSL_init_crypto() explicitly using
1231 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1234 *Matthias St. Pierre*
1236 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1237 used even when parsing explicit parameters, when loading a serialized key
1238 or calling `EC_GROUP_new_from_ecpkparameters()`/
1239 `EC_GROUP_new_from_ecparameters()`.
1240 This prevents bypass of security hardening and performance gains,
1241 especially for curves with specialized EC_METHODs.
1242 By default, if a key encoded with explicit parameters is loaded and later
1243 serialized, the output is still encoded with explicit parameters, even if
1244 internally a "named" EC_GROUP is used for computation.
1248 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1249 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1250 NULL. After this change, only the cofactor parameter can be NULL. It also
1251 does some minimal sanity checks on the passed order.
1256 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1257 An attack is simple, if the first CMS_recipientInfo is valid but the
1258 second CMS_recipientInfo is chosen ciphertext. If the second
1259 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1260 encryption key will be replaced by garbage, and the message cannot be
1261 decoded, but if the RSA decryption fails, the correct encryption key is
1262 used and the recipient will not notice the attack.
1263 As a work around for this potential attack the length of the decrypted
1264 key must be equal to the cipher default key length, in case the
1265 certifiate is not given and all recipientInfo are tried out.
1266 The old behaviour can be re-enabled in the CMS code by setting the
1267 CMS_DEBUG_DECRYPT flag.
1272 * Early start up entropy quality from the DEVRANDOM seed source has been
1273 improved for older Linux systems. The RAND subsystem will wait for
1274 /dev/random to be producing output before seeding from /dev/urandom.
1275 The seeded state is stored for future library initialisations using
1276 a system global shared memory segment. The shared memory identifier
1277 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1278 the desired value. The default identifier is 114.
1282 * Correct the extended master secret constant on EBCDIC systems. Without this
1283 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1284 negotiate EMS will fail. Unfortunately this also means that TLS connections
1285 between EBCDIC systems with this fix, and EBCDIC systems without this
1286 fix will fail if they negotiate EMS.
1290 * Use Windows installation paths in the mingw builds
1292 Mingw isn't a POSIX environment per se, which means that Windows
1293 paths should be used for installation.
1298 * Changed DH_check to accept parameters with order q and 2q subgroups.
1299 With order 2q subgroups the bit 0 of the private key is not secret
1300 but DH_generate_key works around that by clearing bit 0 of the
1301 private key for those. This avoids leaking bit 0 of the private key.
1305 * Significantly reduce secure memory usage by the randomness pools.
1309 * Revert the DEVRANDOM_WAIT feature for Linux systems
1311 The DEVRANDOM_WAIT feature added a select() call to wait for the
1312 /dev/random device to become readable before reading from the
1313 /dev/urandom device.
1315 It turned out that this change had negative side effects on
1316 performance which were not acceptable. After some discussion it
1317 was decided to revert this feature and leave it up to the OS
1318 resp. the platform maintainer to ensure a proper initialization
1319 during early boot time.
1321 *Matthias St. Pierre*
1323 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1325 * Add build tests for C++. These are generated files that only do one
1326 thing, to include one public OpenSSL head file each. This tests that
1327 the public header files can be usefully included in a C++ application.
1329 This test isn't enabled by default. It can be enabled with the option
1330 'enable-buildtest-c++'.
1334 * Enable SHA3 pre-hashing for ECDSA and DSA.
1338 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1339 This changes the size when using the genpkey app when no size is given. It
1340 fixes an omission in earlier changes that changed all RSA, DSA and DH
1341 generation apps to use 2048 bits by default.
1345 * Reorganize the manual pages to consistently have RETURN VALUES,
1346 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1347 util/fix-doc-nits accordingly.
1349 *Paul Yang, Joshua Lock*
1351 * Add the missing accessor EVP_PKEY_get0_engine()
1355 * Have apps like 's_client' and 's_server' output the signature scheme
1356 along with other cipher suite parameters when debugging.
1360 * Make OPENSSL_config() error agnostic again.
1364 * Do the error handling in RSA decryption constant time.
1368 * Prevent over long nonces in ChaCha20-Poly1305.
1370 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1371 for every encryption operation. RFC 7539 specifies that the nonce value
1372 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1373 and front pads the nonce with 0 bytes if it is less than 12
1374 bytes. However it also incorrectly allows a nonce to be set of up to 16
1375 bytes. In this case only the last 12 bytes are significant and any
1376 additional leading bytes are ignored.
1378 It is a requirement of using this cipher that nonce values are
1379 unique. Messages encrypted using a reused nonce value are susceptible to
1380 serious confidentiality and integrity attacks. If an application changes
1381 the default nonce length to be longer than 12 bytes and then makes a
1382 change to the leading bytes of the nonce expecting the new value to be a
1383 new unique nonce then such an application could inadvertently encrypt
1384 messages with a reused nonce.
1386 Additionally the ignored bytes in a long nonce are not covered by the
1387 integrity guarantee of this cipher. Any application that relies on the
1388 integrity of these ignored leading bytes of a long nonce may be further
1389 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1390 is safe because no such use sets such a long nonce value. However user
1391 applications that use this cipher directly and set a non-default nonce
1392 length to be longer than 12 bytes may be vulnerable.
1394 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1400 * Add DEVRANDOM_WAIT feature for Linux systems
1402 On older Linux systems where the getrandom() system call is not available,
1403 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1404 Contrary to getrandom(), the /dev/urandom device will not block during
1405 early boot when the kernel CSPRNG has not been seeded yet.
1407 To mitigate this known weakness, use select() to wait for /dev/random to
1408 become readable before reading from /dev/urandom.
1410 * Ensure that SM2 only uses SM3 as digest algorithm
1414 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1416 * Change the info callback signals for the start and end of a post-handshake
1417 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1418 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1419 confused by this and assume that a TLSv1.2 renegotiation has started. This
1420 can break KeyUpdate handling. Instead we no longer signal the start and end
1421 of a post handshake message exchange (although the messages themselves are
1422 still signalled). This could break some applications that were expecting
1423 the old signals. However without this KeyUpdate is not usable for many
1428 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1430 * Timing vulnerability in DSA signature generation
1432 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1433 timing side channel attack. An attacker could use variations in the signing
1434 algorithm to recover the private key.
1436 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1441 * Timing vulnerability in ECDSA signature generation
1443 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1444 timing side channel attack. An attacker could use variations in the signing
1445 algorithm to recover the private key.
1447 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1452 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1453 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1454 of two gigabytes and the error handling improved.
1456 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1457 categorized as a normal bug, not a security issue, because the DRBG reseeds
1458 automatically and is fully functional even without additional randomness
1459 provided by the application.
1461 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1463 * Add a new ClientHello callback. Provides a callback interface that gives
1464 the application the ability to adjust the nascent SSL object at the
1465 earliest stage of ClientHello processing, immediately after extensions have
1466 been collected but before they have been processed. In particular, this
1467 callback can adjust the supported TLS versions in response to the contents
1472 * Add SM2 base algorithm support.
1476 * s390x assembly pack: add (improved) hardware-support for the following
1477 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1478 aes-cfb/cfb8, aes-ecb.
1482 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1483 parameter is no longer accepted, as it leads to a corrupt table. NULL
1484 pem_str is reserved for alias entries only.
1488 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1489 step for prime curves. The new implementation is based on formulae from
1490 differential addition-and-doubling in homogeneous projective coordinates
1491 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1492 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1493 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1494 to work in projective coordinates.
1496 *Billy Bob Brumley, Nicola Tuveri*
1498 * Change generating and checking of primes so that the error rate of not
1499 being prime depends on the intended use based on the size of the input.
1500 For larger primes this will result in more rounds of Miller-Rabin.
1501 The maximal error rate for primes with more than 1080 bits is lowered
1504 *Kurt Roeckx, Annie Yousar*
1506 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1510 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1511 moving between systems, and to avoid confusion when a Windows build is
1512 done with mingw vs with MSVC. For POSIX installs, there's still a
1513 symlink or copy named 'tsget' to avoid that confusion as well.
1517 * Revert blinding in ECDSA sign and instead make problematic addition
1518 length-invariant. Switch even to fixed-length Montgomery multiplication.
1522 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1523 step for binary curves. The new implementation is based on formulae from
1524 differential addition-and-doubling in mixed Lopez-Dahab projective
1525 coordinates, modified to independently blind the operands.
1527 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1529 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1530 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1531 EC_METHODs to implement their own specialized "ladder step", to take
1532 advantage of more favorable coordinate systems or more efficient
1533 differential addition-and-doubling algorithms.
1535 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1537 * Modified the random device based seed sources to keep the relevant
1538 file descriptors open rather than reopening them on each access.
1539 This allows such sources to operate in a chroot() jail without
1540 the associated device nodes being available. This behaviour can be
1541 controlled using RAND_keep_random_devices_open().
1545 * Numerous side-channel attack mitigations have been applied. This may have
1546 performance impacts for some algorithms for the benefit of improved
1547 security. Specific changes are noted in this change log by their respective
1552 * AIX shared library support overhaul. Switch to AIX "natural" way of
1553 handling shared libraries, which means collecting shared objects of
1554 different versions and bitnesses in one common archive. This allows to
1555 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1556 doesn't affect the way 3rd party applications are linked, only how
1557 multi-version installation is managed.
1561 * Make ec_group_do_inverse_ord() more robust and available to other
1562 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1563 mitigations are applied to the fallback BN_mod_inverse().
1564 When using this function rather than BN_mod_inverse() directly, new
1565 EC cryptosystem implementations are then safer-by-default.
1569 * Add coordinate blinding for EC_POINT and implement projective
1570 coordinate blinding for generic prime curves as a countermeasure to
1571 chosen point SCA attacks.
1573 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1575 * Add blinding to ECDSA and DSA signatures to protect against side channel
1576 attacks discovered by Keegan Ryan (NCC Group).
1580 * Enforce checking in the pkeyutl command line app to ensure that the input
1581 length does not exceed the maximum supported digest length when performing
1582 a sign, verify or verifyrecover operation.
1586 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1587 I/O in combination with something like select() or poll() will hang. This
1588 can be turned off again using SSL_CTX_clear_mode().
1589 Many applications do not properly handle non-application data records, and
1590 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1591 around the problems in those applications, but can also break some.
1592 It's recommended to read the manpages about SSL_read(), SSL_write(),
1593 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1594 SSL_CTX_set_read_ahead() again.
1598 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1599 now allow empty (zero character) pass phrases.
1603 * Apply blinding to binary field modular inversion and remove patent
1604 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1608 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1609 binary and prime elliptic curves.
1613 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1614 constant time fixed point multiplication.
1618 * Revise elliptic curve scalar multiplication with timing attack
1619 defenses: ec_wNAF_mul redirects to a constant time implementation
1620 when computing fixed point and variable point multiplication (which
1621 in OpenSSL are mostly used with secret scalars in keygen, sign,
1622 ECDH derive operations).
1623 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1626 * Updated CONTRIBUTING
1630 * Updated DRBG / RAND to request nonce and additional low entropy
1631 randomness from the system.
1633 *Matthias St. Pierre*
1635 * Updated 'openssl rehash' to use OpenSSL consistent default.
1639 * Moved the load of the ssl_conf module to libcrypto, which helps
1640 loading engines that libssl uses before libssl is initialised.
1644 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1648 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1650 *Ingo Schwarze, Rich Salz*
1652 * Added output of accepting IP address and port for 'openssl s_server'
1656 * Added a new API for TLSv1.3 ciphersuites:
1657 SSL_CTX_set_ciphersuites()
1658 SSL_set_ciphersuites()
1662 * Memory allocation failures consistently add an error to the error
1667 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1668 in libcrypto when run as setuid/setgid.
1672 * Load any config file by default when libssl is used.
1676 * Added new public header file <openssl/rand_drbg.h> and documentation
1677 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1679 *Matthias St. Pierre*
1681 * QNX support removed (cannot find contributors to get their approval
1682 for the license change).
1686 * TLSv1.3 replay protection for early data has been implemented. See the
1687 SSL_read_early_data() man page for further details.
1691 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1692 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1693 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1694 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1695 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1696 configuration has been separated out. See the ciphers man page or the
1697 SSL_CTX_set_ciphersuites() man page for more information.
1701 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1702 in responder mode now supports the new "-multi" option, which
1703 spawns the specified number of child processes to handle OCSP
1704 requests. The "-timeout" option now also limits the OCSP
1705 responder's patience to wait to receive the full client request
1706 on a newly accepted connection. Child processes are respawned
1707 as needed, and the CA index file is automatically reloaded
1708 when changed. This makes it possible to run the "ocsp" responder
1709 as a long-running service, making the OpenSSL CA somewhat more
1710 feature-complete. In this mode, most diagnostic messages logged
1711 after entering the event loop are logged via syslog(3) rather than
1716 * Added support for X448 and Ed448. Heavily based on original work by
1721 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1722 objects loaded. This adds the functions OSSL_STORE_expect() and
1723 OSSL_STORE_find() as well as needed tools to construct searches and
1724 get the search data out of them.
1728 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1729 version of OpenSSL should review their configuration settings to ensure
1730 that they are still appropriate for TLSv1.3. For further information see:
1731 <https://wiki.openssl.org/index.php/TLS1.3>
1735 * Grand redesign of the OpenSSL random generator
1737 The default RAND method now utilizes an AES-CTR DRBG according to
1738 NIST standard SP 800-90Ar1. The new random generator is essentially
1739 a port of the default random generator from the OpenSSL FIPS 2.0
1740 object module. It is a hybrid deterministic random bit generator
1741 using an AES-CTR bit stream and which seeds and reseeds itself
1742 automatically using trusted system entropy sources.
1744 Some of its new features are:
1745 - Support for multiple DRBG instances with seed chaining.
1746 - The default RAND method makes use of a DRBG.
1747 - There is a public and private DRBG instance.
1748 - The DRBG instances are fork-safe.
1749 - Keep all global DRBG instances on the secure heap if it is enabled.
1750 - The public and private DRBG instance are per thread for lock free
1753 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1755 * Changed Configure so it only says what it does and doesn't dump
1756 so much data. Instead, ./configdata.pm should be used as a script
1757 to display all sorts of configuration data.
1761 * Added processing of "make variables" to Configure.
1765 * Added SHA512/224 and SHA512/256 algorithm support.
1769 * The last traces of Netware support, first removed in 1.1.0, have
1774 * Get rid of Makefile.shared, and in the process, make the processing
1775 of certain files (rc.obj, or the .def/.map/.opt files produced from
1776 the ordinal files) more visible and hopefully easier to trace and
1777 debug (or make silent).
1781 * Make it possible to have environment variable assignments as
1782 arguments to config / Configure.
1786 * Add multi-prime RSA (RFC 8017) support.
1790 * Add SM3 implemented according to GB/T 32905-2016
1791 *Jack Lloyd <jack.lloyd@ribose.com>,*
1792 *Ronald Tse <ronald.tse@ribose.com>,*
1793 *Erick Borsboom <erick.borsboom@ribose.com>*
1795 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1796 as documented in RFC6066.
1797 Based on a patch from Tomasz Moń
1799 *Filipe Raimundo da Silva*
1801 * Add SM4 implemented according to GB/T 32907-2016.
1802 *Jack Lloyd <jack.lloyd@ribose.com>,*
1803 *Ronald Tse <ronald.tse@ribose.com>,*
1804 *Erick Borsboom <erick.borsboom@ribose.com>*
1806 * Reimplement -newreq-nodes and ERR_error_string_n; the
1807 original author does not agree with the license change.
1811 * Add ARIA AEAD TLS support.
1815 * Some macro definitions to support VS6 have been removed. Visual
1816 Studio 6 has not worked since 1.1.0
1820 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1821 without clearing the errors.
1825 * Add "atfork" functions. If building on a system that without
1826 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1827 requirements. The RAND facility now uses/requires this.
1835 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1836 not possible to disable entirely. However, it's still possible to
1837 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1840 To disable, configure with 'no-ui-console'. 'no-ui' is still
1841 possible to use as an alias. Check at compile time with the
1842 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1843 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1847 * Add a STORE module, which implements a uniform and URI based reader of
1848 stores that can contain keys, certificates, CRLs and numerous other
1849 objects. The main API is loosely based on a few stdio functions,
1850 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1851 OSSL_STORE_error and OSSL_STORE_close.
1852 The implementation uses backends called "loaders" to implement arbitrary
1853 URI schemes. There is one built in "loader" for the 'file' scheme.
1857 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1858 then adjusted to work on FreeBSD 8.4 as well.
1859 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1860 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1864 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1865 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1866 error code calls like this:
1868 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1870 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1871 that can be encoded in C. For the foreseeable future, this will only
1874 *Richard Levitte and Tim Hudson*
1876 * Removed BSD cryptodev engine.
1880 * Add a build target 'build_all_generated', to build all generated files
1881 and only that. This can be used to prepare everything that requires
1882 things like perl for a system that lacks perl and then move everything
1883 to that system and do the rest of the build there.
1887 * In the UI interface, make it possible to duplicate the user data. This
1888 can be used by engines that need to retain the data for a longer time
1889 than just the call where this user data is passed.
1893 * Ignore the '-named_curve auto' value for compatibility of applications
1896 *Tomas Mraz <tmraz@fedoraproject.org>*
1898 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1899 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1900 alerts across multiple records (some of which could be empty). In practice
1901 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1902 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1903 support this at all. Supporting it adds significant complexity to the
1904 record layer, and its removal is unlikely to cause interoperability
1909 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1910 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1911 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1916 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1917 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1919 *Richard Levitte, Andy Polyakov*
1921 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1926 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1927 platform rather than 'mingw'.
1931 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1932 success if they are asked to add an object which already exists
1933 in the store. This change cascades to other functions which load
1934 certificates and CRLs.
1938 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1939 facilitate stack unwinding even from assembly subroutines.
1943 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1944 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1948 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1949 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1950 which is the minimum version we support.
1954 * Certificate time validation (X509_cmp_time) enforces stricter
1955 compliance with RFC 5280. Fractional seconds and timezone offsets
1956 are no longer allowed.
1960 * Add support for ARIA
1964 * s_client will now send the Server Name Indication (SNI) extension by
1965 default unless the new "-noservername" option is used. The server name is
1966 based on the host provided to the "-connect" option unless overridden by
1967 using "-servername".
1971 * Add support for SipHash
1975 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1976 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1977 prevent issues where no progress is being made and the peer continually
1978 sends unrecognised record types, using up resources processing them.
1982 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1983 using the algorithm defined in
1984 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1988 * Heartbeat support has been removed; the ABI is changed for now.
1990 *Richard Levitte, Rich Salz*
1992 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1996 * The RSA "null" method, which was partially supported to avoid patent
1997 issues, has been replaced to always returns NULL.
2004 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2006 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2007 used even when parsing explicit parameters, when loading a serialized key
2008 or calling `EC_GROUP_new_from_ecpkparameters()`/
2009 `EC_GROUP_new_from_ecparameters()`.
2010 This prevents bypass of security hardening and performance gains,
2011 especially for curves with specialized EC_METHODs.
2012 By default, if a key encoded with explicit parameters is loaded and later
2013 serialized, the output is still encoded with explicit parameters, even if
2014 internally a "named" EC_GROUP is used for computation.
2018 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2019 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2020 NULL. After this change, only the cofactor parameter can be NULL. It also
2021 does some minimal sanity checks on the passed order.
2026 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2027 An attack is simple, if the first CMS_recipientInfo is valid but the
2028 second CMS_recipientInfo is chosen ciphertext. If the second
2029 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2030 encryption key will be replaced by garbage, and the message cannot be
2031 decoded, but if the RSA decryption fails, the correct encryption key is
2032 used and the recipient will not notice the attack.
2033 As a work around for this potential attack the length of the decrypted
2034 key must be equal to the cipher default key length, in case the
2035 certifiate is not given and all recipientInfo are tried out.
2036 The old behaviour can be re-enabled in the CMS code by setting the
2037 CMS_DEBUG_DECRYPT flag.
2042 * Use Windows installation paths in the mingw builds
2044 Mingw isn't a POSIX environment per se, which means that Windows
2045 paths should be used for installation.
2050 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2052 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2053 This changes the size when using the genpkey app when no size is given. It
2054 fixes an omission in earlier changes that changed all RSA, DSA and DH
2055 generation apps to use 2048 bits by default.
2059 * Prevent over long nonces in ChaCha20-Poly1305.
2061 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2062 for every encryption operation. RFC 7539 specifies that the nonce value
2063 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2064 and front pads the nonce with 0 bytes if it is less than 12
2065 bytes. However it also incorrectly allows a nonce to be set of up to 16
2066 bytes. In this case only the last 12 bytes are significant and any
2067 additional leading bytes are ignored.
2069 It is a requirement of using this cipher that nonce values are
2070 unique. Messages encrypted using a reused nonce value are susceptible to
2071 serious confidentiality and integrity attacks. If an application changes
2072 the default nonce length to be longer than 12 bytes and then makes a
2073 change to the leading bytes of the nonce expecting the new value to be a
2074 new unique nonce then such an application could inadvertently encrypt
2075 messages with a reused nonce.
2077 Additionally the ignored bytes in a long nonce are not covered by the
2078 integrity guarantee of this cipher. Any application that relies on the
2079 integrity of these ignored leading bytes of a long nonce may be further
2080 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2081 is safe because no such use sets such a long nonce value. However user
2082 applications that use this cipher directly and set a non-default nonce
2083 length to be longer than 12 bytes may be vulnerable.
2085 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2091 * Added SCA hardening for modular field inversion in EC_GROUP through
2092 a new dedicated field_inv() pointer in EC_METHOD.
2093 This also addresses a leakage affecting conversions from projective
2094 to affine coordinates.
2096 *Billy Bob Brumley, Nicola Tuveri*
2098 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2099 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2103 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2107 * Remove the 'dist' target and add a tarball building script. The
2108 'dist' target has fallen out of use, and it shouldn't be
2109 necessary to configure just to create a source distribution.
2113 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2115 * Timing vulnerability in DSA signature generation
2117 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2118 timing side channel attack. An attacker could use variations in the signing
2119 algorithm to recover the private key.
2121 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2126 * Timing vulnerability in ECDSA signature generation
2128 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2129 timing side channel attack. An attacker could use variations in the signing
2130 algorithm to recover the private key.
2132 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2137 * Add coordinate blinding for EC_POINT and implement projective
2138 coordinate blinding for generic prime curves as a countermeasure to
2139 chosen point SCA attacks.
2141 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2143 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2145 * Client DoS due to large DH parameter
2147 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2148 malicious server can send a very large prime value to the client. This will
2149 cause the client to spend an unreasonably long period of time generating a
2150 key for this prime resulting in a hang until the client has finished. This
2151 could be exploited in a Denial Of Service attack.
2153 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2158 * Cache timing vulnerability in RSA Key Generation
2160 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2161 a cache timing side channel attack. An attacker with sufficient access to
2162 mount cache timing attacks during the RSA key generation process could
2163 recover the private key.
2165 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2166 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2171 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2172 parameter is no longer accepted, as it leads to a corrupt table. NULL
2173 pem_str is reserved for alias entries only.
2177 * Revert blinding in ECDSA sign and instead make problematic addition
2178 length-invariant. Switch even to fixed-length Montgomery multiplication.
2182 * Change generating and checking of primes so that the error rate of not
2183 being prime depends on the intended use based on the size of the input.
2184 For larger primes this will result in more rounds of Miller-Rabin.
2185 The maximal error rate for primes with more than 1080 bits is lowered
2188 *Kurt Roeckx, Annie Yousar*
2190 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2194 * Add blinding to ECDSA and DSA signatures to protect against side channel
2195 attacks discovered by Keegan Ryan (NCC Group).
2199 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2200 now allow empty (zero character) pass phrases.
2204 * Certificate time validation (X509_cmp_time) enforces stricter
2205 compliance with RFC 5280. Fractional seconds and timezone offsets
2206 are no longer allowed.
2210 * Fixed a text canonicalisation bug in CMS
2212 Where a CMS detached signature is used with text content the text goes
2213 through a canonicalisation process first prior to signing or verifying a
2214 signature. This process strips trailing space at the end of lines, converts
2215 line terminators to CRLF and removes additional trailing line terminators
2216 at the end of a file. A bug in the canonicalisation process meant that
2217 some characters, such as form-feed, were incorrectly treated as whitespace
2218 and removed. This is contrary to the specification (RFC5485). This fix
2219 could mean that detached text data signed with an earlier version of
2220 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2221 signed with a fixed OpenSSL may fail to verify with an earlier version of
2222 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2223 and use the "-binary" flag (for the "cms" command line application) or set
2224 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2228 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2230 * Constructed ASN.1 types with a recursive definition could exceed the stack
2232 Constructed ASN.1 types with a recursive definition (such as can be found
2233 in PKCS7) could eventually exceed the stack given malicious input with
2234 excessive recursion. This could result in a Denial Of Service attack. There
2235 are no such structures used within SSL/TLS that come from untrusted sources
2236 so this is considered safe.
2238 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2244 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2246 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2247 effectively reduced to only comparing the least significant bit of each
2248 byte. This allows an attacker to forge messages that would be considered as
2249 authenticated in an amount of tries lower than that guaranteed by the
2250 security claims of the scheme. The module can only be compiled by the
2251 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2253 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2259 * Add a build target 'build_all_generated', to build all generated files
2260 and only that. This can be used to prepare everything that requires
2261 things like perl for a system that lacks perl and then move everything
2262 to that system and do the rest of the build there.
2266 * Backport SSL_OP_NO_RENGOTIATION
2268 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2269 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2270 changes this is no longer possible in 1.1.0. Therefore the new
2271 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2272 1.1.0 to provide equivalent functionality.
2274 Note that if an application built against 1.1.0h headers (or above) is run
2275 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2276 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2280 * Removed the OS390-Unix config target. It relied on a script that doesn't
2285 * rsaz_1024_mul_avx2 overflow bug on x86_64
2287 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2288 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2289 Analysis suggests that attacks against RSA and DSA as a result of this
2290 defect would be very difficult to perform and are not believed likely.
2291 Attacks against DH1024 are considered just feasible, because most of the
2292 work necessary to deduce information about a private key may be performed
2293 offline. The amount of resources required for such an attack would be
2294 significant. However, for an attack on TLS to be meaningful, the server
2295 would have to share the DH1024 private key among multiple clients, which is
2296 no longer an option since CVE-2016-0701.
2298 This only affects processors that support the AVX2 but not ADX extensions
2299 like Intel Haswell (4th generation).
2301 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2302 was originally found via the OSS-Fuzz project.
2307 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2309 * bn_sqrx8x_internal carry bug on x86_64
2311 There is a carry propagating bug in the x86_64 Montgomery squaring
2312 procedure. No EC algorithms are affected. Analysis suggests that attacks
2313 against RSA and DSA as a result of this defect would be very difficult to
2314 perform and are not believed likely. Attacks against DH are considered just
2315 feasible (although very difficult) because most of the work necessary to
2316 deduce information about a private key may be performed offline. The amount
2317 of resources required for such an attack would be very significant and
2318 likely only accessible to a limited number of attackers. An attacker would
2319 additionally need online access to an unpatched system using the target
2320 private key in a scenario with persistent DH parameters and a private
2321 key that is shared between multiple clients.
2323 This only affects processors that support the BMI1, BMI2 and ADX extensions
2324 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2326 This issue was reported to OpenSSL by the OSS-Fuzz project.
2331 * Malformed X.509 IPAddressFamily could cause OOB read
2333 If an X.509 certificate has a malformed IPAddressFamily extension,
2334 OpenSSL could do a one-byte buffer overread. The most likely result
2335 would be an erroneous display of the certificate in text format.
2337 This issue was reported to OpenSSL by the OSS-Fuzz project.
2342 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2344 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2345 platform rather than 'mingw'.
2349 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2350 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2351 which is the minimum version we support.
2355 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2357 * Encrypt-Then-Mac renegotiation crash
2359 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2360 negotiated where it was not in the original handshake (or vice-versa) then
2361 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2362 and servers are affected.
2364 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2369 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2371 * Truncated packet could crash via OOB read
2373 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2374 cipher is being used, then a truncated packet can cause that host to
2375 perform an out-of-bounds read, usually resulting in a crash.
2377 This issue was reported to OpenSSL by Robert Święcki of Google.
2382 * Bad (EC)DHE parameters cause a client crash
2384 If a malicious server supplies bad parameters for a DHE or ECDHE key
2385 exchange then this can result in the client attempting to dereference a
2386 NULL pointer leading to a client crash. This could be exploited in a Denial
2389 This issue was reported to OpenSSL by Guido Vranken.
2394 * BN_mod_exp may produce incorrect results on x86_64
2396 There is a carry propagating bug in the x86_64 Montgomery squaring
2397 procedure. No EC algorithms are affected. Analysis suggests that attacks
2398 against RSA and DSA as a result of this defect would be very difficult to
2399 perform and are not believed likely. Attacks against DH are considered just
2400 feasible (although very difficult) because most of the work necessary to
2401 deduce information about a private key may be performed offline. The amount
2402 of resources required for such an attack would be very significant and
2403 likely only accessible to a limited number of attackers. An attacker would
2404 additionally need online access to an unpatched system using the target
2405 private key in a scenario with persistent DH parameters and a private
2406 key that is shared between multiple clients. For example this can occur by
2407 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2408 similar to CVE-2015-3193 but must be treated as a separate problem.
2410 This issue was reported to OpenSSL by the OSS-Fuzz project.
2415 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2417 * ChaCha20/Poly1305 heap-buffer-overflow
2419 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2420 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2421 crash. This issue is not considered to be exploitable beyond a DoS.
2423 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2428 * CMS Null dereference
2430 Applications parsing invalid CMS structures can crash with a NULL pointer
2431 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2432 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2433 structure callback if an attempt is made to free certain invalid encodings.
2434 Only CHOICE structures using a callback which do not handle NULL value are
2437 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2442 * Montgomery multiplication may produce incorrect results
2444 There is a carry propagating bug in the Broadwell-specific Montgomery
2445 multiplication procedure that handles input lengths divisible by, but
2446 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2447 and DH private keys are impossible. This is because the subroutine in
2448 question is not used in operations with the private key itself and an input
2449 of the attacker's direct choice. Otherwise the bug can manifest itself as
2450 transient authentication and key negotiation failures or reproducible
2451 erroneous outcome of public-key operations with specially crafted input.
2452 Among EC algorithms only Brainpool P-512 curves are affected and one
2453 presumably can attack ECDH key negotiation. Impact was not analyzed in
2454 detail, because pre-requisites for attack are considered unlikely. Namely
2455 multiple clients have to choose the curve in question and the server has to
2456 share the private key among them, neither of which is default behaviour.
2457 Even then only clients that chose the curve will be affected.
2459 This issue was publicly reported as transient failures and was not
2460 initially recognized as a security issue. Thanks to Richard Morgan for
2461 providing reproducible case.
2466 * Removed automatic addition of RPATH in shared libraries and executables,
2467 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2471 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2473 * Fix Use After Free for large message sizes
2475 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2476 message larger than approx 16k is received then the underlying buffer to
2477 store the incoming message is reallocated and moved. Unfortunately a
2478 dangling pointer to the old location is left which results in an attempt to
2479 write to the previously freed location. This is likely to result in a
2480 crash, however it could potentially lead to execution of arbitrary code.
2482 This issue only affects OpenSSL 1.1.0a.
2484 This issue was reported to OpenSSL by Robert Święcki.
2489 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2491 * OCSP Status Request extension unbounded memory growth
2493 A malicious client can send an excessively large OCSP Status Request
2494 extension. If that client continually requests renegotiation, sending a
2495 large OCSP Status Request extension each time, then there will be unbounded
2496 memory growth on the server. This will eventually lead to a Denial Of
2497 Service attack through memory exhaustion. Servers with a default
2498 configuration are vulnerable even if they do not support OCSP. Builds using
2499 the "no-ocsp" build time option are not affected.
2501 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2506 * SSL_peek() hang on empty record
2508 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2509 sends an empty record. This could be exploited by a malicious peer in a
2510 Denial Of Service attack.
2512 This issue was reported to OpenSSL by Alex Gaynor.
2517 * Excessive allocation of memory in tls_get_message_header() and
2518 dtls1_preprocess_fragment()
2520 A (D)TLS message includes 3 bytes for its length in the header for the
2521 message. This would allow for messages up to 16Mb in length. Messages of
2522 this length are excessive and OpenSSL includes a check to ensure that a
2523 peer is sending reasonably sized messages in order to avoid too much memory
2524 being consumed to service a connection. A flaw in the logic of version
2525 1.1.0 means that memory for the message is allocated too early, prior to
2526 the excessive message length check. Due to way memory is allocated in
2527 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2528 to service a connection. This could lead to a Denial of Service through
2529 memory exhaustion. However, the excessive message length check still takes
2530 place, and this would cause the connection to immediately fail. Assuming
2531 that the application calls SSL_free() on the failed connection in a timely
2532 manner then the 21Mb of allocated memory will then be immediately freed
2533 again. Therefore the excessive memory allocation will be transitory in
2534 nature. This then means that there is only a security impact if:
2536 1) The application does not call SSL_free() in a timely manner in the event
2537 that the connection fails
2539 2) The application is working in a constrained environment where there is
2540 very little free memory
2542 3) The attacker initiates multiple connection attempts such that there are
2543 multiple connections in a state where memory has been allocated for the
2544 connection; SSL_free() has not yet been called; and there is insufficient
2545 memory to service the multiple requests.
2547 Except in the instance of (1) above any Denial Of Service is likely to be
2548 transitory because as soon as the connection fails the memory is
2549 subsequently freed again in the SSL_free() call. However there is an
2550 increased risk during this period of application crashes due to the lack of
2551 memory - which would then mean a more serious Denial of Service.
2553 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2554 (CVE-2016-6307 and CVE-2016-6308)
2558 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2559 had to be removed. Primary reason is that vendor assembler can't
2560 assemble our modules with -KPIC flag. As result it, assembly
2561 support, was not even available as option. But its lack means
2562 lack of side-channel resistant code, which is incompatible with
2563 security by todays standards. Fortunately gcc is readily available
2564 prepackaged option, which we firmly point at...
2568 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2570 * Windows command-line tool supports UTF-8 opt-in option for arguments
2571 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2572 (to any value) allows Windows user to access PKCS#12 file generated
2573 with Windows CryptoAPI and protected with non-ASCII password, as well
2574 as files generated under UTF-8 locale on Linux also protected with
2579 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2580 have been disabled by default and removed from DEFAULT, just like RC4.
2581 See the RC4 item below to re-enable both.
2585 * The method for finding the storage location for the Windows RAND seed file
2586 has changed. First we check %RANDFILE%. If that is not set then we check
2587 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2588 all else fails we fall back to C:\.
2592 * The EVP_EncryptUpdate() function has had its return type changed from void
2593 to int. A return of 0 indicates and error while a return of 1 indicates
2598 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2599 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2600 off the constant time implementation for RSA, DSA and DH have been made
2601 no-ops and deprecated.
2605 * Windows RAND implementation was simplified to only get entropy by
2606 calling CryptGenRandom(). Various other RAND-related tickets
2609 *Joseph Wylie Yandle, Rich Salz*
2611 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2612 and `OPENSSL_LH_`, respectively. The old names are available
2613 with API compatibility. They new names are now completely documented.
2617 * Unify TYPE_up_ref(obj) methods signature.
2618 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2619 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2620 int (instead of void) like all others TYPE_up_ref() methods.
2621 So now these methods also check the return value of CRYPTO_atomic_add(),
2622 and the validity of object reference counter.
2624 *fdasilvayy@gmail.com*
2626 * With Windows Visual Studio builds, the .pdb files are installed
2627 alongside the installed libraries and executables. For a static
2628 library installation, ossl_static.pdb is the associate compiler
2629 generated .pdb file to be used when linking programs.
2633 * Remove openssl.spec. Packaging files belong with the packagers.
2637 * Automatic Darwin/OSX configuration has had a refresh, it will now
2638 recognise x86_64 architectures automatically. You can still decide
2639 to build for a different bitness with the environment variable
2640 KERNEL_BITS (can be 32 or 64), for example:
2642 KERNEL_BITS=32 ./config
2646 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2647 256 bit AES and HMAC with SHA256.
2651 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2655 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2659 * To enable users to have their own config files and build file templates,
2660 Configure looks in the directory indicated by the environment variable
2661 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2662 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2663 name and is used as is.
2667 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2668 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2669 X509_CERT_FILE_CTX was removed.
2673 * "shared" builds are now the default. To create only static libraries use
2674 the "no-shared" Configure option.
2678 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2679 All of these option have not worked for some while and are fundamental
2684 * Make various cleanup routines no-ops and mark them as deprecated. Most
2685 global cleanup functions are no longer required because they are handled
2686 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2687 Explicitly de-initing can cause problems (e.g. where a library that uses
2688 OpenSSL de-inits, but an application is still using it). The affected
2689 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2690 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2691 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2692 COMP_zlib_cleanup().
2696 * --strict-warnings no longer enables runtime debugging options
2697 such as REF_DEBUG. Instead, debug options are automatically
2698 enabled with '--debug' builds.
2700 *Andy Polyakov, Emilia Käsper*
2702 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2703 have been moved out of the public header files. New functions for managing
2704 these have been added.
2708 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2709 objects have been moved out of the public header files. New
2710 functions for managing these have been added.
2714 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2715 have been moved out of the public header files. New functions for managing
2716 these have been added.
2720 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2721 moved out of the public header files. New functions for managing these
2726 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2730 * Removed the mk1mf build scripts.
2734 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2735 it is always safe to #include a header now.
2739 * Removed the aged BC-32 config and all its supporting scripts
2743 * Removed support for Ultrix, Netware, and OS/2.
2747 * Add support for HKDF.
2749 *Alessandro Ghedini*
2751 * Add support for blake2b and blake2s
2755 * Added support for "pipelining". Ciphers that have the
2756 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2757 encryptions/decryptions simultaneously. There are currently no built-in
2758 ciphers with this property but the expectation is that engines will be able
2759 to offer it to significantly improve throughput. Support has been extended
2760 into libssl so that multiple records for a single connection can be
2761 processed in one go (for >=TLS 1.1).
2765 * Added the AFALG engine. This is an async capable engine which is able to
2766 offload work to the Linux kernel. In this initial version it only supports
2767 AES128-CBC. The kernel must be version 4.1.0 or greater.
2771 * OpenSSL now uses a new threading API. It is no longer necessary to
2772 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2773 are two supported threading models: pthreads and windows threads. It is
2774 also possible to configure OpenSSL at compile time for "no-threads". The
2775 old threading API should no longer be used. The functions have been
2776 replaced with "no-op" compatibility macros.
2778 *Alessandro Ghedini, Matt Caswell*
2780 * Modify behavior of ALPN to invoke callback after SNI/servername
2781 callback, such that updates to the SSL_CTX affect ALPN.
2785 * Add SSL_CIPHER queries for authentication and key-exchange.
2789 * Changes to the DEFAULT cipherlist:
2790 - Prefer (EC)DHE handshakes over plain RSA.
2791 - Prefer AEAD ciphers over legacy ciphers.
2792 - Prefer ECDSA over RSA when both certificates are available.
2793 - Prefer TLSv1.2 ciphers/PRF.
2794 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2799 * Change the ECC default curve list to be this, in order: x25519,
2800 secp256r1, secp521r1, secp384r1.
2804 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2805 disabled by default. They can be re-enabled using the
2806 enable-weak-ssl-ciphers option to Configure.
2810 * If the server has ALPN configured, but supports no protocols that the
2811 client advertises, send a fatal "no_application_protocol" alert.
2812 This behaviour is SHALL in RFC 7301, though it isn't universally
2813 implemented by other servers.
2817 * Add X25519 support.
2818 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2819 for public and private key encoding using the format documented in
2820 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2821 key generation and key derivation.
2823 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2828 * Deprecate SRP_VBASE_get_by_user.
2829 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2830 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2831 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2832 seed, even if the seed is configured.
2834 Users should use SRP_VBASE_get1_by_user instead. Note that in
2835 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2836 also that even though configuring the SRP seed attempts to hide
2837 invalid usernames by continuing the handshake with fake
2838 credentials, this behaviour is not constant time and no strong
2839 guarantees are made that the handshake is indistinguishable from
2840 that of a valid user.
2844 * Configuration change; it's now possible to build dynamic engines
2845 without having to build shared libraries and vice versa. This
2846 only applies to the engines in engines/, those in crypto/engine/
2847 will always be built into libcrypto (i.e. "static").
2849 Building dynamic engines is enabled by default; to disable, use
2850 the configuration option "disable-dynamic-engine".
2852 The only requirements for building dynamic engines are the
2853 presence of the DSO module and building with position independent
2854 code, so they will also automatically be disabled if configuring
2855 with "disable-dso" or "disable-pic".
2857 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2858 are also taken away from openssl/opensslconf.h, as they are
2863 * Configuration change; if there is a known flag to compile
2864 position independent code, it will always be applied on the
2865 libcrypto and libssl object files, and never on the application
2866 object files. This means other libraries that use routines from
2867 libcrypto / libssl can be made into shared libraries regardless
2868 of how OpenSSL was configured.
2870 If this isn't desirable, the configuration options "disable-pic"
2871 or "no-pic" can be used to disable the use of PIC. This will
2872 also disable building shared libraries and dynamic engines.
2876 * Removed JPAKE code. It was experimental and has no wide use.
2880 * The INSTALL_PREFIX Makefile variable has been renamed to
2881 DESTDIR. That makes for less confusion on what this variable
2882 is for. Also, the configuration option --install_prefix is
2887 * Heartbeat for TLS has been removed and is disabled by default
2888 for DTLS; configure with enable-heartbeats. Code that uses the
2889 old #define's might need to be updated.
2891 *Emilia Käsper, Rich Salz*
2893 * Rename REF_CHECK to REF_DEBUG.
2897 * New "unified" build system
2899 The "unified" build system is aimed to be a common system for all
2900 platforms we support. With it comes new support for VMS.
2902 This system builds supports building in a different directory tree
2903 than the source tree. It produces one Makefile (for unix family
2904 or lookalikes), or one descrip.mms (for VMS).
2906 The source of information to make the Makefile / descrip.mms is
2907 small files called 'build.info', holding the necessary
2908 information for each directory with source to compile, and a
2909 template in Configurations, like unix-Makefile.tmpl or
2912 With this change, the library names were also renamed on Windows
2913 and on VMS. They now have names that are closer to the standard
2914 on Unix, and include the major version number, and in certain
2915 cases, the architecture they are built for. See "Notes on shared
2916 libraries" in INSTALL.
2918 We rely heavily on the perl module Text::Template.
2922 * Added support for auto-initialisation and de-initialisation of the library.
2923 OpenSSL no longer requires explicit init or deinit routines to be called,
2924 except in certain circumstances. See the OPENSSL_init_crypto() and
2925 OPENSSL_init_ssl() man pages for further information.
2929 * The arguments to the DTLSv1_listen function have changed. Specifically the
2930 "peer" argument is now expected to be a BIO_ADDR object.
2932 * Rewrite of BIO networking library. The BIO library lacked consistent
2933 support of IPv6, and adding it required some more extensive
2934 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2935 which hold all types of addresses and chains of address information.
2936 It also introduces a new API, with functions like BIO_socket,
2937 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2938 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2939 have been adapted accordingly.
2943 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2948 * CRIME protection: disable compression by default, even if OpenSSL is
2949 compiled with zlib enabled. Applications can still enable compression
2950 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2951 using the SSL_CONF library to configure compression.
2955 * The signature of the session callback configured with
2956 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2957 was explicitly marked as `const unsigned char*` instead of
2962 * Always DPURIFY. Remove the use of uninitialized memory in the
2963 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2967 * Removed many obsolete configuration items, including
2968 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2969 MD2_CHAR, MD2_INT, MD2_LONG
2971 IDEA_SHORT, IDEA_LONG
2972 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2974 *Rich Salz, with advice from Andy Polyakov*
2976 * Many BN internals have been moved to an internal header file.
2978 *Rich Salz with help from Andy Polyakov*
2980 * Configuration and writing out the results from it has changed.
2981 Files such as Makefile include/openssl/opensslconf.h and are now
2982 produced through general templates, such as Makefile.in and
2983 crypto/opensslconf.h.in and some help from the perl module
2986 Also, the center of configuration information is no longer
2987 Makefile. Instead, Configure produces a perl module in
2988 configdata.pm which holds most of the config data (in the hash
2989 table %config), the target data that comes from the target
2990 configuration in one of the `Configurations/*.conf` files (in
2995 * To clarify their intended purposes, the Configure options
2996 --prefix and --openssldir change their semantics, and become more
2997 straightforward and less interdependent.
2999 --prefix shall be used exclusively to give the location INSTALLTOP
3000 where programs, scripts, libraries, include files and manuals are
3001 going to be installed. The default is now /usr/local.
3003 --openssldir shall be used exclusively to give the default
3004 location OPENSSLDIR where certificates, private keys, CRLs are
3005 managed. This is also where the default openssl.cnf gets
3007 If the directory given with this option is a relative path, the
3008 values of both the --prefix value and the --openssldir value will
3009 be combined to become OPENSSLDIR.
3010 The default for --openssldir is INSTALLTOP/ssl.
3012 Anyone who uses --openssldir to specify where OpenSSL is to be
3013 installed MUST change to use --prefix instead.
3017 * The GOST engine was out of date and therefore it has been removed. An up
3018 to date GOST engine is now being maintained in an external repository.
3019 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3020 support for GOST ciphersuites (these are only activated if a GOST engine
3025 * EGD is no longer supported by default; use enable-egd when
3028 *Ben Kaduk and Rich Salz*
3030 * The distribution now has Makefile.in files, which are used to
3031 create Makefile's when Configure is run. *Configure must be run
3032 before trying to build now.*
3036 * The return value for SSL_CIPHER_description() for error conditions
3041 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3043 Obtaining and performing DNSSEC validation of TLSA records is
3044 the application's responsibility. The application provides
3045 the TLSA records of its choice to OpenSSL, and these are then
3046 used to authenticate the peer.
3048 The TLSA records need not even come from DNS. They can, for
3049 example, be used to implement local end-entity certificate or
3050 trust-anchor "pinning", where the "pin" data takes the form
3051 of TLSA records, which can augment or replace verification
3052 based on the usual WebPKI public certification authorities.
3056 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3057 continues to support deprecated interfaces in default builds.
3058 However, applications are strongly advised to compile their
3059 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3060 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3061 or the 1.1.0 releases.
3063 In environments in which all applications have been ported to
3064 not use any deprecated interfaces OpenSSL's Configure script
3065 should be used with the --api=1.1.0 option to entirely remove
3066 support for the deprecated features from the library and
3067 unconditionally disable them in the installed headers.
3068 Essentially the same effect can be achieved with the "no-deprecated"
3069 argument to Configure, except that this will always restrict
3070 the build to just the latest API, rather than a fixed API
3073 As applications are ported to future revisions of the API,
3074 they should update their compile-time OPENSSL_API_COMPAT define
3075 accordingly, but in most cases should be able to continue to
3076 compile with later releases.
3078 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3079 0x10000000L and 0x00908000L, respectively. However those
3080 versions did not support the OPENSSL_API_COMPAT feature, and
3081 so applications are not typically tested for explicit support
3082 of just the undeprecated features of either release.
3086 * Add support for setting the minimum and maximum supported protocol.
3087 It can bet set via the SSL_set_min_proto_version() and
3088 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3089 MaxProtocol. It's recommended to use the new APIs to disable
3090 protocols instead of disabling individual protocols using
3091 SSL_set_options() or SSL_CONF's Protocol. This change also
3092 removes support for disabling TLS 1.2 in the OpenSSL TLS
3093 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3097 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3101 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3102 and integrates ECDSA and ECDH functionality into EC. Implementations can
3103 now redirect key generation and no longer need to convert to or from
3106 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3107 include the ec.h header file instead.
3111 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3112 ciphers who are no longer supported and drops support the ephemeral RSA key
3113 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3117 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3118 opaque. For HMAC_CTX, the following constructors and destructors
3121 HMAC_CTX *HMAC_CTX_new(void);
3122 void HMAC_CTX_free(HMAC_CTX *ctx);
3124 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3125 destroy such methods has been added. See EVP_MD_meth_new(3) and
3126 EVP_CIPHER_meth_new(3) for documentation.
3129 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3130 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3131 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3132 an already created structure.
3133 2) For consistency with the majority of our object creators and
3134 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3135 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3136 for deprecated builds.
3140 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3141 cryptographic operations to be performed asynchronously as long as an
3142 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3143 further details. Libssl has also had this capability integrated with the
3144 introduction of the new mode SSL_MODE_ASYNC and associated error
3145 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3146 pages. This work was developed in partnership with Intel Corp.
3150 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3151 always enabled now. If you want to disable the support you should
3152 exclude it using the list of supported ciphers. This also means that the
3153 "-no_ecdhe" option has been removed from s_server.
3157 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3158 SSL_{CTX_}set1_curves() which can set a list.
3162 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3163 curve you want to support using SSL_{CTX_}set1_curves().
3167 * State machine rewrite. The state machine code has been significantly
3168 refactored in order to remove much duplication of code and solve issues
3169 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3170 further details). This change does have some associated API changes.
3171 Notably the SSL_state() function has been removed and replaced by
3172 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3173 SSL_set_state() has been removed altogether. The previous handshake states
3174 defined in ssl.h and ssl3.h have also been removed.
3178 * All instances of the string "ssleay" in the public API were replaced
3179 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3180 Some error codes related to internal RSA_eay API's were renamed.
3184 * The demo files in crypto/threads were moved to demo/threads.
3188 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3191 *Matt Caswell, Rich Salz*
3193 * New ASN.1 embed macro.
3195 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3196 structure is not allocated: it is part of the parent. That is instead of
3204 This reduces memory fragmentation and make it impossible to accidentally
3205 set a mandatory field to NULL.
3207 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3208 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3209 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3214 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3218 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3219 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3220 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3221 DES and RC4 ciphersuites.
3225 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3226 This changes the decoding behaviour for some invalid messages,
3227 though the change is mostly in the more lenient direction, and
3228 legacy behaviour is preserved as much as possible.
3232 * Fix no-stdio build.
3233 *David Woodhouse <David.Woodhouse@intel.com> and also*
3234 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3236 * New testing framework
3237 The testing framework has been largely rewritten and is now using
3238 perl and the perl modules Test::Harness and an extended variant of
3239 Test::More called OpenSSL::Test to do its work. All test scripts in
3240 test/ have been rewritten into test recipes, and all direct calls to
3241 executables in test/Makefile have become individual recipes using the
3242 simplified testing OpenSSL::Test::Simple.
3244 For documentation on our testing modules, do:
3246 perldoc test/testlib/OpenSSL/Test/Simple.pm
3247 perldoc test/testlib/OpenSSL/Test.pm
3251 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3252 are used; the latter aborts on memory leaks (usually checked on exit).
3253 Some undocumented "set malloc, etc., hooks" functions were removed
3254 and others were changed. All are now documented.
3258 * In DSA_generate_parameters_ex, if the provided seed is too short,
3261 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3263 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3264 from RFC4279, RFC4785, RFC5487, RFC5489.
3266 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3267 original RSA_PSK patch.
3271 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3272 era flag was never set throughout the codebase (only read). Also removed
3273 SSL3_FLAGS_POP_BUFFER which was only used if
3274 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3278 * Changed the default name options in the "ca", "crl", "req" and "x509"
3279 to be "oneline" instead of "compat".
3283 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3284 not aware of clients that still exhibit this bug, and the workaround
3285 hasn't been working properly for a while.
3289 * The return type of BIO_number_read() and BIO_number_written() as well as
3290 the corresponding num_read and num_write members in the BIO structure has
3291 changed from unsigned long to uint64_t. On platforms where an unsigned
3292 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3297 * Given the pervasive nature of TLS extensions it is inadvisable to run
3298 OpenSSL without support for them. It also means that maintaining
3299 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3300 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3304 * Removed support for the two export grade static DH ciphersuites
3305 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3306 were newly added (along with a number of other static DH ciphersuites) to
3307 1.0.2. However the two export ones have *never* worked since they were
3308 introduced. It seems strange in any case to be adding new export
3309 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3313 * Version negotiation has been rewritten. In particular SSLv23_method(),
3314 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3315 and turned into macros which simply call the new preferred function names
3316 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3317 should use the new names instead. Also as part of this change the ssl23.h
3318 header file has been removed.
3322 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3323 code and the associated standard is no longer considered fit-for-purpose.
3327 * RT2547 was closed. When generating a private key, try to make the
3328 output file readable only by the owner. This behavior change might
3329 be noticeable when interacting with other software.
3331 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3336 * Added HTTP GET support to the ocsp command.
3340 * Changed default digest for the dgst and enc commands from MD5 to
3345 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3349 * Added support for TLS extended master secret from
3350 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3351 initial patch which was a great help during development.
3355 * All libssl internal structures have been removed from the public header
3356 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3357 now redundant). Users should not attempt to access internal structures
3358 directly. Instead they should use the provided API functions.
3362 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3363 Access to deprecated functions can be re-enabled by running config with
3364 "enable-deprecated". In addition applications wishing to use deprecated
3365 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3366 will, by default, disable some transitive includes that previously existed
3367 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3371 * Added support for OCB mode. OpenSSL has been granted a patent license
3372 compatible with the OpenSSL license for use of OCB. Details are available
3373 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3374 for OCB can be removed by calling config with no-ocb.
3378 * SSLv2 support has been removed. It still supports receiving a SSLv2
3379 compatible client hello.
3383 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3384 done while fixing the error code for the key-too-small case.
3386 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3388 * CA.sh has been removed; use CA.pl instead.
3392 * Removed old DES API.
3396 * Remove various unsupported platforms:
3402 Sinix/ReliantUNIX RM400
3407 16-bit platforms such as WIN16
3411 * Clean up OPENSSL_NO_xxx #define's
3412 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3413 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3414 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3415 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3416 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3417 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3418 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3419 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3420 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3421 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3425 * Cleaned up dead code
3426 Remove all but one '#ifdef undef' which is to be looked at.
3430 * Clean up calling of xxx_free routines.
3431 Just like free(), fix most of the xxx_free routines to accept
3432 NULL. Remove the non-null checks from callers. Save much code.
3436 * Add secure heap for storage of private keys (when possible).
3437 Add BIO_s_secmem(), CBIGNUM, etc.
3438 Contributed by Akamai Technologies under our Corporate CLA.
3442 * Experimental support for a new, fast, unbiased prime candidate generator,
3443 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3445 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3447 * New output format NSS in the sess_id command line tool. This allows
3448 exporting the session id and the master key in NSS keylog format.
3450 *Martin Kaiser <martin@kaiser.cx>*
3452 * Harmonize version and its documentation. -f flag is used to display
3455 *mancha <mancha1@zoho.com>*
3457 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3458 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3460 *mancha <mancha1@zoho.com>*
3462 * Fix some double frees. These are not thought to be exploitable.
3464 *mancha <mancha1@zoho.com>*
3466 * A missing bounds check in the handling of the TLS heartbeat extension
3467 can be used to reveal up to 64k of memory to a connected client or
3470 Thanks for Neel Mehta of Google Security for discovering this bug and to
3471 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3472 preparing the fix [CVE-2014-0160][]
3474 *Adam Langley, Bodo Moeller*
3476 * Fix for the attack described in the paper "Recovering OpenSSL
3477 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3478 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3479 <http://eprint.iacr.org/2014/140>
3481 Thanks to Yuval Yarom and Naomi Benger for discovering this
3482 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3484 *Yuval Yarom and Naomi Benger*
3486 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3487 this fixes a limitation in previous versions of OpenSSL.
3491 * Experimental encrypt-then-mac support.
3493 Experimental support for encrypt then mac from
3494 draft-gutmann-tls-encrypt-then-mac-02.txt
3496 To enable it set the appropriate extension number (0x42 for the test
3497 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3499 For non-compliant peers (i.e. just about everything) this should have no
3502 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3506 * Add EVP support for key wrapping algorithms, to avoid problems with
3507 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3508 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3509 algorithms and include tests cases.
3513 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3518 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3519 MGF1 digest and OAEP label.
3523 * Make openssl verify return errors.
3525 *Chris Palmer <palmer@google.com> and Ben Laurie*
3527 * New function ASN1_TIME_diff to calculate the difference between two
3528 ASN1_TIME structures or one structure and the current time.
3532 * Update fips_test_suite to support multiple command line options. New
3533 test to induce all self test errors in sequence and check expected
3538 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3539 sign or verify all in one operation.
3543 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3544 test programs and fips_test_suite. Includes functionality to parse
3545 the minimal script output of fipsalgest.pl directly.
3549 * Add authorisation parameter to FIPS_module_mode_set().
3553 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3557 * Use separate DRBG fields for internal and external flags. New function
3558 FIPS_drbg_health_check() to perform on demand health checking. Add
3559 generation tests to fips_test_suite with reduced health check interval to
3560 demonstrate periodic health checking. Add "nodh" option to
3561 fips_test_suite to skip very slow DH test.
3565 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3570 * More extensive health check for DRBG checking many more failure modes.
3571 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3572 combination: call this in fips_test_suite.
3576 * Add support for canonical generation of DSA parameter 'g'. See
3579 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3580 POST to handle HMAC cases.
3584 * Add functions FIPS_module_version() and FIPS_module_version_text()
3585 to return numerical and string versions of the FIPS module number.
3589 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3590 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3591 outside the validated module in the FIPS capable OpenSSL.
3595 * Minor change to DRBG entropy callback semantics. In some cases
3596 there is no multiple of the block length between min_len and
3597 max_len. Allow the callback to return more than max_len bytes
3598 of entropy but discard any extra: it is the callback's responsibility
3599 to ensure that the extra data discarded does not impact the
3600 requested amount of entropy.
3604 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3605 information in FIPS186-3, SP800-57 and SP800-131A.
3609 * CCM support via EVP. Interface is very similar to GCM case except we
3610 must supply all data in one chunk (i.e. no update, final) and the
3611 message length must be supplied if AAD is used. Add algorithm test
3616 * Initial version of POST overhaul. Add POST callback to allow the status
3617 of POST to be monitored and/or failures induced. Modify fips_test_suite
3618 to use callback. Always run all selftests even if one fails.
3622 * XTS support including algorithm test driver in the fips_gcmtest program.
3623 Note: this does increase the maximum key length from 32 to 64 bytes but
3624 there should be no binary compatibility issues as existing applications
3625 will never use XTS mode.
3629 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3630 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3631 performs algorithm blocking for unapproved PRNG types. Also do not
3632 set PRNG type in FIPS_mode_set(): leave this to the application.
3633 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3634 the standard OpenSSL PRNG: set additional data to a date time vector.
3638 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3639 This shouldn't present any incompatibility problems because applications
3640 shouldn't be using these directly and any that are will need to rethink
3641 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3645 * Extensive self tests and health checking required by SP800-90 DRBG.
3646 Remove strength parameter from FIPS_drbg_instantiate and always
3647 instantiate at maximum supported strength.
3651 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3655 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3659 * New function DH_compute_key_padded() to compute a DH key and pad with
3660 leading zeroes if needed: this complies with SP800-56A et al.
3664 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3665 anything, incomplete, subject to change and largely untested at present.
3669 * Modify fipscanisteronly build option to only build the necessary object
3670 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3674 * Add experimental option FIPSSYMS to give all symbols in
3675 fipscanister.o and FIPS or fips prefix. This will avoid
3676 conflicts with future versions of OpenSSL. Add perl script
3677 util/fipsas.pl to preprocess assembly language source files
3678 and rename any affected symbols.
3682 * Add selftest checks and algorithm block of non-fips algorithms in
3683 FIPS mode. Remove DES2 from selftests.
3687 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3688 return internal method without any ENGINE dependencies. Add new
3689 tiny fips sign and verify functions.
3693 * New build option no-ec2m to disable characteristic 2 code.
3697 * New build option "fipscanisteronly". This only builds fipscanister.o
3698 and (currently) associated fips utilities. Uses the file Makefile.fips
3699 instead of Makefile.org as the prototype.
3703 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3704 Update fips_gcmtest to use IV generator.
3708 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3709 setting output buffer to NULL. The `*Final` function must be
3710 called although it will not retrieve any additional data. The tag
3711 can be set or retrieved with a ctrl. The IV length is by default 12
3712 bytes (96 bits) but can be set to an alternative value. If the IV
3713 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3718 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3719 underlying do_cipher function handles all cipher semantics itself
3720 including padding and finalisation. This is useful if (for example)
3721 an ENGINE cipher handles block padding itself. The behaviour of
3722 do_cipher is subtly changed if this flag is set: the return value
3723 is the number of characters written to the output buffer (zero is
3724 no longer an error code) or a negative error code. Also if the
3725 input buffer is NULL and length 0 finalisation should be performed.
3729 * If a candidate issuer certificate is already part of the constructed
3730 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3734 * Improve forward-security support: add functions
3736 void SSL_CTX_set_not_resumable_session_callback(
3737 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3738 void SSL_set_not_resumable_session_callback(
3739 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3741 for use by SSL/TLS servers; the callback function will be called whenever a
3742 new session is created, and gets to decide whether the session may be
3743 cached to make it resumable (return 0) or not (return 1). (As by the
3744 SSL/TLS protocol specifications, the session_id sent by the server will be
3745 empty to indicate that the session is not resumable; also, the server will
3746 not generate RFC 4507 (RFC 5077) session tickets.)
3748 A simple reasonable callback implementation is to return is_forward_secure.
3749 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3750 by the SSL/TLS server library, indicating whether it can provide forward
3753 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3755 * New -verify_name option in command line utilities to set verification
3760 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3761 Add CMAC pkey methods.
3765 * Experimental renegotiation in s_server -www mode. If the client
3766 browses /reneg connection is renegotiated. If /renegcert it is
3767 renegotiated requesting a certificate.
3771 * Add an "external" session cache for debugging purposes to s_server. This
3772 should help trace issues which normally are only apparent in deployed
3773 multi-process servers.
3777 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3778 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3779 BIO_set_cipher() and some obscure PEM functions were changed so they
3780 can now return an error. The RAND changes required a change to the
3781 RAND_METHOD structure.
3785 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3786 a gcc attribute to warn if the result of a function is ignored. This
3787 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3788 whose return value is often ignored.
3792 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3793 These allow SCTs (signed certificate timestamps) to be requested and
3794 validated when establishing a connection.
3796 *Rob Percival <robpercival@google.com>*
3801 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3803 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3804 used even when parsing explicit parameters, when loading a serialized key
3805 or calling `EC_GROUP_new_from_ecpkparameters()`/
3806 `EC_GROUP_new_from_ecparameters()`.
3807 This prevents bypass of security hardening and performance gains,
3808 especially for curves with specialized EC_METHODs.
3809 By default, if a key encoded with explicit parameters is loaded and later
3810 serialized, the output is still encoded with explicit parameters, even if
3811 internally a "named" EC_GROUP is used for computation.
3815 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3816 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3817 NULL. After this change, only the cofactor parameter can be NULL. It also
3818 does some minimal sanity checks on the passed order.
3823 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3824 An attack is simple, if the first CMS_recipientInfo is valid but the
3825 second CMS_recipientInfo is chosen ciphertext. If the second
3826 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3827 encryption key will be replaced by garbage, and the message cannot be
3828 decoded, but if the RSA decryption fails, the correct encryption key is
3829 used and the recipient will not notice the attack.
3830 As a work around for this potential attack the length of the decrypted
3831 key must be equal to the cipher default key length, in case the
3832 certifiate is not given and all recipientInfo are tried out.
3833 The old behaviour can be re-enabled in the CMS code by setting the
3834 CMS_DEBUG_DECRYPT flag.
3839 * Document issue with installation paths in diverse Windows builds
3841 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3842 binaries and run-time config file.
3847 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3849 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3850 This changes the size when using the genpkey app when no size is given. It
3851 fixes an omission in earlier changes that changed all RSA, DSA and DH
3852 generation apps to use 2048 bits by default.
3856 * Add FIPS support for Android Arm 64-bit
3858 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3859 Module in Version 2.0.10. For some reason, the corresponding target
3860 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3861 built with FIPS support on Android Arm 64-bit. This omission has been
3864 *Matthias St. Pierre*
3866 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3868 * 0-byte record padding oracle
3870 If an application encounters a fatal protocol error and then calls
3871 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3872 then OpenSSL can respond differently to the calling application if a 0 byte
3873 record is received with invalid padding compared to if a 0 byte record is
3874 received with an invalid MAC. If the application then behaves differently
3875 based on that in a way that is detectable to the remote peer, then this
3876 amounts to a padding oracle that could be used to decrypt data.
3878 In order for this to be exploitable "non-stitched" ciphersuites must be in
3879 use. Stitched ciphersuites are optimised implementations of certain
3880 commonly used ciphersuites. Also the application must call SSL_shutdown()
3881 twice even if a protocol error has occurred (applications should not do
3882 this but some do anyway).
3884 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3885 Aviram, with additional investigation by Steven Collison and Andrew
3886 Hourselt. It was reported to OpenSSL on 10th December 2018.
3891 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3895 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3897 * Microarchitecture timing vulnerability in ECC scalar multiplication
3899 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3900 shown to be vulnerable to a microarchitecture timing side channel attack.
3901 An attacker with sufficient access to mount local timing attacks during
3902 ECDSA signature generation could recover the private key.
3904 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3905 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3911 * Timing vulnerability in DSA signature generation
3913 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3914 timing side channel attack. An attacker could use variations in the signing
3915 algorithm to recover the private key.
3917 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3922 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3923 Module, accidentally introduced while backporting security fixes from the
3924 development branch and hindering the use of ECC in FIPS mode.
3928 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3930 * Client DoS due to large DH parameter
3932 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3933 malicious server can send a very large prime value to the client. This will
3934 cause the client to spend an unreasonably long period of time generating a
3935 key for this prime resulting in a hang until the client has finished. This
3936 could be exploited in a Denial Of Service attack.
3938 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3943 * Cache timing vulnerability in RSA Key Generation
3945 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3946 a cache timing side channel attack. An attacker with sufficient access to
3947 mount cache timing attacks during the RSA key generation process could
3948 recover the private key.
3950 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3951 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3956 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3957 parameter is no longer accepted, as it leads to a corrupt table. NULL
3958 pem_str is reserved for alias entries only.
3962 * Revert blinding in ECDSA sign and instead make problematic addition
3963 length-invariant. Switch even to fixed-length Montgomery multiplication.
3967 * Change generating and checking of primes so that the error rate of not
3968 being prime depends on the intended use based on the size of the input.
3969 For larger primes this will result in more rounds of Miller-Rabin.
3970 The maximal error rate for primes with more than 1080 bits is lowered
3973 *Kurt Roeckx, Annie Yousar*
3975 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3979 * Add blinding to ECDSA and DSA signatures to protect against side channel
3980 attacks discovered by Keegan Ryan (NCC Group).
3984 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3985 now allow empty (zero character) pass phrases.
3989 * Certificate time validation (X509_cmp_time) enforces stricter
3990 compliance with RFC 5280. Fractional seconds and timezone offsets
3991 are no longer allowed.
3995 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3997 * Constructed ASN.1 types with a recursive definition could exceed the stack
3999 Constructed ASN.1 types with a recursive definition (such as can be found
4000 in PKCS7) could eventually exceed the stack given malicious input with
4001 excessive recursion. This could result in a Denial Of Service attack. There
4002 are no such structures used within SSL/TLS that come from untrusted sources
4003 so this is considered safe.
4005 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4011 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4013 * Read/write after SSL object in error state
4015 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4016 mechanism. The intent was that if a fatal error occurred during a handshake
4017 then OpenSSL would move into the error state and would immediately fail if
4018 you attempted to continue the handshake. This works as designed for the
4019 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4020 SSL_connect()), however due to a bug it does not work correctly if
4021 SSL_read() or SSL_write() is called directly. In that scenario, if the
4022 handshake fails then a fatal error will be returned in the initial function
4023 call. If SSL_read()/SSL_write() is subsequently called by the application
4024 for the same SSL object then it will succeed and the data is passed without
4025 being decrypted/encrypted directly from the SSL/TLS record layer.
4027 In order to exploit this issue an application bug would have to be present
4028 that resulted in a call to SSL_read()/SSL_write() being issued after having
4029 already received a fatal error.
4031 This issue was reported to OpenSSL by David Benjamin (Google).
4036 * rsaz_1024_mul_avx2 overflow bug on x86_64
4038 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4039 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4040 Analysis suggests that attacks against RSA and DSA as a result of this
4041 defect would be very difficult to perform and are not believed likely.
4042 Attacks against DH1024 are considered just feasible, because most of the
4043 work necessary to deduce information about a private key may be performed
4044 offline. The amount of resources required for such an attack would be
4045 significant. However, for an attack on TLS to be meaningful, the server
4046 would have to share the DH1024 private key among multiple clients, which is
4047 no longer an option since CVE-2016-0701.
4049 This only affects processors that support the AVX2 but not ADX extensions
4050 like Intel Haswell (4th generation).
4052 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4053 was originally found via the OSS-Fuzz project.
4058 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4060 * bn_sqrx8x_internal carry bug on x86_64
4062 There is a carry propagating bug in the x86_64 Montgomery squaring
4063 procedure. No EC algorithms are affected. Analysis suggests that attacks
4064 against RSA and DSA as a result of this defect would be very difficult to
4065 perform and are not believed likely. Attacks against DH are considered just
4066 feasible (although very difficult) because most of the work necessary to
4067 deduce information about a private key may be performed offline. The amount
4068 of resources required for such an attack would be very significant and
4069 likely only accessible to a limited number of attackers. An attacker would
4070 additionally need online access to an unpatched system using the target
4071 private key in a scenario with persistent DH parameters and a private
4072 key that is shared between multiple clients.
4074 This only affects processors that support the BMI1, BMI2 and ADX extensions
4075 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4077 This issue was reported to OpenSSL by the OSS-Fuzz project.
4082 * Malformed X.509 IPAddressFamily could cause OOB read
4084 If an X.509 certificate has a malformed IPAddressFamily extension,
4085 OpenSSL could do a one-byte buffer overread. The most likely result
4086 would be an erroneous display of the certificate in text format.
4088 This issue was reported to OpenSSL by the OSS-Fuzz project.
4093 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4095 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4096 platform rather than 'mingw'.
4100 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4102 * Truncated packet could crash via OOB read
4104 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4105 cipher is being used, then a truncated packet can cause that host to
4106 perform an out-of-bounds read, usually resulting in a crash.
4108 This issue was reported to OpenSSL by Robert Święcki of Google.
4113 * BN_mod_exp may produce incorrect results on x86_64
4115 There is a carry propagating bug in the x86_64 Montgomery squaring
4116 procedure. No EC algorithms are affected. Analysis suggests that attacks
4117 against RSA and DSA as a result of this defect would be very difficult to
4118 perform and are not believed likely. Attacks against DH are considered just
4119 feasible (although very difficult) because most of the work necessary to
4120 deduce information about a private key may be performed offline. The amount
4121 of resources required for such an attack would be very significant and
4122 likely only accessible to a limited number of attackers. An attacker would
4123 additionally need online access to an unpatched system using the target
4124 private key in a scenario with persistent DH parameters and a private
4125 key that is shared between multiple clients. For example this can occur by
4126 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4127 similar to CVE-2015-3193 but must be treated as a separate problem.
4129 This issue was reported to OpenSSL by the OSS-Fuzz project.
4134 * Montgomery multiplication may produce incorrect results
4136 There is a carry propagating bug in the Broadwell-specific Montgomery
4137 multiplication procedure that handles input lengths divisible by, but
4138 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4139 and DH private keys are impossible. This is because the subroutine in
4140 question is not used in operations with the private key itself and an input
4141 of the attacker's direct choice. Otherwise the bug can manifest itself as
4142 transient authentication and key negotiation failures or reproducible
4143 erroneous outcome of public-key operations with specially crafted input.
4144 Among EC algorithms only Brainpool P-512 curves are affected and one
4145 presumably can attack ECDH key negotiation. Impact was not analyzed in
4146 detail, because pre-requisites for attack are considered unlikely. Namely
4147 multiple clients have to choose the curve in question and the server has to
4148 share the private key among them, neither of which is default behaviour.
4149 Even then only clients that chose the curve will be affected.
4151 This issue was publicly reported as transient failures and was not
4152 initially recognized as a security issue. Thanks to Richard Morgan for
4153 providing reproducible case.
4158 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4159 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4160 prevent issues where no progress is being made and the peer continually
4161 sends unrecognised record types, using up resources processing them.
4165 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4167 * Missing CRL sanity check
4169 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4170 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4171 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4173 This issue only affects the OpenSSL 1.0.2i
4178 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4180 * OCSP Status Request extension unbounded memory growth
4182 A malicious client can send an excessively large OCSP Status Request
4183 extension. If that client continually requests renegotiation, sending a
4184 large OCSP Status Request extension each time, then there will be unbounded
4185 memory growth on the server. This will eventually lead to a Denial Of
4186 Service attack through memory exhaustion. Servers with a default
4187 configuration are vulnerable even if they do not support OCSP. Builds using
4188 the "no-ocsp" build time option are not affected.
4190 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4195 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4198 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4204 * OOB write in MDC2_Update()
4206 An overflow can occur in MDC2_Update() either if called directly or
4207 through the EVP_DigestUpdate() function using MDC2. If an attacker
4208 is able to supply very large amounts of input data after a previous
4209 call to EVP_EncryptUpdate() with a partial block then a length check
4210 can overflow resulting in a heap corruption.
4212 The amount of data needed is comparable to SIZE_MAX which is impractical
4215 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4220 * Malformed SHA512 ticket DoS
4222 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4223 DoS attack where a malformed ticket will result in an OOB read which will
4226 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4227 a custom server callback and ticket lookup mechanism.
4229 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4234 * OOB write in BN_bn2dec()
4236 The function BN_bn2dec() does not check the return value of BN_div_word().
4237 This can cause an OOB write if an application uses this function with an
4238 overly large BIGNUM. This could be a problem if an overly large certificate
4239 or CRL is printed out from an untrusted source. TLS is not affected because
4240 record limits will reject an oversized certificate before it is parsed.
4242 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4247 * OOB read in TS_OBJ_print_bio()
4249 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4250 the total length the OID text representation would use and not the amount
4251 of data written. This will result in OOB reads when large OIDs are
4254 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4259 * Pointer arithmetic undefined behaviour
4261 Avoid some undefined pointer arithmetic
4263 A common idiom in the codebase is to check limits in the following manner:
4266 Where "p" points to some malloc'd data of SIZE bytes and
4269 "len" here could be from some externally supplied data (e.g. from a TLS
4272 The rules of C pointer arithmetic are such that "p + len" is only well
4273 defined where len <= SIZE. Therefore the above idiom is actually
4274 undefined behaviour.
4276 For example this could cause problems if some malloc implementation
4277 provides an address for "p" such that "p + len" actually overflows for
4278 values of len that are too big and therefore p + len < limit.
4280 This issue was reported to OpenSSL by Guido Vranken
4285 * Constant time flag not preserved in DSA signing
4287 Operations in the DSA signing algorithm should run in constant time in
4288 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4289 implementation means that a non-constant time codepath is followed for
4290 certain operations. This has been demonstrated through a cache-timing
4291 attack to be sufficient for an attacker to recover the private DSA key.
4293 This issue was reported by César Pereida (Aalto University), Billy Brumley
4294 (Tampere University of Technology), and Yuval Yarom (The University of
4295 Adelaide and NICTA).
4300 * DTLS buffered message DoS
4302 In a DTLS connection where handshake messages are delivered out-of-order
4303 those messages that OpenSSL is not yet ready to process will be buffered
4304 for later use. Under certain circumstances, a flaw in the logic means that
4305 those messages do not get removed from the buffer even though the handshake
4306 has been completed. An attacker could force up to approx. 15 messages to
4307 remain in the buffer when they are no longer required. These messages will
4308 be cleared when the DTLS connection is closed. The default maximum size for
4309 a message is 100k. Therefore the attacker could force an additional 1500k
4310 to be consumed per connection. By opening many simulataneous connections an
4311 attacker could cause a DoS attack through memory exhaustion.
4313 This issue was reported to OpenSSL by Quan Luo.
4318 * DTLS replay protection DoS
4320 A flaw in the DTLS replay attack protection mechanism means that records
4321 that arrive for future epochs update the replay protection "window" before
4322 the MAC for the record has been validated. This could be exploited by an
4323 attacker by sending a record for the next epoch (which does not have to
4324 decrypt or have a valid MAC), with a very large sequence number. This means
4325 that all subsequent legitimate packets are dropped causing a denial of
4326 service for a specific DTLS connection.
4328 This issue was reported to OpenSSL by the OCAP audit team.
4333 * Certificate message OOB reads
4335 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4336 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4337 theoretical DoS risk but this has not been observed in practice on common
4340 The messages affected are client certificate, client certificate request
4341 and server certificate. As a result the attack can only be performed
4342 against a client or a server which enables client authentication.
4344 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4349 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4351 * Prevent padding oracle in AES-NI CBC MAC check
4353 A MITM attacker can use a padding oracle attack to decrypt traffic
4354 when the connection uses an AES CBC cipher and the server support
4357 This issue was introduced as part of the fix for Lucky 13 padding
4358 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4359 constant time by making sure that always the same bytes are read and
4360 compared against either the MAC or padding bytes. But it no longer
4361 checked that there was enough data to have both the MAC and padding
4364 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4369 * Fix EVP_EncodeUpdate overflow
4371 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4372 Base64 encoding of binary data. If an attacker is able to supply very large
4373 amounts of input data then a length check can overflow resulting in a heap
4376 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4377 the `PEM_write_bio*` family of functions. These are mainly used within the
4378 OpenSSL command line applications, so any application which processes data
4379 from an untrusted source and outputs it as a PEM file should be considered
4380 vulnerable to this issue. User applications that call these APIs directly
4381 with large amounts of untrusted data may also be vulnerable.
4383 This issue was reported by Guido Vranken.
4388 * Fix EVP_EncryptUpdate overflow
4390 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4391 is able to supply very large amounts of input data after a previous call to
4392 EVP_EncryptUpdate() with a partial block then a length check can overflow
4393 resulting in a heap corruption. Following an analysis of all OpenSSL
4394 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4395 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4396 the first called function after an EVP_EncryptInit(), and therefore that
4397 specific call must be safe. The second form is where the length passed to
4398 EVP_EncryptUpdate() can be seen from the code to be some small value and
4399 therefore there is no possibility of an overflow. Since all instances are
4400 one of these two forms, it is believed that there can be no overflows in
4401 internal code due to this problem. It should be noted that
4402 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4403 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4404 of these calls have also been analysed too and it is believed there are no
4405 instances in internal usage where an overflow could occur.
4407 This issue was reported by Guido Vranken.
4412 * Prevent ASN.1 BIO excessive memory allocation
4414 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4415 a short invalid encoding can cause allocation of large amounts of memory
4416 potentially consuming excessive resources or exhausting memory.
4418 Any application parsing untrusted data through d2i BIO functions is
4419 affected. The memory based functions such as d2i_X509() are *not* affected.
4420 Since the memory based functions are used by the TLS library, TLS
4421 applications are not affected.
4423 This issue was reported by Brian Carpenter.
4430 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4431 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4432 in arbitrary stack data being returned in the buffer.
4434 This issue was reported by Guido Vranken.
4439 * Modify behavior of ALPN to invoke callback after SNI/servername
4440 callback, such that updates to the SSL_CTX affect ALPN.
4444 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4449 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4450 methods are enabled and ssl2 is disabled the methods return NULL.
4454 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4456 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4457 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4458 provide any "EXPORT" or "LOW" strength ciphers.
4462 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4463 is by default disabled at build-time. Builds that are not configured with
4464 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4465 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4466 will need to explicitly call either of:
4468 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4470 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4472 as appropriate. Even if either of those is used, or the application
4473 explicitly uses the version-specific SSLv2_method() or its client and
4474 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4475 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4476 ciphers, and SSLv2 56-bit DES are no longer available.
4481 * Fix a double-free in DSA code
4483 A double free bug was discovered when OpenSSL parses malformed DSA private
4484 keys and could lead to a DoS attack or memory corruption for applications
4485 that receive DSA private keys from untrusted sources. This scenario is
4488 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4494 * Disable SRP fake user seed to address a server memory leak.
4496 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4498 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4499 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4500 was changed to ignore the "fake user" SRP seed, even if the seed
4503 Users should use SRP_VBASE_get1_by_user instead. Note that in
4504 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4505 also that even though configuring the SRP seed attempts to hide
4506 invalid usernames by continuing the handshake with fake
4507 credentials, this behaviour is not constant time and no strong
4508 guarantees are made that the handshake is indistinguishable from
4509 that of a valid user.
4514 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4516 In the BN_hex2bn function the number of hex digits is calculated using an
4517 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4518 large values of `i` this can result in `bn_expand` not allocating any
4519 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4520 field as NULL leading to a subsequent NULL ptr deref. For very large values
4521 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4522 In this case memory is allocated to the internal BIGNUM data field, but it
4523 is insufficiently sized leading to heap corruption. A similar issue exists
4524 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4525 is ever called by user applications with very large untrusted hex/dec data.
4526 This is anticipated to be a rare occurrence.
4528 All OpenSSL internal usage of these functions use data that is not expected
4529 to be untrusted, e.g. config file data or application command line
4530 arguments. If user developed applications generate config file data based
4531 on untrusted data then it is possible that this could also lead to security
4532 consequences. This is also anticipated to be rare.
4534 This issue was reported to OpenSSL by Guido Vranken.
4539 * Fix memory issues in `BIO_*printf` functions
4541 The internal `fmtstr` function used in processing a "%s" format string in
4542 the `BIO_*printf` functions could overflow while calculating the length of a
4543 string and cause an OOB read when printing very long strings.
4545 Additionally the internal `doapr_outch` function can attempt to write to an
4546 OOB memory location (at an offset from the NULL pointer) in the event of a
4547 memory allocation failure. In 1.0.2 and below this could be caused where
4548 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4549 could be in processing a very long "%s" format string. Memory leaks can
4552 The first issue may mask the second issue dependent on compiler behaviour.
4553 These problems could enable attacks where large amounts of untrusted data
4554 is passed to the `BIO_*printf` functions. If applications use these functions
4555 in this way then they could be vulnerable. OpenSSL itself uses these
4556 functions when printing out human-readable dumps of ASN.1 data. Therefore
4557 applications that print this data could be vulnerable if the data is from
4558 untrusted sources. OpenSSL command line applications could also be
4559 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4560 as command line arguments.
4562 Libssl is not considered directly vulnerable. Additionally certificates etc
4563 received via remote connections via libssl are also unlikely to be able to
4564 trigger these issues because of message size limits enforced within libssl.
4566 This issue was reported to OpenSSL Guido Vranken.
4571 * Side channel attack on modular exponentiation
4573 A side-channel attack was found which makes use of cache-bank conflicts on
4574 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4575 of RSA keys. The ability to exploit this issue is limited as it relies on
4576 an attacker who has control of code in a thread running on the same
4577 hyper-threaded core as the victim thread which is performing decryptions.
4579 This issue was reported to OpenSSL by Yuval Yarom, The University of
4580 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4581 Nadia Heninger, University of Pennsylvania with more information at
4582 <http://cachebleed.info>.
4587 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4588 if no keysize is specified with default_bits. This fixes an
4589 omission in an earlier change that changed all RSA/DSA key generation
4590 apps to use 2048 bits by default.
4594 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4596 * DH small subgroups
4598 Historically OpenSSL only ever generated DH parameters based on "safe"
4599 primes. More recently (in version 1.0.2) support was provided for
4600 generating X9.42 style parameter files such as those required for RFC 5114
4601 support. The primes used in such files may not be "safe". Where an
4602 application is using DH configured with parameters based on primes that are
4603 not "safe" then an attacker could use this fact to find a peer's private
4604 DH exponent. This attack requires that the attacker complete multiple
4605 handshakes in which the peer uses the same private DH exponent. For example
4606 this could be used to discover a TLS server's private DH exponent if it's
4607 reusing the private DH exponent or it's using a static DH ciphersuite.
4609 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4610 TLS. It is not on by default. If the option is not set then the server
4611 reuses the same private DH exponent for the life of the server process and
4612 would be vulnerable to this attack. It is believed that many popular
4613 applications do set this option and would therefore not be at risk.
4615 The fix for this issue adds an additional check where a "q" parameter is
4616 available (as is the case in X9.42 based parameters). This detects the
4617 only known attack, and is the only possible defense for static DH
4618 ciphersuites. This could have some performance impact.
4620 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4621 default and cannot be disabled. This could have some performance impact.
4623 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4628 * SSLv2 doesn't block disabled ciphers
4630 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4631 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4632 been disabled, provided that the SSLv2 protocol was not also disabled via
4635 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4636 and Sebastian Schinzel.
4641 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4643 * BN_mod_exp may produce incorrect results on x86_64
4645 There is a carry propagating bug in the x86_64 Montgomery squaring
4646 procedure. No EC algorithms are affected. Analysis suggests that attacks
4647 against RSA and DSA as a result of this defect would be very difficult to
4648 perform and are not believed likely. Attacks against DH are considered just
4649 feasible (although very difficult) because most of the work necessary to
4650 deduce information about a private key may be performed offline. The amount
4651 of resources required for such an attack would be very significant and
4652 likely only accessible to a limited number of attackers. An attacker would
4653 additionally need online access to an unpatched system using the target
4654 private key in a scenario with persistent DH parameters and a private
4655 key that is shared between multiple clients. For example this can occur by
4656 default in OpenSSL DHE based SSL/TLS ciphersuites.
4658 This issue was reported to OpenSSL by Hanno Böck.
4663 * Certificate verify crash with missing PSS parameter
4665 The signature verification routines will crash with a NULL pointer
4666 dereference if presented with an ASN.1 signature using the RSA PSS
4667 algorithm and absent mask generation function parameter. Since these
4668 routines are used to verify certificate signature algorithms this can be
4669 used to crash any certificate verification operation and exploited in a
4670 DoS attack. Any application which performs certificate verification is
4671 vulnerable including OpenSSL clients and servers which enable client
4674 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4679 * X509_ATTRIBUTE memory leak
4681 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4682 memory. This structure is used by the PKCS#7 and CMS routines so any
4683 application which reads PKCS#7 or CMS data from untrusted sources is
4684 affected. SSL/TLS is not affected.
4686 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4692 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4693 This changes the decoding behaviour for some invalid messages,
4694 though the change is mostly in the more lenient direction, and
4695 legacy behaviour is preserved as much as possible.
4699 * In DSA_generate_parameters_ex, if the provided seed is too short,
4702 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4704 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4706 * Alternate chains certificate forgery
4708 During certificate verification, OpenSSL will attempt to find an
4709 alternative certificate chain if the first attempt to build such a chain
4710 fails. An error in the implementation of this logic can mean that an
4711 attacker could cause certain checks on untrusted certificates to be
4712 bypassed, such as the CA flag, enabling them to use a valid leaf
4713 certificate to act as a CA and "issue" an invalid certificate.
4715 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4720 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4722 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4723 incompatibility in the handling of HMAC. The previous ABI has now been
4728 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4730 * Malformed ECParameters causes infinite loop
4732 When processing an ECParameters structure OpenSSL enters an infinite loop
4733 if the curve specified is over a specially malformed binary polynomial
4736 This can be used to perform denial of service against any
4737 system which processes public keys, certificate requests or
4738 certificates. This includes TLS clients and TLS servers with
4739 client authentication enabled.
4741 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4746 * Exploitable out-of-bounds read in X509_cmp_time
4748 X509_cmp_time does not properly check the length of the ASN1_TIME
4749 string and can read a few bytes out of bounds. In addition,
4750 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4753 An attacker can use this to craft malformed certificates and CRLs of
4754 various sizes and potentially cause a segmentation fault, resulting in
4755 a DoS on applications that verify certificates or CRLs. TLS clients
4756 that verify CRLs are affected. TLS clients and servers with client
4757 authentication enabled may be affected if they use custom verification
4760 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4761 independently by Hanno Böck.
4766 * PKCS7 crash with missing EnvelopedContent
4768 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4769 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4770 with missing content and trigger a NULL pointer dereference on parsing.
4772 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4773 structures from untrusted sources are affected. OpenSSL clients and
4774 servers are not affected.
4776 This issue was reported to OpenSSL by Michal Zalewski (Google).
4781 * CMS verify infinite loop with unknown hash function
4783 When verifying a signedData message the CMS code can enter an infinite loop
4784 if presented with an unknown hash function OID. This can be used to perform
4785 denial of service against any system which verifies signedData messages using
4787 This issue was reported to OpenSSL by Johannes Bauer.
4792 * Race condition handling NewSessionTicket
4794 If a NewSessionTicket is received by a multi-threaded client when attempting to
4795 reuse a previous ticket then a race condition can occur potentially leading to
4796 a double free of the ticket data.
4801 * Only support 256-bit or stronger elliptic curves with the
4802 'ecdh_auto' setting (server) or by default (client). Of supported
4803 curves, prefer P-256 (both).
4807 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4809 * ClientHello sigalgs DoS fix
4811 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4812 invalid signature algorithms extension a NULL pointer dereference will
4813 occur. This can be exploited in a DoS attack against the server.
4815 This issue was was reported to OpenSSL by David Ramos of Stanford
4819 *Stephen Henson and Matt Caswell*
4821 * Multiblock corrupted pointer fix
4823 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4824 feature only applies on 64 bit x86 architecture platforms that support AES
4825 NI instructions. A defect in the implementation of "multiblock" can cause
4826 OpenSSL's internal write buffer to become incorrectly set to NULL when
4827 using non-blocking IO. Typically, when the user application is using a
4828 socket BIO for writing, this will only result in a failed connection.
4829 However if some other BIO is used then it is likely that a segmentation
4830 fault will be triggered, thus enabling a potential DoS attack.
4832 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4837 * Segmentation fault in DTLSv1_listen fix
4839 The DTLSv1_listen function is intended to be stateless and processes the
4840 initial ClientHello from many peers. It is common for user code to loop
4841 over the call to DTLSv1_listen until a valid ClientHello is received with
4842 an associated cookie. A defect in the implementation of DTLSv1_listen means
4843 that state is preserved in the SSL object from one invocation to the next
4844 that can lead to a segmentation fault. Errors processing the initial
4845 ClientHello can trigger this scenario. An example of such an error could be
4846 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4849 This issue was reported to OpenSSL by Per Allansson.
4854 * Segmentation fault in ASN1_TYPE_cmp fix
4856 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4857 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4858 certificate signature algorithm consistency this can be used to crash any
4859 certificate verification operation and exploited in a DoS attack. Any
4860 application which performs certificate verification is vulnerable including
4861 OpenSSL clients and servers which enable client authentication.
4866 * Segmentation fault for invalid PSS parameters fix
4868 The signature verification routines will crash with a NULL pointer
4869 dereference if presented with an ASN.1 signature using the RSA PSS
4870 algorithm and invalid parameters. Since these routines are used to verify
4871 certificate signature algorithms this can be used to crash any
4872 certificate verification operation and exploited in a DoS attack. Any
4873 application which performs certificate verification is vulnerable including
4874 OpenSSL clients and servers which enable client authentication.
4876 This issue was was reported to OpenSSL by Brian Carpenter.
4881 * ASN.1 structure reuse memory corruption fix
4883 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4884 memory corruption via an invalid write. Such reuse is and has been
4885 strongly discouraged and is believed to be rare.
4887 Applications that parse structures containing CHOICE or ANY DEFINED BY
4888 components may be affected. Certificate parsing (d2i_X509 and related
4889 functions) are however not affected. OpenSSL clients and servers are
4895 * PKCS7 NULL pointer dereferences fix
4897 The PKCS#7 parsing code does not handle missing outer ContentInfo
4898 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4899 missing content and trigger a NULL pointer dereference on parsing.
4901 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4902 otherwise parse PKCS#7 structures from untrusted sources are
4903 affected. OpenSSL clients and servers are not affected.
4905 This issue was reported to OpenSSL by Michal Zalewski (Google).
4910 * DoS via reachable assert in SSLv2 servers fix
4912 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4913 servers that both support SSLv2 and enable export cipher suites by sending
4914 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4916 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4917 (OpenSSL development team).
4922 * Empty CKE with client auth and DHE fix
4924 If client auth is used then a server can seg fault in the event of a DHE
4925 ciphersuite being selected and a zero length ClientKeyExchange message
4926 being sent by the client. This could be exploited in a DoS attack.
4931 * Handshake with unseeded PRNG fix
4933 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4934 with an unseeded PRNG. The conditions are:
4935 - The client is on a platform where the PRNG has not been seeded
4936 automatically, and the user has not seeded manually
4937 - A protocol specific client method version has been used (i.e. not
4938 SSL_client_methodv23)
4939 - A ciphersuite is used that does not require additional random data from
4940 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4942 If the handshake succeeds then the client random that has been used will
4943 have been generated from a PRNG with insufficient entropy and therefore the
4944 output may be predictable.
4946 For example using the following command with an unseeded openssl will
4947 succeed on an unpatched platform:
4949 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4954 * Use After Free following d2i_ECPrivatekey error fix
4956 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4957 could cause a use after free condition. This, in turn, could cause a double
4958 free in several private key parsing functions (such as d2i_PrivateKey
4959 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4960 for applications that receive EC private keys from untrusted
4961 sources. This scenario is considered rare.
4963 This issue was discovered by the BoringSSL project and fixed in their
4969 * X509_to_X509_REQ NULL pointer deref fix
4971 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4972 the certificate key is invalid. This function is rarely used in practice.
4974 This issue was discovered by Brian Carpenter.
4979 * Removed the export ciphers from the DEFAULT ciphers
4983 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4985 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4986 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4987 So far those who have to target multiple platforms would compromise
4988 and argue that binary targeting say ARMv5 would still execute on
4989 ARMv8. "Universal" build resolves this compromise by providing
4990 near-optimal performance even on newer platforms.
4994 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4995 (other platforms pending).
4997 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4999 * Add support for the SignedCertificateTimestampList certificate and
5000 OCSP response extensions from RFC6962.
5004 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5005 for corner cases. (Certain input points at infinity could lead to
5006 bogus results, with non-infinity inputs mapped to infinity too.)
5010 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5011 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5012 common cases are optimized and there still is room for further
5013 improvements. Vector Permutation AES for Altivec is also added.
5017 * Add support for little-endian ppc64 Linux target.
5019 *Marcelo Cerri (IBM)*
5021 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5022 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5023 are optimized and there still is room for further improvements.
5024 Both 32- and 64-bit modes are supported.
5026 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5028 * Improved ARMv7 NEON support.
5032 * Support for SPARC Architecture 2011 crypto extensions, first
5033 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5034 SHA256/512, MD5, GHASH and modular exponentiation.
5036 *Andy Polyakov, David Miller*
5038 * Accelerated modular exponentiation for Intel processors, a.k.a.
5041 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5043 * Support for new and upcoming Intel processors, including AVX2,
5044 BMI and SHA ISA extensions. This includes additional "stitched"
5045 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5048 This work was sponsored by Intel Corp.
5052 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5053 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5054 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5058 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5059 this fixes a limitation in previous versions of OpenSSL.
5063 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5064 MGF1 digest and OAEP label.
5068 * Add EVP support for key wrapping algorithms, to avoid problems with
5069 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5070 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5071 algorithms and include tests cases.
5075 * Add functions to allocate and set the fields of an ECDSA_METHOD
5078 *Douglas E. Engert, Steve Henson*
5080 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5081 difference in days and seconds between two tm or ASN1_TIME structures.
5085 * Add -rev test option to s_server to just reverse order of characters
5086 received by client and send back to server. Also prints an abbreviated
5087 summary of the connection parameters.
5091 * New option -brief for s_client and s_server to print out a brief summary
5092 of connection parameters.
5096 * Add callbacks for arbitrary TLS extensions.
5098 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5100 * New option -crl_download in several openssl utilities to download CRLs
5101 from CRLDP extension in certificates.
5105 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5109 * New function X509_CRL_diff to generate a delta CRL from the difference
5110 of two full CRLs. Add support to "crl" utility.
5114 * New functions to set lookup_crls function and to retrieve
5115 X509_STORE from X509_STORE_CTX.
5119 * Print out deprecated issuer and subject unique ID fields in
5124 * Extend OCSP I/O functions so they can be used for simple general purpose
5125 HTTP as well as OCSP. New wrapper function which can be used to download
5126 CRLs using the OCSP API.
5130 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5134 * `SSL_CONF*` functions. These provide a common framework for application
5135 configuration using configuration files or command lines.
5139 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5140 message callback and prints the results. Needs compile time option
5141 "enable-ssl-trace". New options to s_client and s_server to enable
5146 * New ctrl and macro to retrieve supported points extensions.
5147 Print out extension in s_server and s_client.
5151 * New functions to retrieve certificate signature and signature
5156 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5161 * New Suite B modes for TLS code. These use and enforce the requirements
5162 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5163 only use Suite B curves. The Suite B modes can be set by using the
5164 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5168 * New chain verification flags for Suite B levels of security. Check
5169 algorithms are acceptable when flags are set in X509_verify_cert.
5173 * Make tls1_check_chain return a set of flags indicating checks passed
5174 by a certificate chain. Add additional tests to handle client
5175 certificates: checks for matching certificate type and issuer name
5180 * If an attempt is made to use a signature algorithm not in the peer
5181 preference list abort the handshake. If client has no suitable
5182 signature algorithms in response to a certificate request do not
5183 use the certificate.
5187 * If server EC tmp key is not in client preference list abort handshake.
5191 * Add support for certificate stores in CERT structure. This makes it
5192 possible to have different stores per SSL structure or one store in
5193 the parent SSL_CTX. Include distinct stores for certificate chain
5194 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5195 to build and store a certificate chain in CERT structure: returning
5196 an error if the chain cannot be built: this will allow applications
5197 to test if a chain is correctly configured.
5199 Note: if the CERT based stores are not set then the parent SSL_CTX
5200 store is used to retain compatibility with existing behaviour.
5204 * New function ssl_set_client_disabled to set a ciphersuite disabled
5205 mask based on the current session, check mask when sending client
5206 hello and checking the requested ciphersuite.
5210 * New ctrls to retrieve and set certificate types in a certificate
5211 request message. Print out received values in s_client. If certificate
5212 types is not set with custom values set sensible values based on
5213 supported signature algorithms.
5217 * Support for distinct client and server supported signature algorithms.
5221 * Add certificate callback. If set this is called whenever a certificate
5222 is required by client or server. An application can decide which
5223 certificate chain to present based on arbitrary criteria: for example
5224 supported signature algorithms. Add very simple example to s_server.
5225 This fixes many of the problems and restrictions of the existing client
5226 certificate callback: for example you can now clear an existing
5227 certificate and specify the whole chain.
5231 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5232 the certificate can be used for (if anything). Set valid_flags field
5233 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5234 to have similar checks in it.
5236 Add new "cert_flags" field to CERT structure and include a "strict mode".
5237 This enforces some TLS certificate requirements (such as only permitting
5238 certificate signature algorithms contained in the supported algorithms
5239 extension) which some implementations ignore: this option should be used
5240 with caution as it could cause interoperability issues.
5244 * Update and tidy signature algorithm extension processing. Work out
5245 shared signature algorithms based on preferences and peer algorithms
5246 and print them out in s_client and s_server. Abort handshake if no
5247 shared signature algorithms.
5251 * Add new functions to allow customised supported signature algorithms
5252 for SSL and SSL_CTX structures. Add options to s_client and s_server
5257 * New function SSL_certs_clear() to delete all references to certificates
5258 from an SSL structure. Before this once a certificate had been added
5259 it couldn't be removed.
5263 * Integrate hostname, email address and IP address checking with certificate
5264 verification. New verify options supporting checking in openssl utility.
5268 * Fixes and wildcard matching support to hostname and email checking
5269 functions. Add manual page.
5271 *Florian Weimer (Red Hat Product Security Team)*
5273 * New functions to check a hostname email or IP address against a
5274 certificate. Add options x509 utility to print results of checks against
5279 * Fix OCSP checking.
5281 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5283 * Initial experimental support for explicitly trusted non-root CAs.
5284 OpenSSL still tries to build a complete chain to a root but if an
5285 intermediate CA has a trust setting included that is used. The first
5286 setting is used: whether to trust (e.g., -addtrust option to the x509
5291 * Add -trusted_first option which attempts to find certificates in the
5292 trusted store even if an untrusted chain is also supplied.
5296 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5297 platform support for Linux and Android.
5301 * Support for linux-x32, ILP32 environment in x86_64 framework.
5305 * Experimental multi-implementation support for FIPS capable OpenSSL.
5306 When in FIPS mode the approved implementations are used as normal,
5307 when not in FIPS mode the internal unapproved versions are used instead.
5308 This means that the FIPS capable OpenSSL isn't forced to use the
5309 (often lower performance) FIPS implementations outside FIPS mode.
5313 * Transparently support X9.42 DH parameters when calling
5314 PEM_read_bio_DHparameters. This means existing applications can handle
5315 the new parameter format automatically.
5319 * Initial experimental support for X9.42 DH parameter format: mainly
5320 to support use of 'q' parameter for RFC5114 parameters.
5324 * Add DH parameters from RFC5114 including test data to dhtest.
5328 * Support for automatic EC temporary key parameter selection. If enabled
5329 the most preferred EC parameters are automatically used instead of
5330 hardcoded fixed parameters. Now a server just has to call:
5331 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5332 support ECDH and use the most appropriate parameters.
5336 * Enhance and tidy EC curve and point format TLS extension code. Use
5337 static structures instead of allocation if default values are used.
5338 New ctrls to set curves we wish to support and to retrieve shared curves.
5339 Print out shared curves in s_server. New options to s_server and s_client
5340 to set list of supported curves.
5344 * New ctrls to retrieve supported signature algorithms and
5345 supported curve values as an array of NIDs. Extend openssl utility
5346 to print out received values.
5350 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5351 between NIDs and the more common NIST names such as "P-256". Enhance
5352 ecparam utility and ECC method to recognise the NIST names for curves.
5356 * Enhance SSL/TLS certificate chain handling to support different
5357 chains for each certificate instead of one chain in the parent SSL_CTX.
5361 * Support for fixed DH ciphersuite client authentication: where both
5362 server and client use DH certificates with common parameters.
5366 * Support for fixed DH ciphersuites: those requiring DH server
5371 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5373 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5374 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5375 X509_CINF_get_signature were reverted post internal team review.
5380 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5382 * OCSP Status Request extension unbounded memory growth
5384 A malicious client can send an excessively large OCSP Status Request
5385 extension. If that client continually requests renegotiation, sending a
5386 large OCSP Status Request extension each time, then there will be unbounded
5387 memory growth on the server. This will eventually lead to a Denial Of
5388 Service attack through memory exhaustion. Servers with a default
5389 configuration are vulnerable even if they do not support OCSP. Builds using
5390 the "no-ocsp" build time option are not affected.
5392 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5397 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5400 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5406 * OOB write in MDC2_Update()
5408 An overflow can occur in MDC2_Update() either if called directly or
5409 through the EVP_DigestUpdate() function using MDC2. If an attacker
5410 is able to supply very large amounts of input data after a previous
5411 call to EVP_EncryptUpdate() with a partial block then a length check
5412 can overflow resulting in a heap corruption.
5414 The amount of data needed is comparable to SIZE_MAX which is impractical
5417 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5422 * Malformed SHA512 ticket DoS
5424 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5425 DoS attack where a malformed ticket will result in an OOB read which will
5428 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5429 a custom server callback and ticket lookup mechanism.
5431 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5436 * OOB write in BN_bn2dec()
5438 The function BN_bn2dec() does not check the return value of BN_div_word().
5439 This can cause an OOB write if an application uses this function with an
5440 overly large BIGNUM. This could be a problem if an overly large certificate
5441 or CRL is printed out from an untrusted source. TLS is not affected because
5442 record limits will reject an oversized certificate before it is parsed.
5444 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5449 * OOB read in TS_OBJ_print_bio()
5451 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5452 the total length the OID text representation would use and not the amount
5453 of data written. This will result in OOB reads when large OIDs are
5456 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5461 * Pointer arithmetic undefined behaviour
5463 Avoid some undefined pointer arithmetic
5465 A common idiom in the codebase is to check limits in the following manner:
5468 Where "p" points to some malloc'd data of SIZE bytes and
5471 "len" here could be from some externally supplied data (e.g. from a TLS
5474 The rules of C pointer arithmetic are such that "p + len" is only well
5475 defined where len <= SIZE. Therefore the above idiom is actually
5476 undefined behaviour.
5478 For example this could cause problems if some malloc implementation
5479 provides an address for "p" such that "p + len" actually overflows for
5480 values of len that are too big and therefore p + len < limit.
5482 This issue was reported to OpenSSL by Guido Vranken
5487 * Constant time flag not preserved in DSA signing
5489 Operations in the DSA signing algorithm should run in constant time in
5490 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5491 implementation means that a non-constant time codepath is followed for
5492 certain operations. This has been demonstrated through a cache-timing
5493 attack to be sufficient for an attacker to recover the private DSA key.
5495 This issue was reported by César Pereida (Aalto University), Billy Brumley
5496 (Tampere University of Technology), and Yuval Yarom (The University of
5497 Adelaide and NICTA).
5502 * DTLS buffered message DoS
5504 In a DTLS connection where handshake messages are delivered out-of-order
5505 those messages that OpenSSL is not yet ready to process will be buffered
5506 for later use. Under certain circumstances, a flaw in the logic means that
5507 those messages do not get removed from the buffer even though the handshake
5508 has been completed. An attacker could force up to approx. 15 messages to
5509 remain in the buffer when they are no longer required. These messages will
5510 be cleared when the DTLS connection is closed. The default maximum size for
5511 a message is 100k. Therefore the attacker could force an additional 1500k
5512 to be consumed per connection. By opening many simulataneous connections an
5513 attacker could cause a DoS attack through memory exhaustion.
5515 This issue was reported to OpenSSL by Quan Luo.
5520 * DTLS replay protection DoS
5522 A flaw in the DTLS replay attack protection mechanism means that records
5523 that arrive for future epochs update the replay protection "window" before
5524 the MAC for the record has been validated. This could be exploited by an
5525 attacker by sending a record for the next epoch (which does not have to
5526 decrypt or have a valid MAC), with a very large sequence number. This means
5527 that all subsequent legitimate packets are dropped causing a denial of
5528 service for a specific DTLS connection.
5530 This issue was reported to OpenSSL by the OCAP audit team.
5535 * Certificate message OOB reads
5537 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5538 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5539 theoretical DoS risk but this has not been observed in practice on common
5542 The messages affected are client certificate, client certificate request
5543 and server certificate. As a result the attack can only be performed
5544 against a client or a server which enables client authentication.
5546 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5551 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5553 * Prevent padding oracle in AES-NI CBC MAC check
5555 A MITM attacker can use a padding oracle attack to decrypt traffic
5556 when the connection uses an AES CBC cipher and the server support
5559 This issue was introduced as part of the fix for Lucky 13 padding
5560 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5561 constant time by making sure that always the same bytes are read and
5562 compared against either the MAC or padding bytes. But it no longer
5563 checked that there was enough data to have both the MAC and padding
5566 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5571 * Fix EVP_EncodeUpdate overflow
5573 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5574 Base64 encoding of binary data. If an attacker is able to supply very large
5575 amounts of input data then a length check can overflow resulting in a heap
5578 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5579 the `PEM_write_bio*` family of functions. These are mainly used within the
5580 OpenSSL command line applications, so any application which processes data
5581 from an untrusted source and outputs it as a PEM file should be considered
5582 vulnerable to this issue. User applications that call these APIs directly
5583 with large amounts of untrusted data may also be vulnerable.
5585 This issue was reported by Guido Vranken.
5590 * Fix EVP_EncryptUpdate overflow
5592 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5593 is able to supply very large amounts of input data after a previous call to
5594 EVP_EncryptUpdate() with a partial block then a length check can overflow
5595 resulting in a heap corruption. Following an analysis of all OpenSSL
5596 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5597 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5598 the first called function after an EVP_EncryptInit(), and therefore that
5599 specific call must be safe. The second form is where the length passed to
5600 EVP_EncryptUpdate() can be seen from the code to be some small value and
5601 therefore there is no possibility of an overflow. Since all instances are
5602 one of these two forms, it is believed that there can be no overflows in
5603 internal code due to this problem. It should be noted that
5604 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5605 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5606 of these calls have also been analysed too and it is believed there are no
5607 instances in internal usage where an overflow could occur.
5609 This issue was reported by Guido Vranken.
5614 * Prevent ASN.1 BIO excessive memory allocation
5616 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5617 a short invalid encoding can casuse allocation of large amounts of memory
5618 potentially consuming excessive resources or exhausting memory.
5620 Any application parsing untrusted data through d2i BIO functions is
5621 affected. The memory based functions such as d2i_X509() are *not* affected.
5622 Since the memory based functions are used by the TLS library, TLS
5623 applications are not affected.
5625 This issue was reported by Brian Carpenter.
5632 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5633 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5634 in arbitrary stack data being returned in the buffer.
5636 This issue was reported by Guido Vranken.
5641 * Modify behavior of ALPN to invoke callback after SNI/servername
5642 callback, such that updates to the SSL_CTX affect ALPN.
5646 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5651 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5652 methods are enabled and ssl2 is disabled the methods return NULL.
5656 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5658 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5659 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5660 provide any "EXPORT" or "LOW" strength ciphers.
5664 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5665 is by default disabled at build-time. Builds that are not configured with
5666 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5667 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5668 will need to explicitly call either of:
5670 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5672 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5674 as appropriate. Even if either of those is used, or the application
5675 explicitly uses the version-specific SSLv2_method() or its client and
5676 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5677 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5678 ciphers, and SSLv2 56-bit DES are no longer available.
5683 * Fix a double-free in DSA code
5685 A double free bug was discovered when OpenSSL parses malformed DSA private
5686 keys and could lead to a DoS attack or memory corruption for applications
5687 that receive DSA private keys from untrusted sources. This scenario is
5690 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5696 * Disable SRP fake user seed to address a server memory leak.
5698 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5700 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5701 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5702 was changed to ignore the "fake user" SRP seed, even if the seed
5705 Users should use SRP_VBASE_get1_by_user instead. Note that in
5706 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5707 also that even though configuring the SRP seed attempts to hide
5708 invalid usernames by continuing the handshake with fake
5709 credentials, this behaviour is not constant time and no strong
5710 guarantees are made that the handshake is indistinguishable from
5711 that of a valid user.
5716 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5718 In the BN_hex2bn function the number of hex digits is calculated using an
5719 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5720 large values of `i` this can result in `bn_expand` not allocating any
5721 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5722 field as NULL leading to a subsequent NULL ptr deref. For very large values
5723 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5724 In this case memory is allocated to the internal BIGNUM data field, but it
5725 is insufficiently sized leading to heap corruption. A similar issue exists
5726 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5727 is ever called by user applications with very large untrusted hex/dec data.
5728 This is anticipated to be a rare occurrence.
5730 All OpenSSL internal usage of these functions use data that is not expected
5731 to be untrusted, e.g. config file data or application command line
5732 arguments. If user developed applications generate config file data based
5733 on untrusted data then it is possible that this could also lead to security
5734 consequences. This is also anticipated to be rare.
5736 This issue was reported to OpenSSL by Guido Vranken.
5741 * Fix memory issues in `BIO_*printf` functions
5743 The internal `fmtstr` function used in processing a "%s" format string in
5744 the `BIO_*printf` functions could overflow while calculating the length of a
5745 string and cause an OOB read when printing very long strings.
5747 Additionally the internal `doapr_outch` function can attempt to write to an
5748 OOB memory location (at an offset from the NULL pointer) in the event of a
5749 memory allocation failure. In 1.0.2 and below this could be caused where
5750 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5751 could be in processing a very long "%s" format string. Memory leaks can
5754 The first issue may mask the second issue dependent on compiler behaviour.
5755 These problems could enable attacks where large amounts of untrusted data
5756 is passed to the `BIO_*printf` functions. If applications use these functions
5757 in this way then they could be vulnerable. OpenSSL itself uses these
5758 functions when printing out human-readable dumps of ASN.1 data. Therefore
5759 applications that print this data could be vulnerable if the data is from
5760 untrusted sources. OpenSSL command line applications could also be
5761 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5762 as command line arguments.
5764 Libssl is not considered directly vulnerable. Additionally certificates etc
5765 received via remote connections via libssl are also unlikely to be able to
5766 trigger these issues because of message size limits enforced within libssl.
5768 This issue was reported to OpenSSL Guido Vranken.
5773 * Side channel attack on modular exponentiation
5775 A side-channel attack was found which makes use of cache-bank conflicts on
5776 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5777 of RSA keys. The ability to exploit this issue is limited as it relies on
5778 an attacker who has control of code in a thread running on the same
5779 hyper-threaded core as the victim thread which is performing decryptions.
5781 This issue was reported to OpenSSL by Yuval Yarom, The University of
5782 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5783 Nadia Heninger, University of Pennsylvania with more information at
5784 <http://cachebleed.info>.
5789 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5790 if no keysize is specified with default_bits. This fixes an
5791 omission in an earlier change that changed all RSA/DSA key generation
5792 apps to use 2048 bits by default.
5796 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5798 * Protection for DH small subgroup attacks
5800 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5801 switched on by default and cannot be disabled. This could have some
5806 * SSLv2 doesn't block disabled ciphers
5808 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5809 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5810 been disabled, provided that the SSLv2 protocol was not also disabled via
5813 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5814 and Sebastian Schinzel.
5819 * Reject DH handshakes with parameters shorter than 1024 bits.
5823 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5825 * Certificate verify crash with missing PSS parameter
5827 The signature verification routines will crash with a NULL pointer
5828 dereference if presented with an ASN.1 signature using the RSA PSS
5829 algorithm and absent mask generation function parameter. Since these
5830 routines are used to verify certificate signature algorithms this can be
5831 used to crash any certificate verification operation and exploited in a
5832 DoS attack. Any application which performs certificate verification is
5833 vulnerable including OpenSSL clients and servers which enable client
5836 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5841 * X509_ATTRIBUTE memory leak
5843 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5844 memory. This structure is used by the PKCS#7 and CMS routines so any
5845 application which reads PKCS#7 or CMS data from untrusted sources is
5846 affected. SSL/TLS is not affected.
5848 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5854 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5855 This changes the decoding behaviour for some invalid messages,
5856 though the change is mostly in the more lenient direction, and
5857 legacy behaviour is preserved as much as possible.
5861 * In DSA_generate_parameters_ex, if the provided seed is too short,
5862 use a random seed, as already documented.
5864 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5866 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5868 * Alternate chains certificate forgery
5870 During certificate verfification, OpenSSL will attempt to find an
5871 alternative certificate chain if the first attempt to build such a chain
5872 fails. An error in the implementation of this logic can mean that an
5873 attacker could cause certain checks on untrusted certificates to be
5874 bypassed, such as the CA flag, enabling them to use a valid leaf
5875 certificate to act as a CA and "issue" an invalid certificate.
5877 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5883 * Race condition handling PSK identify hint
5885 If PSK identity hints are received by a multi-threaded client then
5886 the values are wrongly updated in the parent SSL_CTX structure. This can
5887 result in a race condition potentially leading to a double free of the
5893 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5895 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5896 incompatibility in the handling of HMAC. The previous ABI has now been
5899 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5901 * Malformed ECParameters causes infinite loop
5903 When processing an ECParameters structure OpenSSL enters an infinite loop
5904 if the curve specified is over a specially malformed binary polynomial
5907 This can be used to perform denial of service against any
5908 system which processes public keys, certificate requests or
5909 certificates. This includes TLS clients and TLS servers with
5910 client authentication enabled.
5912 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5917 * Exploitable out-of-bounds read in X509_cmp_time
5919 X509_cmp_time does not properly check the length of the ASN1_TIME
5920 string and can read a few bytes out of bounds. In addition,
5921 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5924 An attacker can use this to craft malformed certificates and CRLs of
5925 various sizes and potentially cause a segmentation fault, resulting in
5926 a DoS on applications that verify certificates or CRLs. TLS clients
5927 that verify CRLs are affected. TLS clients and servers with client
5928 authentication enabled may be affected if they use custom verification
5931 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5932 independently by Hanno Böck.
5937 * PKCS7 crash with missing EnvelopedContent
5939 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5940 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5941 with missing content and trigger a NULL pointer dereference on parsing.
5943 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5944 structures from untrusted sources are affected. OpenSSL clients and
5945 servers are not affected.
5947 This issue was reported to OpenSSL by Michal Zalewski (Google).
5952 * CMS verify infinite loop with unknown hash function
5954 When verifying a signedData message the CMS code can enter an infinite loop
5955 if presented with an unknown hash function OID. This can be used to perform
5956 denial of service against any system which verifies signedData messages using
5958 This issue was reported to OpenSSL by Johannes Bauer.
5963 * Race condition handling NewSessionTicket
5965 If a NewSessionTicket is received by a multi-threaded client when attempting to
5966 reuse a previous ticket then a race condition can occur potentially leading to
5967 a double free of the ticket data.
5972 * Reject DH handshakes with parameters shorter than 768 bits.
5974 *Kurt Roeckx and Emilia Kasper*
5976 * dhparam: generate 2048-bit parameters by default.
5978 *Kurt Roeckx and Emilia Kasper*
5980 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5982 * Segmentation fault in ASN1_TYPE_cmp fix
5984 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5985 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5986 certificate signature algorithm consistency this can be used to crash any
5987 certificate verification operation and exploited in a DoS attack. Any
5988 application which performs certificate verification is vulnerable including
5989 OpenSSL clients and servers which enable client authentication.
5994 * ASN.1 structure reuse memory corruption fix
5996 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5997 memory corruption via an invalid write. Such reuse is and has been
5998 strongly discouraged and is believed to be rare.
6000 Applications that parse structures containing CHOICE or ANY DEFINED BY
6001 components may be affected. Certificate parsing (d2i_X509 and related
6002 functions) are however not affected. OpenSSL clients and servers are
6008 * PKCS7 NULL pointer dereferences fix
6010 The PKCS#7 parsing code does not handle missing outer ContentInfo
6011 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6012 missing content and trigger a NULL pointer dereference on parsing.
6014 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6015 otherwise parse PKCS#7 structures from untrusted sources are
6016 affected. OpenSSL clients and servers are not affected.
6018 This issue was reported to OpenSSL by Michal Zalewski (Google).
6023 * DoS via reachable assert in SSLv2 servers fix
6025 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6026 servers that both support SSLv2 and enable export cipher suites by sending
6027 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6029 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6030 (OpenSSL development team).
6035 * Use After Free following d2i_ECPrivatekey error fix
6037 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6038 could cause a use after free condition. This, in turn, could cause a double
6039 free in several private key parsing functions (such as d2i_PrivateKey
6040 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6041 for applications that receive EC private keys from untrusted
6042 sources. This scenario is considered rare.
6044 This issue was discovered by the BoringSSL project and fixed in their
6050 * X509_to_X509_REQ NULL pointer deref fix
6052 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6053 the certificate key is invalid. This function is rarely used in practice.
6055 This issue was discovered by Brian Carpenter.
6060 * Removed the export ciphers from the DEFAULT ciphers
6064 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6066 * Build fixes for the Windows and OpenVMS platforms
6068 *Matt Caswell and Richard Levitte*
6070 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6072 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6073 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6074 dereference. This could lead to a Denial Of Service attack. Thanks to
6075 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6080 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6081 dtls1_buffer_record function under certain conditions. In particular this
6082 could occur if an attacker sent repeated DTLS records with the same
6083 sequence number but for the next epoch. The memory leak could be exploited
6084 by an attacker in a Denial of Service attack through memory exhaustion.
6085 Thanks to Chris Mueller for reporting this issue.
6090 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6091 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6092 method would be set to NULL which could later result in a NULL pointer
6093 dereference. Thanks to Frank Schmirler for reporting this issue.
6098 * Abort handshake if server key exchange message is omitted for ephemeral
6101 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6102 reporting this issue.
6107 * Remove non-export ephemeral RSA code on client and server. This code
6108 violated the TLS standard by allowing the use of temporary RSA keys in
6109 non-export ciphersuites and could be used by a server to effectively
6110 downgrade the RSA key length used to a value smaller than the server
6111 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6112 INRIA or reporting this issue.
6117 * Fixed issue where DH client certificates are accepted without verification.
6118 An OpenSSL server will accept a DH certificate for client authentication
6119 without the certificate verify message. This effectively allows a client to
6120 authenticate without the use of a private key. This only affects servers
6121 which trust a client certificate authority which issues certificates
6122 containing DH keys: these are extremely rare and hardly ever encountered.
6123 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6129 * Ensure that the session ID context of an SSL is updated when its
6130 SSL_CTX is updated via SSL_set_SSL_CTX.
6132 The session ID context is typically set from the parent SSL_CTX,
6133 and can vary with the CTX.
6137 * Fix various certificate fingerprint issues.
6139 By using non-DER or invalid encodings outside the signed portion of a
6140 certificate the fingerprint can be changed without breaking the signature.
6141 Although no details of the signed portion of the certificate can be changed
6142 this can cause problems with some applications: e.g. those using the
6143 certificate fingerprint for blacklists.
6145 1. Reject signatures with non zero unused bits.
6147 If the BIT STRING containing the signature has non zero unused bits reject
6148 the signature. All current signature algorithms require zero unused bits.
6150 2. Check certificate algorithm consistency.
6152 Check the AlgorithmIdentifier inside TBS matches the one in the
6153 certificate signature. NB: this will result in signature failure
6154 errors for some broken certificates.
6156 Thanks to Konrad Kraszewski from Google for reporting this issue.
6158 3. Check DSA/ECDSA signatures use DER.
6160 Re-encode DSA/ECDSA signatures and compare with the original received
6161 signature. Return an error if there is a mismatch.
6163 This will reject various cases including garbage after signature
6164 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6165 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6166 (negative or with leading zeroes).
6168 Further analysis was conducted and fixes were developed by Stephen Henson
6169 of the OpenSSL core team.
6175 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6176 results on some platforms, including x86_64. This bug occurs at random
6177 with a very low probability, and is not known to be exploitable in any
6178 way, though its exact impact is difficult to determine. Thanks to Pieter
6179 Wuille (Blockstream) who reported this issue and also suggested an initial
6180 fix. Further analysis was conducted by the OpenSSL development team and
6181 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6182 the OpenSSL core team.
6187 * Do not resume sessions on the server if the negotiated protocol
6188 version does not match the session's version. Resuming with a different
6189 version, while not strictly forbidden by the RFC, is of questionable
6190 sanity and breaks all known clients.
6192 *David Benjamin, Emilia Käsper*
6194 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6195 early CCS messages during renegotiation. (Note that because
6196 renegotiation is encrypted, this early CCS was not exploitable.)
6200 * Tighten client-side session ticket handling during renegotiation:
6201 ensure that the client only accepts a session ticket if the server sends
6202 the extension anew in the ServerHello. Previously, a TLS client would
6203 reuse the old extension state and thus accept a session ticket if one was
6204 announced in the initial ServerHello.
6206 Similarly, ensure that the client requires a session ticket if one
6207 was advertised in the ServerHello. Previously, a TLS client would
6208 ignore a missing NewSessionTicket message.
6212 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6216 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6217 sends a carefully crafted handshake message, to cause OpenSSL to fail
6218 to free up to 64k of memory causing a memory leak. This could be
6219 exploited in a Denial Of Service attack. This issue affects OpenSSL
6220 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6221 whether SRTP is used or configured. Implementations of OpenSSL that
6222 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6224 The fix was developed by the OpenSSL team.
6229 * Session Ticket Memory Leak.
6231 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6232 integrity of that ticket is first verified. In the event of a session
6233 ticket integrity check failing, OpenSSL will fail to free memory
6234 causing a memory leak. By sending a large number of invalid session
6235 tickets an attacker could exploit this issue in a Denial Of Service
6241 * Build option no-ssl3 is incomplete.
6243 When OpenSSL is configured with "no-ssl3" as a build option, servers
6244 could accept and complete a SSL 3.0 handshake, and clients could be
6245 configured to send them.
6248 *Akamai and the OpenSSL team*
6250 * Add support for TLS_FALLBACK_SCSV.
6251 Client applications doing fallback retries should call
6252 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6255 *Adam Langley, Bodo Moeller*
6257 * Add additional DigestInfo checks.
6259 Re-encode DigestInto in DER and check against the original when
6260 verifying RSA signature: this will reject any improperly encoded
6261 DigestInfo structures.
6263 Note: this is a precautionary measure and no attacks are currently known.
6267 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6269 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6270 SRP code can be overrun an internal buffer. Add sanity check that
6271 g, A, B < N to SRP code.
6273 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6274 Group for discovering this issue.
6279 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6280 TLS 1.0 instead of higher protocol versions when the ClientHello message
6281 is badly fragmented. This allows a man-in-the-middle attacker to force a
6282 downgrade to TLS 1.0 even if both the server and the client support a
6283 higher protocol version, by modifying the client's TLS records.
6285 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6286 researching this issue.
6291 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6292 to a denial of service attack. A malicious server can crash the client
6293 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6294 ciphersuite and sending carefully crafted handshake messages.
6296 Thanks to Felix Gröbert (Google) for discovering and researching this
6302 * By sending carefully crafted DTLS packets an attacker could cause openssl
6303 to leak memory. This can be exploited through a Denial of Service attack.
6304 Thanks to Adam Langley for discovering and researching this issue.
6309 * An attacker can force openssl to consume large amounts of memory whilst
6310 processing DTLS handshake messages. This can be exploited through a
6311 Denial of Service attack.
6312 Thanks to Adam Langley for discovering and researching this issue.
6317 * An attacker can force an error condition which causes openssl to crash
6318 whilst processing DTLS packets due to memory being freed twice. This
6319 can be exploited through a Denial of Service attack.
6320 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6326 * If a multithreaded client connects to a malicious server using a resumed
6327 session and the server sends an ec point format extension it could write
6328 up to 255 bytes to freed memory.
6330 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6336 * A malicious server can crash an OpenSSL client with a null pointer
6337 dereference (read) by specifying an SRP ciphersuite even though it was not
6338 properly negotiated with the client. This can be exploited through a
6339 Denial of Service attack.
6341 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6342 discovering and researching this issue.
6347 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6348 X509_name_oneline, X509_name_print_ex et al. to leak some information
6349 from the stack. Applications may be affected if they echo pretty printing
6350 output to the attacker.
6352 Thanks to Ivan Fratric (Google) for discovering this issue.
6355 *Emilia Käsper, and Steve Henson*
6357 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6358 for corner cases. (Certain input points at infinity could lead to
6359 bogus results, with non-infinity inputs mapped to infinity too.)
6363 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6365 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6366 handshake can force the use of weak keying material in OpenSSL
6367 SSL/TLS clients and servers.
6369 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6370 researching this issue. [CVE-2014-0224][]
6372 *KIKUCHI Masashi, Steve Henson*
6374 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6375 OpenSSL DTLS client the code can be made to recurse eventually crashing
6378 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6381 *Imre Rad, Steve Henson*
6383 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6384 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6385 client or server. This is potentially exploitable to run arbitrary
6386 code on a vulnerable client or server.
6388 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6390 *Jüri Aedla, Steve Henson*
6392 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6393 are subject to a denial of service attack.
6395 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6396 this issue. [CVE-2014-3470][]
6398 *Felix Gröbert, Ivan Fratric, Steve Henson*
6400 * Harmonize version and its documentation. -f flag is used to display
6403 *mancha <mancha1@zoho.com>*
6405 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6406 in i2d_ECPrivateKey.
6408 *mancha <mancha1@zoho.com>*
6410 * Fix some double frees. These are not thought to be exploitable.
6412 *mancha <mancha1@zoho.com>*
6414 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6416 * A missing bounds check in the handling of the TLS heartbeat extension
6417 can be used to reveal up to 64k of memory to a connected client or
6420 Thanks for Neel Mehta of Google Security for discovering this bug and to
6421 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6422 preparing the fix [CVE-2014-0160][]
6424 *Adam Langley, Bodo Moeller*
6426 * Fix for the attack described in the paper "Recovering OpenSSL
6427 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6428 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6429 <http://eprint.iacr.org/2014/140>
6431 Thanks to Yuval Yarom and Naomi Benger for discovering this
6432 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6434 *Yuval Yarom and Naomi Benger*
6436 * TLS pad extension: draft-agl-tls-padding-03
6438 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6439 TLS client Hello record length value would otherwise be > 255 and
6440 less that 512 pad with a dummy extension containing zeroes so it
6441 is at least 512 bytes long.
6443 *Adam Langley, Steve Henson*
6445 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6447 * Fix for TLS record tampering bug. A carefully crafted invalid
6448 handshake could crash OpenSSL with a NULL pointer exception.
6449 Thanks to Anton Johansson for reporting this issues.
6452 * Keep original DTLS digest and encryption contexts in retransmission
6453 structures so we can use the previous session parameters if they need
6454 to be resent. [CVE-2013-6450][]
6458 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6459 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6460 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6461 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6462 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6463 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6465 *Rob Stradling, Adam Langley*
6467 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6469 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6470 supporting platforms or when small records were transferred.
6472 *Andy Polyakov, Steve Henson*
6474 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6476 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6478 This addresses the flaw in CBC record processing discovered by
6479 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6480 at: <http://www.isg.rhul.ac.uk/tls/>
6482 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6483 Security Group at Royal Holloway, University of London
6484 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6485 Emilia Käsper for the initial patch.
6488 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6490 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6491 ciphersuites which can be exploited in a denial of service attack.
6492 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6493 and detecting this bug and to Wolfgang Ettlinger
6494 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6499 * Return an error when checking OCSP signatures when key is NULL.
6500 This fixes a DoS attack. [CVE-2013-0166][]
6504 * Make openssl verify return errors.
6506 *Chris Palmer <palmer@google.com> and Ben Laurie*
6508 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6509 the right response is stapled. Also change SSL_get_certificate()
6510 so it returns the certificate actually sent.
6511 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6513 *Rob Stradling <rob.stradling@comodo.com>*
6515 * Fix possible deadlock when decoding public keys.
6519 * Don't use TLS 1.0 record version number in initial client hello
6524 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6526 * Sanity check record length before skipping explicit IV in TLS
6527 1.2, 1.1 and DTLS to fix DoS attack.
6529 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6530 fuzzing as a service testing platform.
6535 * Initialise tkeylen properly when encrypting CMS messages.
6536 Thanks to Solar Designer of Openwall for reporting this issue.
6540 * In FIPS mode don't try to use composite ciphers as they are not
6545 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6547 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6548 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6549 mean any application compiled against OpenSSL 1.0.0 headers setting
6550 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6551 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6552 0x10000000L Any application which was previously compiled against
6553 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6554 will need to be recompiled as a result. Letting be results in
6555 inability to disable specifically TLS 1.1 and in client context,
6556 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6560 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6561 disable just protocol X, but all protocols above X *if* there are
6562 protocols *below* X still enabled. In more practical terms it means
6563 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6564 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6565 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6570 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6572 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6573 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6574 in CRYPTO_realloc_clean.
6576 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6577 issue and to Adam Langley <agl@chromium.org> for fixing it.
6580 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6582 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6586 * Workarounds for some broken servers that "hang" if a client hello
6587 record length exceeds 255 bytes.
6589 1. Do not use record version number > TLS 1.0 in initial client
6590 hello: some (but not all) hanging servers will now work.
6591 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6592 the number of ciphers sent in the client hello. This should be
6593 set to an even number, such as 50, for example by passing:
6594 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6595 Most broken servers should now work.
6596 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6597 TLS 1.2 client support entirely.
6601 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6605 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6607 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6608 STRING form instead of a DigestInfo.
6612 * The format used for MDC2 RSA signatures is inconsistent between EVP
6613 and the RSA_sign/RSA_verify functions. This was made more apparent when
6614 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6615 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6616 the correct format in RSA_verify so both forms transparently work.
6620 * Some servers which support TLS 1.0 can choke if we initially indicate
6621 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6622 encrypted premaster secret. As a workaround use the maximum permitted
6623 client version in client hello, this should keep such servers happy
6624 and still work with previous versions of OpenSSL.
6628 * Add support for TLS/DTLS heartbeats.
6630 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6632 * Add support for SCTP.
6634 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6636 * Improved PRNG seeding for VOS.
6638 *Paul Green <Paul.Green@stratus.com>*
6640 * Extensive assembler packs updates, most notably:
6642 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6643 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6644 - x86_64: bit-sliced AES implementation;
6645 - ARM: NEON support, contemporary platforms optimizations;
6646 - s390x: z196 support;
6647 - `*`: GHASH and GF(2^m) multiplication implementations;
6651 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6652 (removal of unnecessary code)
6654 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6656 * Add TLS key material exporter from RFC 5705.
6660 * Add DTLS-SRTP negotiation from RFC 5764.
6664 * Add Next Protocol Negotiation,
6665 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6666 disabled with a no-npn flag to config or Configure. Code donated
6669 *Adam Langley <agl@google.com> and Ben Laurie*
6671 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6672 NIST-P256, NIST-P521, with constant-time single point multiplication on
6673 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6674 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6675 Code made available under Apache License version 2.0.
6677 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6678 line to include this in your build of OpenSSL, and run "make depend" (or
6679 "make update"). This enables the following EC_METHODs:
6681 EC_GFp_nistp224_method()
6682 EC_GFp_nistp256_method()
6683 EC_GFp_nistp521_method()
6685 EC_GROUP_new_by_curve_name() will automatically use these (while
6686 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6689 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6691 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6692 all platforms. Move ssize_t definition from e_os.h to the public
6693 header file e_os2.h as it now appears in public header file cms.h
6697 * New -sigopt option to the ca, req and x509 utilities. Additional
6698 signature parameters can be passed using this option and in
6703 * Add RSA PSS signing function. This will generate and set the
6704 appropriate AlgorithmIdentifiers for PSS based on those in the
6705 corresponding EVP_MD_CTX structure. No application support yet.
6709 * Support for companion algorithm specific ASN1 signing routines.
6710 New function ASN1_item_sign_ctx() signs a pre-initialised
6711 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6712 the appropriate parameters.
6716 * Add new algorithm specific ASN1 verification initialisation function
6717 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6718 handling will be the same no matter what EVP_PKEY_METHOD is used.
6719 Add a PSS handler to support verification of PSS signatures: checked
6720 against a number of sample certificates.
6724 * Add signature printing for PSS. Add PSS OIDs.
6726 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6728 * Add algorithm specific signature printing. An individual ASN1 method
6729 can now print out signatures instead of the standard hex dump.
6731 More complex signatures (e.g. PSS) can print out more meaningful
6732 information. Include DSA version that prints out the signature
6737 * Password based recipient info support for CMS library: implementing
6742 * Split password based encryption into PBES2 and PBKDF2 functions. This
6743 neatly separates the code into cipher and PBE sections and is required
6744 for some algorithms that split PBES2 into separate pieces (such as
6745 password based CMS).
6749 * Session-handling fixes:
6750 - Fix handling of connections that are resuming with a session ID,
6751 but also support Session Tickets.
6752 - Fix a bug that suppressed issuing of a new ticket if the client
6753 presented a ticket with an expired session.
6754 - Try to set the ticket lifetime hint to something reasonable.
6755 - Make tickets shorter by excluding irrelevant information.
6756 - On the client side, don't ignore renewed tickets.
6758 *Adam Langley, Bodo Moeller (Google)*
6760 * Fix PSK session representation.
6764 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6766 This work was sponsored by Intel.
6770 * Add GCM support to TLS library. Some custom code is needed to split
6771 the IV between the fixed (from PRF) and explicit (from TLS record)
6772 portions. This adds all GCM ciphersuites supported by RFC5288 and
6773 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6774 add a special AESGCM string for GCM only.
6778 * Expand range of ctrls for AES GCM. Permit setting invocation
6779 field on decrypt and retrieval of invocation field only on encrypt.
6783 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6784 As required by RFC5289 these ciphersuites cannot be used if for
6785 versions of TLS earlier than 1.2.
6789 * For FIPS capable OpenSSL interpret a NULL default public key method
6790 as unset and return the appropriate default but do *not* set the default.
6791 This means we can return the appropriate method in applications that
6792 switch between FIPS and non-FIPS modes.
6796 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6797 ENGINE is used then we cannot handle that in the FIPS module so we
6798 keep original code iff non-FIPS operations are allowed.
6802 * Add -attime option to openssl utilities.
6804 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6806 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6810 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6811 FIPS EC methods unconditionally for now.
6815 * New build option no-ec2m to disable characteristic 2 code.
6819 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6820 all cases can be covered as some introduce binary incompatibilities.
6824 * Redirect RSA operations to FIPS module including keygen,
6825 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6829 * Add similar low level API blocking to ciphers.
6833 * Low level digest APIs are not approved in FIPS mode: any attempt
6834 to use these will cause a fatal error. Applications that *really* want
6835 to use them can use the `private_*` version instead.
6839 * Redirect cipher operations to FIPS module for FIPS builds.
6843 * Redirect digest operations to FIPS module for FIPS builds.
6847 * Update build system to add "fips" flag which will link in fipscanister.o
6848 for static and shared library builds embedding a signature if needed.
6852 * Output TLS supported curves in preference order instead of numerical
6853 order. This is currently hardcoded for the highest order curves first.
6854 This should be configurable so applications can judge speed vs strength.
6858 * Add TLS v1.2 server support for client authentication.
6862 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6867 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6868 FIPS modules versions.
6872 * Add TLS v1.2 client side support for client authentication. Keep cache
6873 of handshake records longer as we don't know the hash algorithm to use
6874 until after the certificate request message is received.
6878 * Initial TLS v1.2 client support. Add a default signature algorithms
6879 extension including all the algorithms we support. Parse new signature
6880 format in client key exchange. Relax some ECC signing restrictions for
6881 TLS v1.2 as indicated in RFC5246.
6885 * Add server support for TLS v1.2 signature algorithms extension. Switch
6886 to new signature format when needed using client digest preference.
6887 All server ciphersuites should now work correctly in TLS v1.2. No client
6888 support yet and no support for client certificates.
6892 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6893 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6894 ciphersuites. At present only RSA key exchange ciphersuites work with
6895 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6896 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6897 and version checking.
6901 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6902 with this defined it will not be affected by any changes to ssl internal
6903 structures. Add several utility functions to allow openssl application
6904 to work with OPENSSL_NO_SSL_INTERN defined.
6908 * A long standing patch to add support for SRP from EdelWeb (Peter
6909 Sylvester and Christophe Renou) was integrated.
6910 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6911 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6914 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6918 * Permit abbreviated handshakes when renegotiating using the function
6919 SSL_renegotiate_abbreviated().
6921 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6923 * Add call to ENGINE_register_all_complete() to
6924 ENGINE_load_builtin_engines(), so some implementations get used
6925 automatically instead of needing explicit application support.
6929 * Add support for TLS key exporter as described in RFC5705.
6931 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6933 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6934 a few changes are required:
6936 Add SSL_OP_NO_TLSv1_1 flag.
6937 Add TLSv1_1 methods.
6938 Update version checking logic to handle version 1.1.
6939 Add explicit IV handling (ported from DTLS code).
6940 Add command line options to s_client/s_server.
6947 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6949 * X509_ATTRIBUTE memory leak
6951 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6952 memory. This structure is used by the PKCS#7 and CMS routines so any
6953 application which reads PKCS#7 or CMS data from untrusted sources is
6954 affected. SSL/TLS is not affected.
6956 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6962 * Race condition handling PSK identify hint
6964 If PSK identity hints are received by a multi-threaded client then
6965 the values are wrongly updated in the parent SSL_CTX structure. This can
6966 result in a race condition potentially leading to a double free of the
6972 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6974 * Malformed ECParameters causes infinite loop
6976 When processing an ECParameters structure OpenSSL enters an infinite loop
6977 if the curve specified is over a specially malformed binary polynomial
6980 This can be used to perform denial of service against any
6981 system which processes public keys, certificate requests or
6982 certificates. This includes TLS clients and TLS servers with
6983 client authentication enabled.
6985 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6990 * Exploitable out-of-bounds read in X509_cmp_time
6992 X509_cmp_time does not properly check the length of the ASN1_TIME
6993 string and can read a few bytes out of bounds. In addition,
6994 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6997 An attacker can use this to craft malformed certificates and CRLs of
6998 various sizes and potentially cause a segmentation fault, resulting in
6999 a DoS on applications that verify certificates or CRLs. TLS clients
7000 that verify CRLs are affected. TLS clients and servers with client
7001 authentication enabled may be affected if they use custom verification
7004 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7005 independently by Hanno Böck.
7010 * PKCS7 crash with missing EnvelopedContent
7012 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7013 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7014 with missing content and trigger a NULL pointer dereference on parsing.
7016 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7017 structures from untrusted sources are affected. OpenSSL clients and
7018 servers are not affected.
7020 This issue was reported to OpenSSL by Michal Zalewski (Google).
7025 * CMS verify infinite loop with unknown hash function
7027 When verifying a signedData message the CMS code can enter an infinite loop
7028 if presented with an unknown hash function OID. This can be used to perform
7029 denial of service against any system which verifies signedData messages using
7031 This issue was reported to OpenSSL by Johannes Bauer.
7036 * Race condition handling NewSessionTicket
7038 If a NewSessionTicket is received by a multi-threaded client when attempting to
7039 reuse a previous ticket then a race condition can occur potentially leading to
7040 a double free of the ticket data.
7045 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7047 * Segmentation fault in ASN1_TYPE_cmp fix
7049 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7050 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7051 certificate signature algorithm consistency this can be used to crash any
7052 certificate verification operation and exploited in a DoS attack. Any
7053 application which performs certificate verification is vulnerable including
7054 OpenSSL clients and servers which enable client authentication.
7059 * ASN.1 structure reuse memory corruption fix
7061 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7062 memory corruption via an invalid write. Such reuse is and has been
7063 strongly discouraged and is believed to be rare.
7065 Applications that parse structures containing CHOICE or ANY DEFINED BY
7066 components may be affected. Certificate parsing (d2i_X509 and related
7067 functions) are however not affected. OpenSSL clients and servers are
7073 * PKCS7 NULL pointer dereferences fix
7075 The PKCS#7 parsing code does not handle missing outer ContentInfo
7076 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7077 missing content and trigger a NULL pointer dereference on parsing.
7079 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7080 otherwise parse PKCS#7 structures from untrusted sources are
7081 affected. OpenSSL clients and servers are not affected.
7083 This issue was reported to OpenSSL by Michal Zalewski (Google).
7088 * DoS via reachable assert in SSLv2 servers fix
7090 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7091 servers that both support SSLv2 and enable export cipher suites by sending
7092 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7094 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7095 (OpenSSL development team).
7100 * Use After Free following d2i_ECPrivatekey error fix
7102 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7103 could cause a use after free condition. This, in turn, could cause a double
7104 free in several private key parsing functions (such as d2i_PrivateKey
7105 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7106 for applications that receive EC private keys from untrusted
7107 sources. This scenario is considered rare.
7109 This issue was discovered by the BoringSSL project and fixed in their
7115 * X509_to_X509_REQ NULL pointer deref fix
7117 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7118 the certificate key is invalid. This function is rarely used in practice.
7120 This issue was discovered by Brian Carpenter.
7125 * Removed the export ciphers from the DEFAULT ciphers
7129 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7131 * Build fixes for the Windows and OpenVMS platforms
7133 *Matt Caswell and Richard Levitte*
7135 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7137 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7138 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7139 dereference. This could lead to a Denial Of Service attack. Thanks to
7140 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7145 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7146 dtls1_buffer_record function under certain conditions. In particular this
7147 could occur if an attacker sent repeated DTLS records with the same
7148 sequence number but for the next epoch. The memory leak could be exploited
7149 by an attacker in a Denial of Service attack through memory exhaustion.
7150 Thanks to Chris Mueller for reporting this issue.
7155 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7156 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7157 method would be set to NULL which could later result in a NULL pointer
7158 dereference. Thanks to Frank Schmirler for reporting this issue.
7163 * Abort handshake if server key exchange message is omitted for ephemeral
7166 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7167 reporting this issue.
7172 * Remove non-export ephemeral RSA code on client and server. This code
7173 violated the TLS standard by allowing the use of temporary RSA keys in
7174 non-export ciphersuites and could be used by a server to effectively
7175 downgrade the RSA key length used to a value smaller than the server
7176 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7177 INRIA or reporting this issue.
7182 * Fixed issue where DH client certificates are accepted without verification.
7183 An OpenSSL server will accept a DH certificate for client authentication
7184 without the certificate verify message. This effectively allows a client to
7185 authenticate without the use of a private key. This only affects servers
7186 which trust a client certificate authority which issues certificates
7187 containing DH keys: these are extremely rare and hardly ever encountered.
7188 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7194 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7195 results on some platforms, including x86_64. This bug occurs at random
7196 with a very low probability, and is not known to be exploitable in any
7197 way, though its exact impact is difficult to determine. Thanks to Pieter
7198 Wuille (Blockstream) who reported this issue and also suggested an initial
7199 fix. Further analysis was conducted by the OpenSSL development team and
7200 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7201 the OpenSSL core team.
7206 * Fix various certificate fingerprint issues.
7208 By using non-DER or invalid encodings outside the signed portion of a
7209 certificate the fingerprint can be changed without breaking the signature.
7210 Although no details of the signed portion of the certificate can be changed
7211 this can cause problems with some applications: e.g. those using the
7212 certificate fingerprint for blacklists.
7214 1. Reject signatures with non zero unused bits.
7216 If the BIT STRING containing the signature has non zero unused bits reject
7217 the signature. All current signature algorithms require zero unused bits.
7219 2. Check certificate algorithm consistency.
7221 Check the AlgorithmIdentifier inside TBS matches the one in the
7222 certificate signature. NB: this will result in signature failure
7223 errors for some broken certificates.
7225 Thanks to Konrad Kraszewski from Google for reporting this issue.
7227 3. Check DSA/ECDSA signatures use DER.
7229 Reencode DSA/ECDSA signatures and compare with the original received
7230 signature. Return an error if there is a mismatch.
7232 This will reject various cases including garbage after signature
7233 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7234 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7235 (negative or with leading zeroes).
7237 Further analysis was conducted and fixes were developed by Stephen Henson
7238 of the OpenSSL core team.
7244 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7246 * Session Ticket Memory Leak.
7248 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7249 integrity of that ticket is first verified. In the event of a session
7250 ticket integrity check failing, OpenSSL will fail to free memory
7251 causing a memory leak. By sending a large number of invalid session
7252 tickets an attacker could exploit this issue in a Denial Of Service
7258 * Build option no-ssl3 is incomplete.
7260 When OpenSSL is configured with "no-ssl3" as a build option, servers
7261 could accept and complete a SSL 3.0 handshake, and clients could be
7262 configured to send them.
7265 *Akamai and the OpenSSL team*
7267 * Add support for TLS_FALLBACK_SCSV.
7268 Client applications doing fallback retries should call
7269 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7272 *Adam Langley, Bodo Moeller*
7274 * Add additional DigestInfo checks.
7276 Reencode DigestInto in DER and check against the original when
7277 verifying RSA signature: this will reject any improperly encoded
7278 DigestInfo structures.
7280 Note: this is a precautionary measure and no attacks are currently known.
7284 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7286 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7287 to a denial of service attack. A malicious server can crash the client
7288 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7289 ciphersuite and sending carefully crafted handshake messages.
7291 Thanks to Felix Gröbert (Google) for discovering and researching this
7297 * By sending carefully crafted DTLS packets an attacker could cause openssl
7298 to leak memory. This can be exploited through a Denial of Service attack.
7299 Thanks to Adam Langley for discovering and researching this issue.
7304 * An attacker can force openssl to consume large amounts of memory whilst
7305 processing DTLS handshake messages. This can be exploited through a
7306 Denial of Service attack.
7307 Thanks to Adam Langley for discovering and researching this issue.
7312 * An attacker can force an error condition which causes openssl to crash
7313 whilst processing DTLS packets due to memory being freed twice. This
7314 can be exploited through a Denial of Service attack.
7315 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7321 * If a multithreaded client connects to a malicious server using a resumed
7322 session and the server sends an ec point format extension it could write
7323 up to 255 bytes to freed memory.
7325 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7331 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7332 X509_name_oneline, X509_name_print_ex et al. to leak some information
7333 from the stack. Applications may be affected if they echo pretty printing
7334 output to the attacker.
7336 Thanks to Ivan Fratric (Google) for discovering this issue.
7339 *Emilia Käsper, and Steve Henson*
7341 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7342 for corner cases. (Certain input points at infinity could lead to
7343 bogus results, with non-infinity inputs mapped to infinity too.)
7347 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7349 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7350 handshake can force the use of weak keying material in OpenSSL
7351 SSL/TLS clients and servers.
7353 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7354 researching this issue. [CVE-2014-0224][]
7356 *KIKUCHI Masashi, Steve Henson*
7358 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7359 OpenSSL DTLS client the code can be made to recurse eventually crashing
7362 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7365 *Imre Rad, Steve Henson*
7367 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7368 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7369 client or server. This is potentially exploitable to run arbitrary
7370 code on a vulnerable client or server.
7372 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7374 *Jüri Aedla, Steve Henson*
7376 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7377 are subject to a denial of service attack.
7379 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7380 this issue. [CVE-2014-3470][]
7382 *Felix Gröbert, Ivan Fratric, Steve Henson*
7384 * Harmonize version and its documentation. -f flag is used to display
7387 *mancha <mancha1@zoho.com>*
7389 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7390 in i2d_ECPrivateKey.
7392 *mancha <mancha1@zoho.com>*
7394 * Fix some double frees. These are not thought to be exploitable.
7396 *mancha <mancha1@zoho.com>*
7398 * Fix for the attack described in the paper "Recovering OpenSSL
7399 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7400 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7401 <http://eprint.iacr.org/2014/140>
7403 Thanks to Yuval Yarom and Naomi Benger for discovering this
7404 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7406 *Yuval Yarom and Naomi Benger*
7408 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7410 * Keep original DTLS digest and encryption contexts in retransmission
7411 structures so we can use the previous session parameters if they need
7412 to be resent. [CVE-2013-6450][]
7416 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7417 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7418 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7419 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7420 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7421 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7423 *Rob Stradling, Adam Langley*
7425 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7427 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7429 This addresses the flaw in CBC record processing discovered by
7430 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7431 at: <http://www.isg.rhul.ac.uk/tls/>
7433 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7434 Security Group at Royal Holloway, University of London
7435 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7436 Emilia Käsper for the initial patch.
7439 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7441 * Return an error when checking OCSP signatures when key is NULL.
7442 This fixes a DoS attack. [CVE-2013-0166][]
7446 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7447 the right response is stapled. Also change SSL_get_certificate()
7448 so it returns the certificate actually sent.
7449 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7450 (This is a backport)
7452 *Rob Stradling <rob.stradling@comodo.com>*
7454 * Fix possible deadlock when decoding public keys.
7458 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7460 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7463 * Sanity check record length before skipping explicit IV in DTLS
7466 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7467 fuzzing as a service testing platform.
7472 * Initialise tkeylen properly when encrypting CMS messages.
7473 Thanks to Solar Designer of Openwall for reporting this issue.
7477 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7479 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7480 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7481 in CRYPTO_realloc_clean.
7483 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7484 issue and to Adam Langley <agl@chromium.org> for fixing it.
7487 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7489 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7491 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7492 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7493 content decryption and always return the same error. Note: this attack
7494 needs on average 2^20 messages so it only affects automated senders. The
7495 old behaviour can be re-enabled in the CMS code by setting the
7496 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7497 an MMA defence is not necessary.
7498 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7499 this issue. [CVE-2012-0884][]
7503 * Fix CVE-2011-4619: make sure we really are receiving a
7504 client hello before rejecting multiple SGC restarts. Thanks to
7505 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7509 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7511 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7512 Thanks to Antonio Martin, Enterprise Secure Access Research and
7513 Development, Cisco Systems, Inc. for discovering this bug and
7514 preparing a fix. [CVE-2012-0050][]
7518 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7520 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7521 of the Vaudenay padding oracle attack on CBC mode encryption
7522 which enables an efficient plaintext recovery attack against
7523 the OpenSSL implementation of DTLS. Their attack exploits timing
7524 differences arising during decryption processing. A research
7525 paper describing this attack can be found at:
7526 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7527 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7528 Security Group at Royal Holloway, University of London
7529 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7530 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7531 for preparing the fix. [CVE-2011-4108][]
7533 *Robin Seggelmann, Michael Tuexen*
7535 * Clear bytes used for block padding of SSL 3.0 records.
7538 *Adam Langley (Google)*
7540 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7541 Kadianakis <desnacked@gmail.com> for discovering this issue and
7542 Adam Langley for preparing the fix. [CVE-2011-4619][]
7544 *Adam Langley (Google)*
7546 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7548 *Andrey Kulikov <amdeich@gmail.com>*
7550 * Prevent malformed RFC3779 data triggering an assertion failure.
7551 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7552 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7554 *Rob Austein <sra@hactrn.net>*
7556 * Improved PRNG seeding for VOS.
7558 *Paul Green <Paul.Green@stratus.com>*
7560 * Fix ssl_ciph.c set-up race.
7562 *Adam Langley (Google)*
7564 * Fix spurious failures in ecdsatest.c.
7566 *Emilia Käsper (Google)*
7568 * Fix the BIO_f_buffer() implementation (which was mixing different
7569 interpretations of the `..._len` fields).
7571 *Adam Langley (Google)*
7573 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7574 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7575 threads won't reuse the same blinding coefficients.
7577 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7578 lock to call BN_BLINDING_invert_ex, and avoids one use of
7579 BN_BLINDING_update for each BN_BLINDING structure (previously,
7580 the last update always remained unused).
7582 *Emilia Käsper (Google)*
7584 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7586 *Bob Buckholz (Google)*
7588 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7590 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7591 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7593 *Kaspar Brand <ossl@velox.ch>*
7595 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7596 for multi-threaded use of ECDH. [CVE-2011-3210][]
7598 *Adam Langley (Google)*
7600 * Fix x509_name_ex_d2i memory leak on bad inputs.
7604 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7605 signature public key algorithm by using OID xref utilities instead.
7606 Before this you could only use some ECC ciphersuites with SHA1 only.
7610 * Add protection against ECDSA timing attacks as mentioned in the paper
7611 by Billy Bob Brumley and Nicola Tuveri, see:
7612 <http://eprint.iacr.org/2011/232.pdf>
7614 *Billy Bob Brumley and Nicola Tuveri*
7616 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7618 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7620 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7622 * Fix bug in string printing code: if *any* escaping is enabled we must
7623 escape the escape character (backslash) or the resulting string is
7628 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7630 * Disable code workaround for ancient and obsolete Netscape browsers
7631 and servers: an attacker can use it in a ciphersuite downgrade attack.
7632 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7636 * Fixed J-PAKE implementation error, originally discovered by
7637 Sebastien Martini, further info and confirmation from Stefan
7638 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7642 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7644 * Fix extension code to avoid race conditions which can result in a buffer
7645 overrun vulnerability: resumed sessions must not be modified as they can
7646 be shared by multiple threads. CVE-2010-3864
7650 * Fix WIN32 build system to correctly link an ENGINE directory into
7655 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7657 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7660 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7662 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7664 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7665 context. The operation can be customised via the ctrl mechanism in
7666 case ENGINEs want to include additional functionality.
7670 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7674 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7675 output hashes compatible with older versions of OpenSSL.
7677 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7679 * Fix compression algorithm handling: if resuming a session use the
7680 compression algorithm of the resumed session instead of determining
7681 it from client hello again. Don't allow server to change algorithm.
7685 * Add load_crls() function to apps tidying load_certs() too. Add option
7686 to verify utility to allow additional CRLs to be included.
7690 * Update OCSP request code to permit adding custom headers to the request:
7691 some responders need this.
7695 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7698 *Julia Lawall <julia@diku.dk>*
7700 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7701 needlessly dereferenced structures, used obsolete functions and
7702 didn't handle all updated verify codes correctly.
7706 * Disable MD2 in the default configuration.
7710 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7711 indicate the initial BIO being pushed or popped. This makes it possible
7712 to determine whether the BIO is the one explicitly called or as a result
7713 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7714 it handles reference counts correctly and doesn't zero out the I/O bio
7715 when it is not being explicitly popped. WARNING: applications which
7716 included workarounds for the old buggy behaviour will need to be modified
7717 or they could free up already freed BIOs.
7721 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7722 renaming to all platforms (within the 0.9.8 branch, this was
7723 done conditionally on Netware platforms to avoid a name clash).
7725 *Guenter <lists@gknw.net>*
7727 * Add ECDHE and PSK support to DTLS.
7729 *Michael Tuexen <tuexen@fh-muenster.de>*
7731 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7736 * Add "missing" function EVP_MD_flags() (without this the only way to
7737 retrieve a digest flags is by accessing the structure directly. Update
7738 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7739 or cipher is registered as in the "from" argument. Print out all
7740 registered digests in the dgst usage message instead of manually
7741 attempting to work them out.
7745 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7746 this allows the use of compression and extensions. Change default cipher
7747 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7748 by default unless an application cipher string requests it.
7752 * Alter match criteria in PKCS12_parse(). It used to try to use local
7753 key ids to find matching certificates and keys but some PKCS#12 files
7754 don't follow the (somewhat unwritten) rules and this strategy fails.
7755 Now just gather all certificates together and the first private key
7756 then look for the first certificate that matches the key.
7760 * Support use of registered digest and cipher names for dgst and cipher
7761 commands instead of having to add each one as a special case. So now
7768 openssl dgst -sha256 foo
7770 and this works for ENGINE based algorithms too.
7774 * Update Gost ENGINE to support parameter files.
7776 *Victor B. Wagner <vitus@cryptocom.ru>*
7778 * Support GeneralizedTime in ca utility.
7780 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7782 * Enhance the hash format used for certificate directory links. The new
7783 form uses the canonical encoding (meaning equivalent names will work
7784 even if they aren't identical) and uses SHA1 instead of MD5. This form
7785 is incompatible with the older format and as a result c_rehash should
7786 be used to rebuild symbolic links.
7790 * Make PKCS#8 the default write format for private keys, replacing the
7791 traditional format. This form is standardised, more secure and doesn't
7792 include an implicit MD5 dependency.
7796 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7797 committed to OpenSSL should pass this lot as a minimum.
7801 * Add session ticket override functionality for use by EAP-FAST.
7803 *Jouni Malinen <j@w1.fi>*
7805 * Modify HMAC functions to return a value. Since these can be implemented
7806 in an ENGINE errors can occur.
7810 * Type-checked OBJ_bsearch_ex.
7814 * Type-checked OBJ_bsearch. Also some constification necessitated
7815 by type-checking. Still to come: TXT_DB, bsearch(?),
7816 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7821 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7822 seconds to a tm structure directly, instead of going through OS
7823 specific date routines. This avoids any issues with OS routines such
7824 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7825 and X509_time_adj_ex() to cover the extended range. The existing
7826 X509_time_adj() is still usable and will no longer have any date issues.
7830 * Delta CRL support. New use deltas option which will attempt to locate
7831 and search any appropriate delta CRLs available.
7833 This work was sponsored by Google.
7837 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7838 code and add additional score elements. Validate alternate CRL paths
7839 as part of the CRL checking and indicate a new error "CRL path validation
7840 error" in this case. Applications wanting additional details can use
7841 the verify callback and check the new "parent" field. If this is not
7842 NULL CRL path validation is taking place. Existing applications won't
7843 see this because it requires extended CRL support which is off by
7846 This work was sponsored by Google.
7850 * Support for freshest CRL extension.
7852 This work was sponsored by Google.
7856 * Initial indirect CRL support. Currently only supported in the CRLs
7857 passed directly and not via lookup. Process certificate issuer
7858 CRL entry extension and lookup CRL entries by bother issuer name
7859 and serial number. Check and process CRL issuer entry in IDP extension.
7861 This work was sponsored by Google.
7865 * Add support for distinct certificate and CRL paths. The CRL issuer
7866 certificate is validated separately in this case. Only enabled if
7867 an extended CRL support flag is set: this flag will enable additional
7868 CRL functionality in future.
7870 This work was sponsored by Google.
7874 * Add support for policy mappings extension.
7876 This work was sponsored by Google.
7880 * Fixes to pathlength constraint, self issued certificate handling,
7881 policy processing to align with RFC3280 and PKITS tests.
7883 This work was sponsored by Google.
7887 * Support for name constraints certificate extension. DN, email, DNS
7888 and URI types are currently supported.
7890 This work was sponsored by Google.
7894 * To cater for systems that provide a pointer-based thread ID rather
7895 than numeric, deprecate the current numeric thread ID mechanism and
7896 replace it with a structure and associated callback type. This
7897 mechanism allows a numeric "hash" to be extracted from a thread ID in
7898 either case, and on platforms where pointers are larger than 'long',
7899 mixing is done to help ensure the numeric 'hash' is usable even if it
7900 can't be guaranteed unique. The default mechanism is to use "&errno"
7901 as a pointer-based thread ID to distinguish between threads.
7903 Applications that want to provide their own thread IDs should now use
7904 CRYPTO_THREADID_set_callback() to register a callback that will call
7905 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7907 Note that ERR_remove_state() is now deprecated, because it is tied
7908 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7909 to free the current thread's error state should be replaced by
7910 ERR_remove_thread_state(NULL).
7912 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7913 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7914 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7915 application was previously providing a numeric thread callback that
7916 was inappropriate for distinguishing threads, then uniqueness might
7917 have been obtained with &errno that happened immediately in the
7918 intermediate development versions of OpenSSL; this is no longer the
7919 case, the numeric thread callback will now override the automatic use
7922 *Geoff Thorpe, with help from Bodo Moeller*
7924 * Initial support for different CRL issuing certificates. This covers a
7925 simple case where the self issued certificates in the chain exist and
7926 the real CRL issuer is higher in the existing chain.
7928 This work was sponsored by Google.
7932 * Removed effectively defunct crypto/store from the build.
7936 * Revamp of STACK to provide stronger type-checking. Still to come:
7937 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7938 ASN1_STRING, CONF_VALUE.
7942 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7943 RAM on SSL connections. This option can save about 34k per idle SSL.
7947 * Revamp of LHASH to provide stronger type-checking. Still to come:
7948 STACK, TXT_DB, bsearch, qsort.
7952 * Initial support for Cryptographic Message Syntax (aka CMS) based
7953 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7954 support for data, signedData, compressedData, digestedData and
7955 encryptedData, envelopedData types included. Scripts to check against
7956 RFC4134 examples draft and interop and consistency checks of many
7957 content types and variants.
7961 * Add options to enc utility to support use of zlib compression BIO.
7965 * Extend mk1mf to support importing of options and assembly language
7966 files from Configure script, currently only included in VC-WIN32.
7967 The assembly language rules can now optionally generate the source
7968 files from the associated perl scripts.
7972 * Implement remaining functionality needed to support GOST ciphersuites.
7973 Interop testing has been performed using CryptoPro implementations.
7975 *Victor B. Wagner <vitus@cryptocom.ru>*
7977 * s390x assembler pack.
7981 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7986 * Implement Opaque PRF Input TLS extension as specified in
7987 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7988 official specification yet and no extension type assignment by
7989 IANA exists, this extension (for now) will have to be explicitly
7990 enabled when building OpenSSL by providing the extension number
7991 to use. For example, specify an option
7993 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7995 to the "config" or "Configure" script to enable the extension,
7996 assuming extension number 0x9527 (which is a completely arbitrary
7997 and unofficial assignment based on the MD5 hash of the Internet
7998 Draft). Note that by doing so, you potentially lose
7999 interoperability with other TLS implementations since these might
8000 be using the same extension number for other purposes.
8002 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8003 opaque PRF input value to use in the handshake. This will create
8004 an internal copy of the length-'len' string at 'src', and will
8005 return non-zero for success.
8007 To get more control and flexibility, provide a callback function
8010 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8011 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8015 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8018 Callback function 'cb' will be called in handshakes, and is
8019 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8020 Argument 'arg' is for application purposes (the value as given to
8021 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8022 be provided to the callback function). The callback function
8023 has to return non-zero to report success: usually 1 to use opaque
8024 PRF input just if possible, or 2 to enforce use of the opaque PRF
8025 input. In the latter case, the library will abort the handshake
8026 if opaque PRF input is not successfully negotiated.
8028 Arguments 'peerinput' and 'len' given to the callback function
8029 will always be NULL and 0 in the case of a client. A server will
8030 see the client's opaque PRF input through these variables if
8031 available (NULL and 0 otherwise). Note that if the server
8032 provides an opaque PRF input, the length must be the same as the
8033 length of the client's opaque PRF input.
8035 Note that the callback function will only be called when creating
8036 a new session (session resumption can resume whatever was
8037 previously negotiated), and will not be called in SSL 2.0
8038 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8039 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8040 for applications that need to enforce opaque PRF input.
8044 * Update ssl code to support digests other than SHA1+MD5 for handshake
8047 *Victor B. Wagner <vitus@cryptocom.ru>*
8049 * Add RFC4507 support to OpenSSL. This includes the corrections in
8050 RFC4507bis. The encrypted ticket format is an encrypted encoded
8051 SSL_SESSION structure, that way new session features are automatically
8054 If a client application caches session in an SSL_SESSION structure
8055 support is transparent because tickets are now stored in the encoded
8058 The SSL_CTX structure automatically generates keys for ticket
8059 protection in servers so again support should be possible
8060 with no application modification.
8062 If a client or server wishes to disable RFC4507 support then the option
8063 SSL_OP_NO_TICKET can be set.
8065 Add a TLS extension debugging callback to allow the contents of any client
8066 or server extensions to be examined.
8068 This work was sponsored by Google.
8072 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8073 OpenSSL should now compile cleanly on gcc 4.2
8075 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8077 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8078 support including streaming MAC support: this is required for GOST
8079 ciphersuite support.
8081 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8083 * Add option -stream to use PKCS#7 streaming in smime utility. New
8084 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8085 to output in BER and PEM format.
8089 * Experimental support for use of HMAC via EVP_PKEY interface. This
8090 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8091 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8092 ENGINE support for HMAC keys which are unextractable. New -mac and
8093 -macopt options to dgst utility.
8097 * New option -sigopt to dgst utility. Update dgst to use
8098 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8099 alternative signing parameters such as X9.31 or PSS in the dgst
8104 * Change ssl_cipher_apply_rule(), the internal function that does
8105 the work each time a ciphersuite string requests enabling
8106 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8107 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8108 the order of disabled ciphersuites such that those ciphersuites
8109 that most recently went from enabled to disabled not only stay
8110 in order with respect to each other, but also have higher priority
8111 than other disabled ciphersuites the next time ciphersuites are
8114 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8115 the same ciphersuites as with "HIGH" alone, but in a specific
8116 order where the PSK ciphersuites come first (since they are the
8117 most recently disabled ciphersuites when "HIGH" is parsed).
8119 Also, change ssl_create_cipher_list() (using this new
8120 functionality) such that between otherwise identical
8121 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8126 * Change ssl_create_cipher_list() so that it automatically
8127 arranges the ciphersuites in reasonable order before starting
8128 to process the rule string. Thus, the definition for "DEFAULT"
8129 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8130 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8131 This makes it much easier to arrive at a reasonable default order
8132 in applications for which anonymous ciphers are OK (meaning
8133 that you can't actually use DEFAULT).
8135 *Bodo Moeller; suggested by Victor Duchovni*
8137 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8138 processing) into multiple integers instead of setting
8139 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8140 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8141 (These masks as well as the individual bit definitions are hidden
8142 away into the non-exported interface ssl/ssl_locl.h, so this
8143 change to the definition of the SSL_CIPHER structure shouldn't
8144 affect applications.) This give us more bits for each of these
8145 categories, so there is no longer a need to coagulate AES128 and
8146 AES256 into a single algorithm bit, and to coagulate Camellia128
8147 and Camellia256 into a single algorithm bit, which has led to all
8150 Thus, among other things, the kludge introduced in 0.9.7m and
8151 0.9.8e for masking out AES256 independently of AES128 or masking
8152 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8154 With the change, we also introduce new ciphersuite aliases that
8155 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8160 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8161 Use the leftmost N bytes of the signature input if the input is
8162 larger than the prime q (with N being the size in bytes of q).
8166 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8167 it yet and it is largely untested.
8171 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8175 * Initial incomplete changes to avoid need for function casts in OpenSSL
8176 some compilers (gcc 4.2 and later) reject their use. Safestack is
8177 reimplemented. Update ASN1 to avoid use of legacy functions.
8181 * Win32/64 targets are linked with Winsock2.
8185 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8186 to external functions. This can be used to increase CRL handling
8187 efficiency especially when CRLs are very large by (for example) storing
8188 the CRL revoked certificates in a database.
8192 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8193 new CRLs added to a directory can be used. New command line option
8194 -verify_return_error to s_client and s_server. This causes real errors
8195 to be returned by the verify callback instead of carrying on no matter
8196 what. This reflects the way a "real world" verify callback would behave.
8200 * GOST engine, supporting several GOST algorithms and public key formats.
8201 Kindly donated by Cryptocom.
8205 * Partial support for Issuing Distribution Point CRL extension. CRLs
8206 partitioned by DP are handled but no indirect CRL or reason partitioning
8207 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8208 selected via a scoring technique which handles IDP and AKID in CRLs.
8212 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8213 will ultimately be used for all verify operations: this will remove the
8214 X509_STORE dependency on certificate verification and allow alternative
8215 lookup methods. X509_STORE based implementations of these two callbacks.
8219 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8220 Modify get_crl() to find a valid (unexpired) CRL if possible.
8224 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8225 this would be called X509_CRL_cmp() but that name is already used by
8226 a function that just compares CRL issuer names. Cache several CRL
8227 extensions in X509_CRL structure and cache CRLDP in X509.
8231 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8232 this maps equivalent X509_NAME structures into a consistent structure.
8233 Name comparison can then be performed rapidly using memcmp().
8237 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8242 * Allow digests to supply their own micalg string for S/MIME type using
8243 the ctrl EVP_MD_CTRL_MICALG.
8247 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8248 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8249 ctrl. It can then customise the structure before and/or after signing
8254 * New function OBJ_add_sigid() to allow application defined signature OIDs
8255 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8256 to free up any added signature OIDs.
8260 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8261 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8262 digest and cipher tables. New options added to openssl utility:
8263 list-message-digest-algorithms and list-cipher-algorithms.
8267 * Change the array representation of binary polynomials: the list
8268 of degrees of non-zero coefficients is now terminated with -1.
8269 Previously it was terminated with 0, which was also part of the
8270 value; thus, the array representation was not applicable to
8271 polynomials where t^0 has coefficient zero. This change makes
8272 the array representation useful in a more general context.
8276 * Various modifications and fixes to SSL/TLS cipher string
8277 handling. For ECC, the code now distinguishes between fixed ECDH
8278 with RSA certificates on the one hand and with ECDSA certificates
8279 on the other hand, since these are separate ciphersuites. The
8280 unused code for Fortezza ciphersuites has been removed.
8282 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8283 (not "ECDHE"). For consistency with the code for DH
8284 certificates, use of ECDH certificates is now considered ECDH
8285 authentication, not RSA or ECDSA authentication (the latter is
8286 merely the CA's signing algorithm and not actively used in the
8289 The temporary ciphersuite alias "ECCdraft" is no longer
8290 available, and ECC ciphersuites are no longer excluded from "ALL"
8291 and "DEFAULT". The following aliases now exist for RFC 4492
8292 ciphersuites, most of these by analogy with the DH case:
8294 kECDHr - ECDH cert, signed with RSA
8295 kECDHe - ECDH cert, signed with ECDSA
8296 kECDH - ECDH cert (signed with either RSA or ECDSA)
8297 kEECDH - ephemeral ECDH
8298 ECDH - ECDH cert or ephemeral ECDH
8304 AECDH - anonymous ECDH
8305 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8309 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8310 Use correct micalg parameters depending on digest(s) in signed message.
8314 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8315 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8319 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8320 an engine to register a method. Add ENGINE lookups for methods and
8321 functional reference processing.
8325 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8326 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8331 * New -resign option to smime utility. This adds one or more signers
8332 to an existing PKCS#7 signedData structure. Also -md option to use an
8333 alternative message digest algorithm for signing.
8337 * Tidy up PKCS#7 routines and add new functions to make it easier to
8338 create PKCS7 structures containing multiple signers. Update smime
8339 application to support multiple signers.
8343 * New -macalg option to pkcs12 utility to allow setting of an alternative
8348 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8349 Reorganize PBE internals to lookup from a static table using NIDs,
8350 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8351 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8352 PRF which will be automatically used with PBES2.
8356 * Replace the algorithm specific calls to generate keys in "req" with the
8361 * Update PKCS#7 enveloped data routines to use new API. This is now
8362 supported by any public key method supporting the encrypt operation. A
8363 ctrl is added to allow the public key algorithm to examine or modify
8364 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8369 * Add a ctrl to asn1 method to allow a public key algorithm to express
8370 a default digest type to use. In most cases this will be SHA1 but some
8371 algorithms (such as GOST) need to specify an alternative digest. The
8372 return value indicates how strong the preference is 1 means optional and
8373 2 is mandatory (that is it is the only supported type). Modify
8374 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8375 use the default md. Update openssl utilities to use the default digest
8376 type for signing if it is not explicitly indicated.
8380 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8381 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8382 signing method from the key type. This effectively removes the link
8383 between digests and public key types.
8387 * Add an OID cross reference table and utility functions. Its purpose is to
8388 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8389 rsaEncryption. This will allow some of the algorithm specific hackery
8390 needed to use the correct OID to be removed.
8394 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8395 structures for PKCS7_sign(). They are now set up by the relevant public
8400 * Add provisional EC pkey method with support for ECDSA and ECDH.
8404 * Add support for key derivation (agreement) in the API, DH method and
8409 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8410 public and private key formats. As a side effect these add additional
8411 command line functionality not previously available: DSA signatures can be
8412 generated and verified using pkeyutl and DH key support and generation in
8419 *Oliver Tappe <zooey@hirschkaefer.de>*
8421 * New make target "install_html_docs" installs HTML renditions of the
8424 *Oliver Tappe <zooey@hirschkaefer.de>*
8426 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8427 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8428 support key and parameter generation and add initial key generation
8429 functionality for RSA.
8433 * Add functions for main EVP_PKEY_method operations. The undocumented
8434 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8435 `EVP_PKEY_{encrypt,decrypt}_old`.
8439 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8440 key API, doesn't do much yet.
8444 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8445 public key algorithms. New option to openssl utility:
8446 "list-public-key-algorithms" to print out info.
8450 * Implement the Supported Elliptic Curves Extension for
8451 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8455 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8456 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8460 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8461 utilities such as rsa, dsa, dsaparam etc except they process any key
8466 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8467 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8468 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8473 * Initial support for pluggable public key ASN1.
8474 De-spaghettify the public key ASN1 handling. Move public and private
8475 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8476 algorithm specific handling to a single module within the relevant
8477 algorithm directory. Add functions to allow (near) opaque processing
8478 of public and private key structures.
8482 * Implement the Supported Point Formats Extension for
8483 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8487 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8488 for the psk identity [hint] and the psk callback functions to the
8489 SSL_SESSION, SSL and SSL_CTX structure.
8492 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8496 SSL_CTX_use_psk_identity_hint
8497 SSL_get_psk_identity_hint
8498 SSL_get_psk_identity
8499 SSL_use_psk_identity_hint
8501 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8503 * Add RFC 3161 compliant time stamp request creation, response generation
8504 and response verification functionality.
8506 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8508 * Add initial support for TLS extensions, specifically for the server_name
8509 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8510 have new members for a host name. The SSL data structure has an
8511 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8512 stored in that context to allow for session resumption, even after the
8513 SSL has been switched to a new SSL_CTX in reaction to a client's
8514 server_name extension.
8516 New functions (subject to change):
8518 SSL_get_servername()
8519 SSL_get_servername_type()
8522 New CTRL codes and macros (subject to change):
8524 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8525 - SSL_CTX_set_tlsext_servername_callback()
8526 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8527 - SSL_CTX_set_tlsext_servername_arg()
8528 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8530 openssl s_client has a new '-servername ...' option.
8532 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8533 '-key2 ...', '-servername_fatal' (subject to change). This allows
8534 testing the HostName extension for a specific single host name ('-cert'
8535 and '-key' remain fallbacks for handshakes without HostName
8536 negotiation). If the unrecognized_name alert has to be sent, this by
8537 default is a warning; it becomes fatal with the '-servername_fatal'
8540 *Peter Sylvester, Remy Allais, Christophe Renou*
8542 * Whirlpool hash implementation is added.
8546 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8547 bn(64,32). Because of instruction set limitations it doesn't have
8548 any negative impact on performance. This was done mostly in order
8549 to make it possible to share assembler modules, such as bn_mul_mont
8550 implementations, between 32- and 64-bit builds without hassle.
8554 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8555 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8560 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8561 dedicated Montgomery multiplication procedure, is introduced.
8562 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8563 "64-bit" performance on certain 32-bit targets.
8567 * New option SSL_OP_NO_COMP to disable use of compression selectively
8568 in SSL structures. New SSL ctrl to set maximum send fragment size.
8569 Save memory by setting the I/O buffer sizes dynamically instead of
8570 using the maximum available value.
8574 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8575 in addition to the text details.
8579 * Very, very preliminary EXPERIMENTAL support for printing of general
8580 ASN1 structures. This currently produces rather ugly output and doesn't
8581 handle several customised structures at all.
8585 * Integrated support for PVK file format and some related formats such
8586 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8587 these in the 'rsa' and 'dsa' utilities.
8591 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8595 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8596 place for the (very old) "NETSCAPE" format certificates which are now
8597 handled using new ASN1 code equivalents.
8601 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8602 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8603 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8607 * Modify CRL distribution points extension code to print out previously
8608 unsupported fields. Enhance extension setting code to allow setting of
8613 * Add print and set support for Issuing Distribution Point CRL extension.
8617 * Change 'Configure' script to enable Camellia by default.
8624 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8626 * When rejecting SSL/TLS records due to an incorrect version number, never
8627 update s->server with a new major version number. As of
8628 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8629 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8630 the previous behavior could result in a read attempt at NULL when
8631 receiving specific incorrect SSL/TLS records once record payload
8632 protection is active. [CVE-2010-0740][]
8634 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8636 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8637 could be crashed if the relevant tables were not present (e.g. chrooted).
8639 *Tomas Hoger <thoger@redhat.com>*
8641 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8643 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8645 *Martin Olsson, Neel Mehta*
8647 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8648 accommodate for stack sorting, always a write lock!).
8652 * On some versions of WIN32 Heap32Next is very slow. This can cause
8653 excessive delays in the RAND_poll(): over a minute. As a workaround
8654 include a time check in the inner Heap32Next loop too.
8658 * The code that handled flushing of data in SSL/TLS originally used the
8659 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8660 the problem outlined in PR#1949. The fix suggested there however can
8661 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8662 of Apache). So instead simplify the code to flush unconditionally.
8663 This should be fine since flushing with no data to flush is a no op.
8667 * Handle TLS versions 2.0 and later properly and correctly use the
8668 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8669 off ancient servers have a habit of sticking around for a while...
8673 * Modify compression code so it frees up structures without using the
8674 ex_data callbacks. This works around a problem where some applications
8675 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8676 restarting) then use compression (e.g. SSL with compression) later.
8677 This results in significant per-connection memory leaks and
8678 has caused some security issues including CVE-2008-1678 and
8683 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8684 change when encrypting or decrypting.
8688 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8689 connect and renegotiate with servers which do not support RI.
8690 Until RI is more widely deployed this option is enabled by default.
8694 * Add "missing" ssl ctrls to clear options and mode.
8698 * If client attempts to renegotiate and doesn't support RI respond with
8699 a no_renegotiation alert as required by RFC5746. Some renegotiating
8700 TLS clients will continue a connection gracefully when they receive
8701 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8702 waiting for a server hello which it will never receive. Now we treat a
8703 received no_renegotiation alert as a fatal error. This is because
8704 applications requesting a renegotiation might well expect it to succeed
8705 and would have no code in place to handle the server denying it so the
8706 only safe thing to do is to terminate the connection.
8710 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8711 peer supports secure renegotiation and 0 otherwise. Print out peer
8712 renegotiation support in s_client/s_server.
8716 * Replace the highly broken and deprecated SPKAC certification method with
8717 the updated NID creation version. This should correctly handle UTF8.
8721 * Implement RFC5746. Re-enable renegotiation but require the extension
8722 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8723 turns out to be a bad idea. It has been replaced by
8724 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8725 SSL_CTX_set_options(). This is really not recommended unless you
8726 know what you are doing.
8728 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8730 * Fixes to stateless session resumption handling. Use initial_ctx when
8731 issuing and attempting to decrypt tickets in case it has changed during
8732 servername handling. Use a non-zero length session ID when attempting
8733 stateless session resumption: this makes it possible to determine if
8734 a resumption has occurred immediately after receiving server hello
8735 (several places in OpenSSL subtly assume this) instead of later in
8740 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8741 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8742 fixes for a few places where the return code is not checked
8745 *Julia Lawall <julia@diku.dk>*
8747 * Add --strict-warnings option to Configure script to include devteam
8748 warnings in other configurations.
8752 * Add support for --libdir option and LIBDIR variable in makefiles. This
8753 makes it possible to install openssl libraries in locations which
8754 have names other than "lib", for example "/usr/lib64" which some
8757 *Steve Henson, based on patch from Jeremy Utley*
8759 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8760 X690 8.9.12 and can produce some misleading textual output of OIDs.
8762 *Steve Henson, reported by Dan Kaminsky*
8764 * Delete MD2 from algorithm tables. This follows the recommendation in
8765 several standards that it is not used in new applications due to
8766 several cryptographic weaknesses. For binary compatibility reasons
8767 the MD2 API is still compiled in by default.
8771 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8776 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8777 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8780 *Guenter <lists@gknw.net>*
8782 * Fix the server certificate chain building code to use X509_verify_cert(),
8783 it used to have an ad-hoc builder which was unable to cope with anything
8784 other than a simple chain.
8786 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8788 * Don't check self signed certificate signatures in X509_verify_cert()
8789 by default (a flag can override this): it just wastes time without
8790 adding any security. As a useful side effect self signed root CAs
8791 with non-FIPS digests are now usable in FIPS mode.
8795 * In dtls1_process_out_of_seq_message() the check if the current message
8796 is already buffered was missing. For every new message was memory
8797 allocated, allowing an attacker to perform an denial of service attack
8798 with sending out of seq handshake messages until there is no memory
8799 left. Additionally every future message was buffered, even if the
8800 sequence number made no sense and would be part of another handshake.
8801 So only messages with sequence numbers less than 10 in advance will be
8802 buffered. [CVE-2009-1378][]
8804 *Robin Seggelmann, discovered by Daniel Mentz*
8806 * Records are buffered if they arrive with a future epoch to be
8807 processed after finishing the corresponding handshake. There is
8808 currently no limitation to this buffer allowing an attacker to perform
8809 a DOS attack with sending records with future epochs until there is no
8810 memory left. This patch adds the pqueue_size() function to determine
8811 the size of a buffer and limits the record buffer to 100 entries.
8814 *Robin Seggelmann, discovered by Daniel Mentz*
8816 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8817 parent structure is freed. [CVE-2009-1379][]
8821 * Handle non-blocking I/O properly in SSL_shutdown() call.
8823 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8825 * Add `2.5.4.*` OIDs
8827 *Ilya O. <vrghost@gmail.com>*
8829 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8831 * Disable renegotiation completely - this fixes a severe security
8832 problem [CVE-2009-3555][] at the cost of breaking all
8833 renegotiation. Renegotiation can be re-enabled by setting
8834 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8835 run-time. This is really not recommended unless you know what
8840 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8842 * Don't set val to NULL when freeing up structures, it is freed up by
8843 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8844 zeroing past the valid field. [CVE-2009-0789][]
8846 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8848 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8849 checked correctly. This would allow some invalid signed attributes to
8850 appear to verify correctly. [CVE-2009-0591][]
8852 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8854 * Reject UniversalString and BMPString types with invalid lengths. This
8855 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8856 a legal length. [CVE-2009-0590][]
8860 * Set S/MIME signing as the default purpose rather than setting it
8861 unconditionally. This allows applications to override it at the store
8866 * Permit restricted recursion of ASN1 strings. This is needed in practice
8867 to handle some structures.
8871 * Improve efficiency of mem_gets: don't search whole buffer each time
8874 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8876 * New -hex option for openssl rand.
8880 * Print out UTF8String and NumericString when parsing ASN1.
8884 * Support NumericString type for name components.
8888 * Allow CC in the environment to override the automatically chosen
8889 compiler. Note that nothing is done to ensure flags work with the
8894 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8896 * Properly check EVP_VerifyFinal() and similar return values
8899 *Ben Laurie, Bodo Moeller, Google Security Team*
8901 * Enable TLS extensions by default.
8905 * Allow the CHIL engine to be loaded, whether the application is
8906 multithreaded or not. (This does not release the developer from the
8907 obligation to set up the dynamic locking callbacks.)
8909 *Sander Temme <sander@temme.net>*
8911 * Use correct exit code if there is an error in dgst command.
8913 *Steve Henson; problem pointed out by Roland Dirlewanger*
8915 * Tweak Configure so that you need to say "experimental-jpake" to enable
8916 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8920 * Add experimental JPAKE support, including demo authentication in
8921 s_client and s_server.
8925 * Set the comparison function in v3_addr_canonize().
8927 *Rob Austein <sra@hactrn.net>*
8929 * Add support for XMPP STARTTLS in s_client.
8931 *Philip Paeps <philip@freebsd.org>*
8933 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8934 to ensure that even with this option, only ciphersuites in the
8935 server's preference list will be accepted. (Note that the option
8936 applies only when resuming a session, so the earlier behavior was
8937 just about the algorithm choice for symmetric cryptography.)
8941 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8943 * Fix NULL pointer dereference if a DTLS server received
8944 ChangeCipherSpec as first record [CVE-2009-1386][].
8948 * Fix a state transition in s3_srvr.c and d1_srvr.c
8949 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8953 * The fix in 0.9.8c that supposedly got rid of unsafe
8954 double-checked locking was incomplete for RSA blinding,
8955 addressing just one layer of what turns out to have been
8956 doubly unsafe triple-checked locking.
8958 So now fix this for real by retiring the MONT_HELPER macro
8959 in crypto/rsa/rsa_eay.c.
8961 *Bodo Moeller; problem pointed out by Marius Schilder*
8963 * Various precautionary measures:
8965 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8967 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8968 (NB: This would require knowledge of the secret session ticket key
8969 to exploit, in which case you'd be SOL either way.)
8971 - Change bn_nist.c so that it will properly handle input BIGNUMs
8972 outside the expected range.
8974 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8977 *Neel Mehta, Bodo Moeller*
8979 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8980 the load fails. Useful for distros.
8982 *Ben Laurie and the FreeBSD team*
8984 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8988 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8992 * Expand ENGINE to support engine supplied SSL client certificate functions.
8994 This work was sponsored by Logica.
8998 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8999 keystores. Support for SSL/TLS client authentication too.
9000 Not compiled unless enable-capieng specified to Configure.
9002 This work was sponsored by Logica.
9006 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9007 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9008 attribute creation routines such as certificate requests and PKCS#12
9013 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9015 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9016 handshake which could lead to a client crash as found using the
9017 Codenomicon TLS test suite [CVE-2008-1672][]
9019 *Steve Henson, Mark Cox*
9021 * Fix double free in TLS server name extensions which could lead to
9022 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
9026 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9028 Clear the error queue to ensure that error entries left from
9029 older function calls do not interfere with the correct operation.
9031 *Lutz Jaenicke, Erik de Castro Lopo*
9033 * Remove root CA certificates of commercial CAs:
9035 The OpenSSL project does not recommend any specific CA and does not
9036 have any policy with respect to including or excluding any CA.
9037 Therefore it does not make any sense to ship an arbitrary selection
9038 of root CA certificates with the OpenSSL software.
9042 * RSA OAEP patches to fix two separate invalid memory reads.
9043 The first one involves inputs when 'lzero' is greater than
9044 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9045 before the beginning of from). The second one involves inputs where
9046 the 'db' section contains nothing but zeroes (there is a one-byte
9047 invalid read after the end of 'db').
9049 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9051 * Partial backport from 0.9.9-dev:
9053 Introduce bn_mul_mont (dedicated Montgomery multiplication
9054 procedure) as a candidate for BIGNUM assembler implementation.
9055 While 0.9.9-dev uses assembler for various architectures, only
9056 x86_64 is available by default here in the 0.9.8 branch, and
9057 32-bit x86 is available through a compile-time setting.
9059 To try the 32-bit x86 assembler implementation, use Configure
9060 option "enable-montasm" (which exists only for this backport).
9062 As "enable-montasm" for 32-bit x86 disclaims code stability
9063 anyway, in this constellation we activate additional code
9064 backported from 0.9.9-dev for further performance improvements,
9065 namely BN_from_montgomery_word. (To enable this otherwise,
9066 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9068 *Andy Polyakov (backport partially by Bodo Moeller)*
9070 * Add TLS session ticket callback. This allows an application to set
9071 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9072 values. This is useful for key rollover for example where several key
9073 sets may exist with different names.
9077 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9078 This was broken until now in 0.9.8 releases, such that the only way
9079 a registered ENGINE could be used (assuming it initialises
9080 successfully on the host) was to explicitly set it as the default
9081 for the relevant algorithms. This is in contradiction with 0.9.7
9082 behaviour and the documentation. With this fix, when an ENGINE is
9083 registered into a given algorithm's table of implementations, the
9084 'uptodate' flag is reset so that auto-discovery will be used next
9085 time a new context for that algorithm attempts to select an
9088 *Ian Lister (tweaked by Geoff Thorpe)*
9090 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9091 implementation in the following ways:
9093 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9096 Lack of BER streaming support means one pass streaming processing is
9097 only supported if data is detached: setting the streaming flag is
9098 ignored for embedded content.
9100 CMS support is disabled by default and must be explicitly enabled
9101 with the enable-cms configuration option.
9105 * Update the GMP engine glue to do direct copies between BIGNUM and
9106 mpz_t when openssl and GMP use the same limb size. Otherwise the
9107 existing "conversion via a text string export" trick is still used.
9109 *Paul Sheer <paulsheer@gmail.com>*
9111 * Zlib compression BIO. This is a filter BIO which compressed and
9112 uncompresses any data passed through it.
9116 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9117 RFC3394 compatible AES key wrapping.
9121 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9122 sets string data without copying. X509_ALGOR_set0() and
9123 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9124 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9125 from an X509_ATTRIBUTE structure optionally checking it occurs only
9126 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9131 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9132 to get the expected BN_FLG_CONSTTIME behavior.
9134 *Bodo Moeller (Google)*
9138 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9139 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9140 - added some more tests to do_tests.pl
9141 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9142 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9143 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9144 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9145 - various changes to netware.pl to enable gcc-cross builds on Win32
9147 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9148 - various changes to fix missing prototype warnings
9149 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9150 - added AES, WHIRLPOOL and CPUID assembler code to build files
9151 - added missing AES assembler make rules to mk1mf.pl
9152 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9154 *Guenter Knauf <eflash@gmx.net>*
9156 * Implement certificate status request TLS extension defined in RFC3546.
9157 A client can set the appropriate parameters and receive the encoded
9158 OCSP response via a callback. A server can query the supplied parameters
9159 and set the encoded OCSP response in the callback. Add simplified examples
9160 to s_client and s_server.
9164 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9167 + Binary incompatibility of ssl_ctx_st structure
9168 + DTLS interoperation with non-compliant servers
9169 + Don't call get_session_cb() without proposed session
9170 + Fix ia64 assembler code
9172 *Andy Polyakov, Steve Henson*
9174 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9176 * DTLS Handshake overhaul. There were longstanding issues with
9177 OpenSSL DTLS implementation, which were making it impossible for
9178 RFC 4347 compliant client to communicate with OpenSSL server.
9179 Unfortunately just fixing these incompatibilities would "cut off"
9180 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9181 server keeps tolerating non RFC compliant syntax. The opposite is
9182 not true, 0.9.8f client can not communicate with earlier server.
9183 This update even addresses CVE-2007-4995.
9187 * Changes to avoid need for function casts in OpenSSL: some compilers
9188 (gcc 4.2 and later) reject their use.
9189 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9192 * Add RFC4507 support to OpenSSL. This includes the corrections in
9193 RFC4507bis. The encrypted ticket format is an encrypted encoded
9194 SSL_SESSION structure, that way new session features are automatically
9197 If a client application caches session in an SSL_SESSION structure
9198 support is transparent because tickets are now stored in the encoded
9201 The SSL_CTX structure automatically generates keys for ticket
9202 protection in servers so again support should be possible
9203 with no application modification.
9205 If a client or server wishes to disable RFC4507 support then the option
9206 SSL_OP_NO_TICKET can be set.
9208 Add a TLS extension debugging callback to allow the contents of any client
9209 or server extensions to be examined.
9211 This work was sponsored by Google.
9215 * Add initial support for TLS extensions, specifically for the server_name
9216 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9217 have new members for a host name. The SSL data structure has an
9218 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9219 stored in that context to allow for session resumption, even after the
9220 SSL has been switched to a new SSL_CTX in reaction to a client's
9221 server_name extension.
9223 New functions (subject to change):
9225 SSL_get_servername()
9226 SSL_get_servername_type()
9229 New CTRL codes and macros (subject to change):
9231 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9232 - SSL_CTX_set_tlsext_servername_callback()
9233 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9234 - SSL_CTX_set_tlsext_servername_arg()
9235 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9237 openssl s_client has a new '-servername ...' option.
9239 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9240 '-key2 ...', '-servername_fatal' (subject to change). This allows
9241 testing the HostName extension for a specific single host name ('-cert'
9242 and '-key' remain fallbacks for handshakes without HostName
9243 negotiation). If the unrecognized_name alert has to be sent, this by
9244 default is a warning; it becomes fatal with the '-servername_fatal'
9247 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9249 * Add AES and SSE2 assembly language support to VC++ build.
9253 * Mitigate attack on final subtraction in Montgomery reduction.
9257 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9258 (which previously caused an internal error).
9262 * Squeeze another 10% out of IGE mode when in != out.
9266 * AES IGE mode speedup.
9268 *Dean Gaudet (Google)*
9270 * Add the Korean symmetric 128-bit cipher SEED (see
9271 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9272 add SEED ciphersuites from RFC 4162:
9274 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9275 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9276 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9277 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9279 To minimize changes between patchlevels in the OpenSSL 0.9.8
9280 series, SEED remains excluded from compilation unless OpenSSL
9281 is configured with 'enable-seed'.
9283 *KISA, Bodo Moeller*
9285 * Mitigate branch prediction attacks, which can be practical if a
9286 single processor is shared, allowing a spy process to extract
9287 information. For detailed background information, see
9288 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9289 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9290 and Necessary Software Countermeasures"). The core of the change
9291 are new versions BN_div_no_branch() and
9292 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9293 respectively, which are slower, but avoid the security-relevant
9294 conditional branches. These are automatically called by BN_div()
9295 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9296 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9297 remove a conditional branch.
9299 BN_FLG_CONSTTIME is the new name for the previous
9300 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9301 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9302 in the exponent causes BN_mod_exp_mont() to use the alternative
9303 implementation in BN_mod_exp_mont_consttime().) The old name
9304 remains as a deprecated alias.
9306 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9307 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9308 constant-time implementations for more than just exponentiation.
9309 Here too the old name is kept as a deprecated alias.
9311 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9312 the BN_BLINDING structure gets an independent copy of the
9313 modulus. This means that the previous `BIGNUM *m` argument to
9314 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9315 essentially becomes `const BIGNUM *m`, although we can't actually
9316 change this in the header file before 0.9.9. It allows
9317 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9318 enable BN_FLG_CONSTTIME.
9320 *Matthew D Wood (Intel Corp)*
9322 * In the SSL/TLS server implementation, be strict about session ID
9323 context matching (which matters if an application uses a single
9324 external cache for different purposes). Previously,
9325 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9326 set. This did ensure strict client verification, but meant that,
9327 with applications using a single external cache for quite
9328 different requirements, clients could circumvent ciphersuite
9329 restrictions for a given session ID context by starting a session
9330 in a different context.
9334 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9335 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9336 authentication-only ciphersuites.
9340 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9341 not complete and could lead to a possible single byte overflow
9342 [CVE-2007-5135][] [Ben Laurie]
9344 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9346 * Since AES128 and AES256 (and similarly Camellia128 and
9347 Camellia256) share a single mask bit in the logic of
9348 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9349 kludge to work properly if AES128 is available and AES256 isn't
9350 (or if Camellia128 is available and Camellia256 isn't).
9354 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9355 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9356 When a point or a seed is encoded in a BIT STRING, we need to
9357 prevent the removal of trailing zero bits to get the proper DER
9358 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9359 of a NamedBitList, for which trailing 0 bits need to be removed.)
9363 * Have SSL/TLS server implementation tolerate "mismatched" record
9364 protocol version while receiving ClientHello even if the
9365 ClientHello is fragmented. (The server can't insist on the
9366 particular protocol version it has chosen before the ServerHello
9367 message has informed the client about his choice.)
9371 * Add RFC 3779 support.
9373 *Rob Austein for ARIN, Ben Laurie*
9375 * Load error codes if they are not already present instead of using a
9376 static variable. This allows them to be cleanly unloaded and reloaded.
9377 Improve header file function name parsing.
9381 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9382 or CAPABILITY handshake as required by RFCs.
9386 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9388 * Introduce limits to prevent malicious keys being able to
9389 cause a denial of service. [CVE-2006-2940][]
9391 *Steve Henson, Bodo Moeller*
9393 * Fix ASN.1 parsing of certain invalid structures that can result
9394 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9396 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9397 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9399 * Fix SSL client code which could crash if connecting to a
9400 malicious SSLv2 server. [CVE-2006-4343][]
9402 *Tavis Ormandy and Will Drewry, Google Security Team*
9404 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9405 match only those. Before that, "AES256-SHA" would be interpreted
9406 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9407 the same strength classification in 0.9.7h) as we currently only
9408 have a single AES bit in the ciphersuite description bitmap.
9409 That change, however, also applied to ciphersuite strings such as
9410 "RC4-MD5" that intentionally matched multiple ciphersuites --
9411 namely, SSL 2.0 ciphersuites in addition to the more common ones
9412 from SSL 3.0/TLS 1.0.
9414 So we change the selection algorithm again: Naming an explicit
9415 ciphersuite selects this one ciphersuite, and any other similar
9416 ciphersuite (same bitmap) from *other* protocol versions.
9417 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9418 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9420 Since SSL 2.0 does not have any ciphersuites for which the
9421 128/256 bit distinction would be relevant, this works for now.
9422 The proper fix will be to use different bits for AES128 and
9423 AES256, which would have avoided the problems from the beginning;
9424 however, bits are scarce, so we can only do this in a new release
9425 (not just a patchlevel) when we can change the SSL_CIPHER
9426 definition to split the single 'unsigned long mask' bitmap into
9427 multiple values to extend the available space.
9431 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9433 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9434 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9436 * Add AES IGE and biIGE modes.
9440 * Change the Unix randomness entropy gathering to use poll() when
9441 possible instead of select(), since the latter has some
9442 undesirable limitations.
9444 *Darryl Miles via Richard Levitte and Bodo Moeller*
9446 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9447 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9448 cannot be implicitly activated as part of, e.g., the "AES" alias.
9449 However, please upgrade to OpenSSL 0.9.9[-dev] for
9450 non-experimental use of the ECC ciphersuites to get TLS extension
9451 support, which is required for curve and point format negotiation
9452 to avoid potential handshake problems.
9456 * Disable rogue ciphersuites:
9458 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9459 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9460 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9462 The latter two were purportedly from
9463 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9466 Also deactivate the remaining ciphersuites from
9467 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9468 unofficial, and the ID has long expired.
9472 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9473 dual-core machines) and other potential thread-safety issues.
9477 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9478 versions), which is now available for royalty-free use
9479 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9480 Also, add Camellia TLS ciphersuites from RFC 4132.
9482 To minimize changes between patchlevels in the OpenSSL 0.9.8
9483 series, Camellia remains excluded from compilation unless OpenSSL
9484 is configured with 'enable-camellia'.
9488 * Disable the padding bug check when compression is in use. The padding
9489 bug check assumes the first packet is of even length, this is not
9490 necessarily true if compression is enabled and can result in false
9491 positives causing handshake failure. The actual bug test is ancient
9492 code so it is hoped that implementations will either have fixed it by
9493 now or any which still have the bug do not support compression.
9497 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9499 * When applying a cipher rule check to see if string match is an explicit
9500 cipher suite and only match that one cipher suite if it is.
9504 * Link in manifests for VC++ if needed.
9506 *Austin Ziegler <halostatue@gmail.com>*
9508 * Update support for ECC-based TLS ciphersuites according to
9509 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9510 TLS extensions, which are supported starting with the 0.9.9
9511 branch, not in the OpenSSL 0.9.8 branch).
9515 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9516 opaque EVP_CIPHER_CTX handling.
9520 * Fixes and enhancements to zlib compression code. We now only use
9521 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9522 to conform with the standards mentioned here:
9523 <http://www.zlib.net/DLL_FAQ.txt>
9524 Static zlib linking now works on Windows and the new --with-zlib-include
9525 --with-zlib-lib options to Configure can be used to supply the location
9526 of the headers and library. Gracefully handle case where zlib library
9531 * Several fixes and enhancements to the OID generation code. The old code
9532 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9533 handle numbers larger than ULONG_MAX, truncated printing and had a
9534 non standard OBJ_obj2txt() behaviour.
9538 * Add support for building of engines under engine/ as shared libraries
9539 under VC++ build system.
9543 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9544 Hopefully, we will not see any false combination of paths any more.
9548 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9550 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9551 (part of SSL_OP_ALL). This option used to disable the
9552 countermeasure against man-in-the-middle protocol-version
9553 rollback in the SSL 2.0 server implementation, which is a bad
9554 idea. [CVE-2005-2969][]
9556 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9557 for Information Security, National Institute of Advanced Industrial
9558 Science and Technology [AIST], Japan)*
9560 * Add two function to clear and return the verify parameter flags.
9564 * Keep cipherlists sorted in the source instead of sorting them at
9565 runtime, thus removing the need for a lock.
9569 * Avoid some small subgroup attacks in Diffie-Hellman.
9571 *Nick Mathewson and Ben Laurie*
9573 * Add functions for well-known primes.
9577 * Extended Windows CE support.
9579 *Satoshi Nakamura and Andy Polyakov*
9581 * Initialize SSL_METHOD structures at compile time instead of during
9582 runtime, thus removing the need for a lock.
9586 * Make PKCS7_decrypt() work even if no certificate is supplied by
9587 attempting to decrypt each encrypted key in turn. Add support to
9592 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9594 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9597 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9601 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9602 key into the same file any more.
9606 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9610 * Add -utf8 command line and config file option to 'ca'.
9612 *Stefan <stf@udoma.org*
9614 * Removed the macro des_crypt(), as it seems to conflict with some
9615 libraries. Use DES_crypt().
9619 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9620 involves renaming the source and generated shared-libs for
9621 both. The engines will accept the corrected or legacy ids
9622 ('ncipher' and '4758_cca' respectively) when binding. NB,
9623 this only applies when building 'shared'.
9625 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9627 * Add attribute functions to EVP_PKEY structure. Modify
9628 PKCS12_create() to recognize a CSP name attribute and
9629 use it. Make -CSP option work again in pkcs12 utility.
9633 * Add new functionality to the bn blinding code:
9634 - automatic re-creation of the BN_BLINDING parameters after
9635 a fixed number of uses (currently 32)
9636 - add new function for parameter creation
9637 - introduce flags to control the update behaviour of the
9638 BN_BLINDING parameters
9639 - hide BN_BLINDING structure
9640 Add a second BN_BLINDING slot to the RSA structure to improve
9641 performance when a single RSA object is shared among several
9646 * Add support for DTLS.
9648 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9650 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9651 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9655 * Remove buggy and incomplete DH cert support from
9656 ssl/ssl_rsa.c and ssl/s3_both.c
9660 * Use SHA-1 instead of MD5 as the default digest algorithm for
9661 the apps/openssl applications.
9665 * Compile clean with "-Wall -Wmissing-prototypes
9666 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9667 DEBUG_SAFESTACK must also be set.
9671 * Change ./Configure so that certain algorithms can be disabled by default.
9672 The new counterpiece to "no-xxx" is "enable-xxx".
9674 The patented RC5 and MDC2 algorithms will now be disabled unless
9675 "enable-rc5" and "enable-mdc2", respectively, are specified.
9677 (IDEA remains enabled despite being patented. This is because IDEA
9678 is frequently required for interoperability, and there is no license
9679 fee for non-commercial use. As before, "no-idea" can be used to
9680 avoid this algorithm.)
9684 * Add processing of proxy certificates (see RFC 3820). This work was
9685 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9686 EGEE (Enabling Grids for E-science in Europe).
9690 * RC4 performance overhaul on modern architectures/implementations, such
9691 as Intel P4, IA-64 and AMD64.
9695 * New utility extract-section.pl. This can be used specify an alternative
9696 section number in a pod file instead of having to treat each file as
9697 a separate case in Makefile. This can be done by adding two lines to the
9700 =for comment openssl_section:XXX
9702 The blank line is mandatory.
9706 * New arguments -certform, -keyform and -pass for s_client and s_server
9707 to allow alternative format key and certificate files and passphrase
9712 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9713 update associated structures and add various utility functions.
9715 Add new policy related verify parameters, include policy checking in
9716 standard verify code. Enhance 'smime' application with extra parameters
9717 to support policy checking and print out.
9721 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9722 Nehemiah processors. These extensions support AES encryption in hardware
9723 as well as RNG (though RNG support is currently disabled).
9725 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9727 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9731 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9733 *Andy Polyakov and a number of other people*
9735 * Improved PowerPC platform support. Most notably BIGNUM assembler
9736 implementation contributed by IBM.
9738 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9740 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9741 exponent rather than 'unsigned long'. There is a corresponding change to
9742 the new 'rsa_keygen' element of the RSA_METHOD structure.
9744 *Jelte Jansen, Geoff Thorpe*
9746 * Functionality for creating the initial serial number file is now
9747 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9749 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9750 number file to 1, which is bound to cause problems. To avoid
9751 the problems while respecting compatibility between different 0.9.7
9752 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9753 CA.pl for serial number initialization. With the new release 0.9.8,
9754 we can fix the problem directly in the 'ca' utility.)
9758 * Reduced header interdependencies by declaring more opaque objects in
9759 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9760 give fewer recursive includes, which could break lazy source code - so
9761 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9762 developers should define this symbol when building and using openssl to
9763 ensure they track the recommended behaviour, interfaces, [etc], but
9764 backwards-compatible behaviour prevails when this isn't defined.
9768 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9772 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9773 This will generate a random key of the appropriate length based on the
9774 cipher context. The EVP_CIPHER can provide its own random key generation
9775 routine to support keys of a specific form. This is used in the des and
9776 3des routines to generate a key of the correct parity. Update S/MIME
9777 code to use new functions and hence generate correct parity DES keys.
9778 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9779 valid (weak or incorrect parity).
9783 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9784 as looking them up. This is useful when the verified structure may contain
9785 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9786 present unless the new PKCS7_NO_CRL flag is asserted.
9790 * Extend ASN1 oid configuration module. It now additionally accepts the
9793 shortName = some long name, 1.2.3.4
9797 * Reimplemented the BN_CTX implementation. There is now no more static
9798 limitation on the number of variables it can handle nor the depth of the
9799 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9800 information can now expand as required, and rather than having a single
9801 static array of bignums, BN_CTX now uses a linked-list of such arrays
9802 allowing it to expand on demand whilst maintaining the usefulness of
9803 BN_CTX's "bundling".
9807 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9808 to allow all RSA operations to function using a single BN_CTX.
9812 * Preliminary support for certificate policy evaluation and checking. This
9813 is initially intended to pass the tests outlined in "Conformance Testing
9814 of Relying Party Client Certificate Path Processing Logic" v1.07.
9818 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9819 remained unused and not that useful. A variety of other little bignum
9820 tweaks and fixes have also been made continuing on from the audit (see
9825 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9826 associated ASN1, EVP and SSL functions and old ASN1 macros.
9830 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9831 and this should never fail. So the return value from the use of
9832 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9833 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9837 * BN_CTX_get() should return zero-valued bignums, providing the same
9838 initialised value as BN_new().
9840 *Geoff Thorpe, suggested by Ulf Möller*
9842 * Support for inhibitAnyPolicy certificate extension.
9846 * An audit of the BIGNUM code is underway, for which debugging code is
9847 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9848 is considered valid when processing BIGNUMs, and causes execution to
9849 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9850 further steps are taken to deliberately pollute unused data in BIGNUM
9851 structures to try and expose faulty code further on. For now, openssl will
9852 (in its default mode of operation) continue to tolerate the inconsistent
9853 forms that it has tolerated in the past, but authors and packagers should
9854 consider trying openssl and their own applications when compiled with
9855 these debugging symbols defined. It will help highlight potential bugs in
9856 their own code, and will improve the test coverage for OpenSSL itself. At
9857 some point, these tighter rules will become openssl's default to improve
9858 maintainability, though the assert()s and other overheads will remain only
9859 in debugging configurations. See bn.h for more details.
9861 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9863 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9864 that can only be obtained through BN_CTX_new() (which implicitly
9865 initialises it). The presence of this function only made it possible
9866 to overwrite an existing structure (and cause memory leaks).
9870 * Because of the callback-based approach for implementing LHASH as a
9871 template type, lh_insert() adds opaque objects to hash-tables and
9872 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9873 to clean up those corresponding objects before destroying the hash table
9874 (and losing the object pointers). So some over-zealous constifications in
9875 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9876 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9877 prototyped to have "const" restrictions on the object pointers they are
9878 given (and so aren't required to cast them away any more).
9882 * The tmdiff.h API was so ugly and minimal that our own timing utility
9883 (speed) prefers to use its own implementation. The two implementations
9884 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9885 its object type properly exposed (MS_TM) instead of casting to/from
9886 `char *`. This may still change yet if someone realises MS_TM and
9888 aren't necessarily the greatest nomenclatures - but this is what was used
9889 internally to the implementation so I've used that for now.
9893 * Ensure that deprecated functions do not get compiled when
9894 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9895 the self-tests were still using deprecated key-generation functions so
9896 these have been updated also.
9900 * Reorganise PKCS#7 code to separate the digest location functionality
9901 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9902 New function PKCS7_set_digest() to set the digest type for PKCS#7
9903 digestedData type. Add additional code to correctly generate the
9904 digestedData type and add support for this type in PKCS7 initialization
9909 * New function PKCS7_set0_type_other() this initializes a PKCS7
9910 structure of type "other".
9914 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9915 sure the loop does correctly stop and breaking ("division by zero")
9916 modulus operations are not performed. The (pre-generated) prime
9917 table crypto/bn/bn_prime.h was already correct, but it could not be
9918 re-generated on some platforms because of the "division by zero"
9919 situation in the script.
9921 *Ralf S. Engelschall*
9923 * Update support for ECC-based TLS ciphersuites according to
9924 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9925 SHA-1 now is only used for "small" curves (where the
9926 representation of a field element takes up to 24 bytes); for
9927 larger curves, the field element resulting from ECDH is directly
9928 used as premaster secret.
9930 *Douglas Stebila (Sun Microsystems Laboratories)*
9932 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9933 curve secp160r1 to the tests.
9935 *Douglas Stebila (Sun Microsystems Laboratories)*
9937 * Add the possibility to load symbols globally with DSO.
9939 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9941 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9942 control of the error stack.
9946 * Add support for STORE in ENGINE.
9950 * Add the STORE type. The intention is to provide a common interface
9951 to certificate and key stores, be they simple file-based stores, or
9952 HSM-type store, or LDAP stores, or...
9953 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9957 * Add a generic structure called OPENSSL_ITEM. This can be used to
9958 pass a list of arguments to any function as well as provide a way
9959 for a function to pass data back to the caller.
9963 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9964 works like BUF_strdup() but can be used to duplicate a portion of
9965 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9970 * Add the function sk_find_ex() which works like sk_find(), but will
9971 return an index to an element even if an exact match couldn't be
9972 found. The index is guaranteed to point at the element where the
9973 searched-for key would be inserted to preserve sorting order.
9977 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9978 takes an extra flags argument for optional functionality. Currently,
9979 the following flags are defined:
9981 OBJ_BSEARCH_VALUE_ON_NOMATCH
9982 This one gets OBJ_bsearch_ex() to return a pointer to the first
9983 element where the comparing function returns a negative or zero
9986 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9987 This one gets OBJ_bsearch_ex() to return a pointer to the first
9988 element where the comparing function returns zero. This is useful
9989 if there are more than one element where the comparing function
9994 * Make it possible to create self-signed certificates with 'openssl ca'
9995 in such a way that the self-signed certificate becomes part of the
9996 CA database and uses the same mechanisms for serial number generation
9997 as all other certificate signing. The new flag '-selfsign' enables
9998 this functionality. Adapt CA.sh and CA.pl.in.
10002 * Add functionality to check the public key of a certificate request
10003 against a given private. This is useful to check that a certificate
10004 request can be signed by that key (self-signing).
10008 * Make it possible to have multiple active certificates with the same
10009 subject in the CA index file. This is done only if the keyword
10010 'unique_subject' is set to 'no' in the main CA section (default
10011 if 'CA_default') of the configuration file. The value is saved
10012 with the database itself in a separate index attribute file,
10013 named like the index file with '.attr' appended to the name.
10017 * Generate multi-valued AVAs using '+' notation in config files for
10022 * Support for nameConstraints certificate extension.
10026 * Support for policyConstraints certificate extension.
10030 * Support for policyMappings certificate extension.
10034 * Make sure the default DSA_METHOD implementation only uses its
10035 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10036 and change its own handlers to be NULL so as to remove unnecessary
10037 indirection. This lets alternative implementations fallback to the
10038 default implementation more easily.
10042 * Support for directoryName in GeneralName related extensions
10047 * Make it possible to link applications using Makefile.shared.
10048 Make that possible even when linking against static libraries!
10052 * Support for single pass processing for S/MIME signing. This now
10053 means that S/MIME signing can be done from a pipe, in addition
10054 cleartext signing (multipart/signed type) is effectively streaming
10055 and the signed data does not need to be all held in memory.
10057 This is done with a new flag PKCS7_STREAM. When this flag is set
10058 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10059 is done after the data is output (and digests calculated) in
10060 SMIME_write_PKCS7().
10064 * Add full support for -rpath/-R, both in shared libraries and
10065 applications, at least on the platforms where it's known how
10070 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10071 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10072 will now compute a table of multiples of the generator that
10073 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10074 faster (notably in the case of a single point multiplication,
10075 scalar * generator).
10077 *Nils Larsch, Bodo Moeller*
10079 * IPv6 support for certificate extensions. The various extensions
10080 which use the IP:a.b.c.d can now take IPv6 addresses using the
10081 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10086 * Added an ENGINE that implements RSA by performing private key
10087 exponentiations with the GMP library. The conversions to and from
10088 GMP's mpz_t format aren't optimised nor are any montgomery forms
10089 cached, and on x86 it appears OpenSSL's own performance has caught up.
10090 However there are likely to be other architectures where GMP could
10091 provide a boost. This ENGINE is not built in by default, but it can be
10092 specified at Configure time and should be accompanied by the necessary
10093 linker additions, eg;
10094 ./config -DOPENSSL_USE_GMP -lgmp
10098 * "openssl engine" will not display ENGINE/DSO load failure errors when
10099 testing availability of engines with "-t" - the old behaviour is
10100 produced by increasing the feature's verbosity with "-tt".
10104 * ECDSA routines: under certain error conditions uninitialized BN objects
10105 could be freed. Solution: make sure initialization is performed early
10106 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10111 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10112 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10113 software implementations. For DSA and DH, parameter generation can
10114 also be overridden by providing the appropriate method callbacks.
10118 * Change the "progress" mechanism used in key-generation and
10119 primality testing to functions that take a new BN_GENCB pointer in
10120 place of callback/argument pairs. The new API functions have `_ex`
10121 postfixes and the older functions are reimplemented as wrappers for
10122 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10123 declarations of the old functions to help (graceful) attempts to
10124 migrate to the new functions. Also, the new key-generation API
10125 functions operate on a caller-supplied key-structure and return
10126 success/failure rather than returning a key or NULL - this is to
10127 help make "keygen" another member function of RSA_METHOD etc.
10129 Example for using the new callback interface:
10131 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10132 void *my_arg = ...;
10135 BN_GENCB_set(&my_cb, my_callback, my_arg);
10137 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10138 /* For the meaning of a, b in calls to my_callback(), see the
10139 * documentation of the function that calls the callback.
10140 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10141 * my_callback should return 1 if it wants BN_is_prime_ex()
10142 * to continue, or 0 to stop.
10147 * Change the ZLIB compression method to be stateful, and make it
10148 available to TLS with the number defined in
10149 draft-ietf-tls-compression-04.txt.
10153 * Add the ASN.1 structures and functions for CertificatePair, which
10154 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10156 CertificatePair ::= SEQUENCE {
10157 forward [0] Certificate OPTIONAL,
10158 reverse [1] Certificate OPTIONAL,
10159 -- at least one of the pair shall be present -- }
10161 Also implement the PEM functions to read and write certificate
10162 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10164 This needed to be defined, mostly for the sake of the LDAP
10165 attribute crossCertificatePair, but may prove useful elsewhere as
10170 * Make it possible to inhibit symlinking of shared libraries in
10171 Makefile.shared, for Cygwin's sake.
10175 * Extend the BIGNUM API by creating a function
10176 void BN_set_negative(BIGNUM *a, int neg);
10177 and a macro that behave like
10178 int BN_is_negative(const BIGNUM *a);
10180 to avoid the need to access 'a->neg' directly in applications.
10184 * Implement fast modular reduction for pseudo-Mersenne primes
10185 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10186 EC_GROUP_new_curve_GFp() will now automatically use this
10189 *Nils Larsch <nla@trustcenter.de>*
10191 * Add new lock type (CRYPTO_LOCK_BN).
10195 * Change the ENGINE framework to automatically load engines
10196 dynamically from specific directories unless they could be
10197 found to already be built in or loaded. Move all the
10198 current engines except for the cryptodev one to a new
10199 directory engines/.
10200 The engines in engines/ are built as shared libraries if
10201 the "shared" options was given to ./Configure or ./config.
10202 Otherwise, they are inserted in libcrypto.a.
10203 /usr/local/ssl/engines is the default directory for dynamic
10204 engines, but that can be overridden at configure time through
10205 the usual use of --prefix and/or --openssldir, and at run
10206 time with the environment variable OPENSSL_ENGINES.
10208 *Geoff Thorpe and Richard Levitte*
10210 * Add Makefile.shared, a helper makefile to build shared
10211 libraries. Adapt Makefile.org.
10215 * Add version info to Win32 DLLs.
10217 *Peter 'Luna' Runestig" <peter@runestig.com>*
10219 * Add new 'medium level' PKCS#12 API. Certificates and keys
10220 can be added using this API to created arbitrary PKCS#12
10221 files while avoiding the low level API.
10223 New options to PKCS12_create(), key or cert can be NULL and
10224 will then be omitted from the output file. The encryption
10225 algorithm NIDs can be set to -1 for no encryption, the mac
10226 iteration count can be set to 0 to omit the mac.
10228 Enhance pkcs12 utility by making the -nokeys and -nocerts
10229 options work when creating a PKCS#12 file. New option -nomac
10230 to omit the mac, NONE can be set for an encryption algorithm.
10231 New code is modified to use the enhanced PKCS12_create()
10232 instead of the low level API.
10236 * Extend ASN1 encoder to support indefinite length constructed
10237 encoding. This can output sequences tags and octet strings in
10238 this form. Modify pk7_asn1.c to support indefinite length
10239 encoding. This is experimental and needs additional code to
10240 be useful, such as an ASN1 bio and some enhanced streaming
10243 Extend template encode functionality so that tagging is passed
10244 down to the template encoder.
10248 * Let 'openssl req' fail if an argument to '-newkey' is not
10249 recognized instead of using RSA as a default.
10253 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10254 As these are not official, they are not included in "ALL";
10255 the "ECCdraft" ciphersuite group alias can be used to select them.
10257 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10259 * Add ECDH engine support.
10261 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10263 * Add ECDH in new directory crypto/ecdh/.
10265 *Douglas Stebila (Sun Microsystems Laboratories)*
10267 * Let BN_rand_range() abort with an error after 100 iterations
10268 without success (which indicates a broken PRNG).
10272 * Change BN_mod_sqrt() so that it verifies that the input value
10273 is really the square of the return value. (Previously,
10274 BN_mod_sqrt would show GIGO behaviour.)
10278 * Add named elliptic curves over binary fields from X9.62, SECG,
10279 and WAP/WTLS; add OIDs that were still missing.
10281 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10283 * Extend the EC library for elliptic curves over binary fields
10284 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10287 EC_GF2m_simple_method
10291 EC_GROUP_new_curve_GF2m
10292 EC_GROUP_set_curve_GF2m
10293 EC_GROUP_get_curve_GF2m
10294 EC_POINT_set_affine_coordinates_GF2m
10295 EC_POINT_get_affine_coordinates_GF2m
10296 EC_POINT_set_compressed_coordinates_GF2m
10298 Point compression for binary fields is disabled by default for
10299 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10302 As binary polynomials are represented as BIGNUMs, various members
10303 of the EC_GROUP and EC_POINT data structures can be shared
10304 between the implementations for prime fields and binary fields;
10305 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10306 are essentially identical to their `..._GFp` counterparts.
10307 (For simplicity, the `..._GFp` prefix has been dropped from
10308 various internal method names.)
10310 An internal 'field_div' method (similar to 'field_mul' and
10311 'field_sqr') has been added; this is used only for binary fields.
10313 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10315 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10316 through methods ('mul', 'precompute_mult').
10318 The generic implementations (now internally called 'ec_wNAF_mul'
10319 and 'ec_wNAF_precomputed_mult') remain the default if these
10320 methods are undefined.
10322 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10324 * New function EC_GROUP_get_degree, which is defined through
10325 EC_METHOD. For curves over prime fields, this returns the bit
10326 length of the modulus.
10328 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10330 * New functions EC_GROUP_dup, EC_POINT_dup.
10331 (These simply call ..._new and ..._copy).
10333 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10335 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10336 Polynomials are represented as BIGNUMs (where the sign bit is not
10337 used) in the following functions [macros]:
10340 BN_GF2m_sub [= BN_GF2m_add]
10341 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10342 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10343 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10345 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10346 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10347 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10348 BN_GF2m_cmp [= BN_ucmp]
10350 (Note that only the 'mod' functions are actually for fields GF(2^m).
10351 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10353 For some functions, an the irreducible polynomial defining a
10354 field can be given as an 'unsigned int[]' with strictly
10355 decreasing elements giving the indices of those bits that are set;
10356 i.e., p[] represents the polynomial
10357 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10359 p[0] > p[1] > ... > p[k] = 0.
10360 This applies to the following functions:
10363 BN_GF2m_mod_mul_arr
10364 BN_GF2m_mod_sqr_arr
10365 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10366 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10367 BN_GF2m_mod_exp_arr
10368 BN_GF2m_mod_sqrt_arr
10369 BN_GF2m_mod_solve_quad_arr
10373 Conversion can be performed by the following functions:
10378 bntest.c has additional tests for binary polynomial arithmetic.
10380 Two implementations for BN_GF2m_mod_div() are available.
10381 The default algorithm simply uses BN_GF2m_mod_inv() and
10382 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10383 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10384 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10386 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10388 * Add new error code 'ERR_R_DISABLED' that can be used when some
10389 functionality is disabled at compile-time.
10391 *Douglas Stebila <douglas.stebila@sun.com>*
10393 * Change default behaviour of 'openssl asn1parse' so that more
10394 information is visible when viewing, e.g., a certificate:
10396 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10397 mode the content of non-printable OCTET STRINGs is output in a
10398 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10399 avoid the appearance of a printable string.
10401 *Nils Larsch <nla@trustcenter.de>*
10403 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10405 EC_GROUP_set_asn1_flag()
10406 EC_GROUP_get_asn1_flag()
10407 EC_GROUP_set_point_conversion_form()
10408 EC_GROUP_get_point_conversion_form()
10409 These control ASN1 encoding details:
10410 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10411 has been set to OPENSSL_EC_NAMED_CURVE.
10412 - Points are encoded in uncompressed form by default; options for
10413 asn1_for are as for point2oct, namely
10414 POINT_CONVERSION_COMPRESSED
10415 POINT_CONVERSION_UNCOMPRESSED
10416 POINT_CONVERSION_HYBRID
10418 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10420 EC_GROUP_set_seed()
10421 EC_GROUP_get0_seed()
10422 EC_GROUP_get_seed_len()
10423 This is used only for ASN1 purposes (so far).
10425 *Nils Larsch <nla@trustcenter.de>*
10427 * Add 'field_type' member to EC_METHOD, which holds the NID
10428 of the appropriate field type OID. The new function
10429 EC_METHOD_get_field_type() returns this value.
10431 *Nils Larsch <nla@trustcenter.de>*
10434 EC_POINT_point2bn()
10435 EC_POINT_bn2point()
10436 EC_POINT_point2hex()
10437 EC_POINT_hex2point()
10438 providing useful interfaces to EC_POINT_point2oct() and
10439 EC_POINT_oct2point().
10441 *Nils Larsch <nla@trustcenter.de>*
10443 * Change internals of the EC library so that the functions
10444 EC_GROUP_set_generator()
10445 EC_GROUP_get_generator()
10446 EC_GROUP_get_order()
10447 EC_GROUP_get_cofactor()
10448 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10449 to methods, which would lead to unnecessary code duplication when
10450 adding different types of curves.
10452 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10454 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10455 arithmetic, and such that modified wNAFs are generated
10456 (which avoid length expansion in many cases).
10460 * Add a function EC_GROUP_check_discriminant() (defined via
10461 EC_METHOD) that verifies that the curve discriminant is non-zero.
10463 Add a function EC_GROUP_check() that makes some sanity tests
10464 on a EC_GROUP, its generator and order. This includes
10465 EC_GROUP_check_discriminant().
10467 *Nils Larsch <nla@trustcenter.de>*
10469 * Add ECDSA in new directory crypto/ecdsa/.
10471 Add applications 'openssl ecparam' and 'openssl ecdsa'
10472 (these are based on 'openssl dsaparam' and 'openssl dsa').
10474 ECDSA support is also included in various other files across the
10475 library. Most notably,
10476 - 'openssl req' now has a '-newkey ecdsa:file' option;
10477 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10478 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10479 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10480 them suitable for ECDSA where domain parameters must be
10481 extracted before the specific public key;
10482 - ECDSA engine support has been added.
10484 *Nils Larsch <nla@trustcenter.de>*
10486 * Include some named elliptic curves, and add OIDs from X9.62,
10487 SECG, and WAP/WTLS. Each curve can be obtained from the new
10489 EC_GROUP_new_by_curve_name(),
10490 and the list of available named curves can be obtained with
10491 EC_get_builtin_curves().
10492 Also add a 'curve_name' member to EC_GROUP objects, which can be
10494 EC_GROUP_set_curve_name()
10495 EC_GROUP_get_curve_name()
10497 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10499 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10500 was actually never needed) and in BN_mul(). The removal in BN_mul()
10501 required a small change in bn_mul_part_recursive() and the addition
10502 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10503 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10504 bn_sub_words() and bn_add_words() except they take arrays with
10509 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10511 * Cleanse PEM buffers before freeing them since they may contain
10514 *Benjamin Bennett <ben@psc.edu>*
10516 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10517 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10518 authentication-only ciphersuites.
10522 * Since AES128 and AES256 share a single mask bit in the logic of
10523 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10524 kludge to work properly if AES128 is available and AES256 isn't.
10528 * Expand security boundary to match 1.1.1 module.
10532 * Remove redundant features: hash file source, editing of test vectors
10533 modify fipsld to use external fips_premain.c signature.
10537 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10538 run algorithm test programs.
10542 * Make algorithm test programs more tolerant of whitespace.
10546 * Have SSL/TLS server implementation tolerate "mismatched" record
10547 protocol version while receiving ClientHello even if the
10548 ClientHello is fragmented. (The server can't insist on the
10549 particular protocol version it has chosen before the ServerHello
10550 message has informed the client about his choice.)
10554 * Load error codes if they are not already present instead of using a
10555 static variable. This allows them to be cleanly unloaded and reloaded.
10559 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10561 * Introduce limits to prevent malicious keys being able to
10562 cause a denial of service. [CVE-2006-2940][]
10564 *Steve Henson, Bodo Moeller*
10566 * Fix ASN.1 parsing of certain invalid structures that can result
10567 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10569 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10570 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10572 * Fix SSL client code which could crash if connecting to a
10573 malicious SSLv2 server. [CVE-2006-4343][]
10575 *Tavis Ormandy and Will Drewry, Google Security Team*
10577 * Change ciphersuite string processing so that an explicit
10578 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10579 will no longer include "AES128-SHA"), and any other similar
10580 ciphersuite (same bitmap) from *other* protocol versions (so that
10581 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10582 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10583 changes from 0.9.8b and 0.9.8d.
10587 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10589 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10590 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10592 * Change the Unix randomness entropy gathering to use poll() when
10593 possible instead of select(), since the latter has some
10594 undesirable limitations.
10596 *Darryl Miles via Richard Levitte and Bodo Moeller*
10598 * Disable rogue ciphersuites:
10600 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10601 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10602 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10604 The latter two were purportedly from
10605 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10608 Also deactivate the remaining ciphersuites from
10609 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10610 unofficial, and the ID has long expired.
10614 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10615 dual-core machines) and other potential thread-safety issues.
10619 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10621 * Adapt fipsld and the build system to link against the validated FIPS
10622 module in FIPS mode.
10626 * Fixes for VC++ 2005 build under Windows.
10630 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10631 from a Windows bash shell such as MSYS. It is autodetected from the
10632 "config" script when run from a VC++ environment. Modify standard VC++
10633 build to use fipscanister.o from the GNU make build.
10637 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10639 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10640 The value now differs depending on if you build for FIPS or not.
10641 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10642 safely run with a non-FIPSed libcrypto, as it may crash because of
10643 the difference induced by this change.
10647 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10649 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10650 (part of SSL_OP_ALL). This option used to disable the
10651 countermeasure against man-in-the-middle protocol-version
10652 rollback in the SSL 2.0 server implementation, which is a bad
10653 idea. [CVE-2005-2969][]
10655 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10656 for Information Security, National Institute of Advanced Industrial
10657 Science and Technology [AIST, Japan)]*
10659 * Minimal support for X9.31 signatures and PSS padding modes. This is
10660 mainly for FIPS compliance and not fully integrated at this stage.
10664 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10665 the exponentiation using a fixed-length exponent. (Otherwise,
10666 the information leaked through timing could expose the secret key
10667 after many signatures; cf. Bleichenbacher's attack on DSA with
10672 * Make a new fixed-window mod_exp implementation the default for
10673 RSA, DSA, and DH private-key operations so that the sequence of
10674 squares and multiplies and the memory access pattern are
10675 independent of the particular secret key. This will mitigate
10676 cache-timing and potential related attacks.
10678 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10679 and this is automatically used by BN_mod_exp_mont() if the new flag
10680 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10681 will use this BN flag for private exponents unless the flag
10682 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10683 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10685 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10687 * Change the client implementation for SSLv23_method() and
10688 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10689 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10690 (Previously, the SSL 2.0 backwards compatible Client Hello
10691 message format would be used even with SSL_OP_NO_SSLv2.)
10695 * Add support for smime-type MIME parameter in S/MIME messages which some
10700 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10701 a threadsafe manner. Modify rsa code to use new function and add calls
10702 to dsa and dh code (which had race conditions before).
10706 * Include the fixed error library code in the C error file definitions
10707 instead of fixing them up at runtime. This keeps the error code
10708 structures constant.
10712 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10714 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10717 * Fixes for newer kerberos headers. NB: the casts are needed because
10718 the 'length' field is signed on one version and unsigned on another
10719 with no (?) obvious way to tell the difference, without these VC++
10720 complains. Also the "definition" of FAR (blank) is no longer included
10721 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10722 some needed definitions.
10726 * Undo Cygwin change.
10730 * Added support for proxy certificates according to RFC 3820.
10731 Because they may be a security thread to unaware applications,
10732 they must be explicitly allowed in run-time. See
10733 docs/HOWTO/proxy_certificates.txt for further information.
10737 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10739 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10740 server and client random values. Previously
10741 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10742 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10744 This change has negligible security impact because:
10746 1. Server and client random values still have 24 bytes of pseudo random
10749 2. Server and client random values are sent in the clear in the initial
10752 3. The master secret is derived using the premaster secret (48 bytes in
10753 size for static RSA ciphersuites) as well as client server and random
10756 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10759 *Stephen Henson, reported by UK NISCC*
10761 * Use Windows randomness collection on Cygwin.
10765 * Fix hang in EGD/PRNGD query when communication socket is closed
10766 prematurely by EGD/PRNGD.
10768 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10770 * Prompt for pass phrases when appropriate for PKCS12 input format.
10774 * Back-port of selected performance improvements from development
10775 branch, as well as improved support for PowerPC platforms.
10779 * Add lots of checks for memory allocation failure, error codes to indicate
10780 failure and freeing up memory if a failure occurs.
10782 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10784 * Add new -passin argument to dgst.
10788 * Perform some character comparisons of different types in X509_NAME_cmp:
10789 this is needed for some certificates that re-encode DNs into UTF8Strings
10790 (in violation of RFC3280) and can't or won't issue name rollover
10795 * Make an explicit check during certificate validation to see that
10796 the CA setting in each certificate on the chain is correct. As a
10797 side effect always do the following basic checks on extensions,
10798 not just when there's an associated purpose to the check:
10800 - if there is an unhandled critical extension (unless the user
10801 has chosen to ignore this fault)
10802 - if the path length has been exceeded (if one is set at all)
10803 - that certain extensions fit the associated purpose (if one has
10808 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10810 * Avoid a race condition when CRLs are checked in a multi threaded
10811 environment. This would happen due to the reordering of the revoked
10812 entries during signature checking and serial number lookup. Now the
10813 encoding is cached and the serial number sort performed under a lock.
10814 Add new STACK function sk_is_sorted().
10818 * Add Delta CRL to the extension code.
10822 * Various fixes to s3_pkt.c so alerts are sent properly.
10824 *David Holmes <d.holmes@f5.com>*
10826 * Reduce the chances of duplicate issuer name and serial numbers (in
10827 violation of RFC3280) using the OpenSSL certificate creation utilities.
10828 This is done by creating a random 64 bit value for the initial serial
10829 number when a serial number file is created or when a self signed
10830 certificate is created using 'openssl req -x509'. The initial serial
10831 number file is created using 'openssl x509 -next_serial' in CA.pl
10832 rather than being initialized to 1.
10836 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10838 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10839 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10841 *Joe Orton, Steve Henson*
10843 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10846 *Joe Orton, Steve Henson*
10848 * Make it possible to have multiple active certificates with the same
10849 subject in the CA index file. This is done only if the keyword
10850 'unique_subject' is set to 'no' in the main CA section (default
10851 if 'CA_default') of the configuration file. The value is saved
10852 with the database itself in a separate index attribute file,
10853 named like the index file with '.attr' appended to the name.
10857 * X509 verify fixes. Disable broken certificate workarounds when
10858 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10859 keyUsage extension present. Don't accept CRLs with unhandled critical
10860 extensions: since verify currently doesn't process CRL extensions this
10861 rejects a CRL with *any* critical extensions. Add new verify error codes
10866 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10867 A clarification of RFC2560 will require the use of OCTET STRINGs and
10868 some implementations cannot handle the current raw format. Since OpenSSL
10869 copies and compares OCSP nonces as opaque blobs without any attempt at
10870 parsing them this should not create any compatibility issues.
10874 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10875 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10876 this HMAC (and other) operations are several times slower than OpenSSL
10881 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10883 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10885 * Use the correct content when signing type "other".
10889 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10891 * Fix various bugs revealed by running the NISCC test suite:
10893 Stop out of bounds reads in the ASN1 code when presented with
10894 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10896 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10898 If verify callback ignores invalid public key errors don't try to check
10899 certificate signature with the NULL public key.
10903 * New -ignore_err option in ocsp application to stop the server
10904 exiting on the first error in a request.
10908 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10909 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10914 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10915 extra data after the compression methods not only for TLS 1.0
10916 but also for SSL 3.0 (as required by the specification).
10918 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10920 * Change X509_certificate_type() to mark the key as exported/exportable
10921 when it's 512 *bits* long, not 512 bytes.
10925 * Change AES_cbc_encrypt() so it outputs exact multiple of
10926 blocks during encryption.
10930 * Various fixes to base64 BIO and non blocking I/O. On write
10931 flushes were not handled properly if the BIO retried. On read
10932 data was not being buffered properly and had various logic bugs.
10933 This also affects blocking I/O when the data being decoded is a
10938 * Various S/MIME bugfixes and compatibility changes:
10939 output correct application/pkcs7 MIME type if
10940 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10941 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10942 of files as .eml work). Correctly handle very long lines in MIME
10947 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10949 * Countermeasure against the Klima-Pokorny-Rosa extension of
10950 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10951 a protocol version number mismatch like a decryption error
10952 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10956 * Turn on RSA blinding by default in the default implementation
10957 to avoid a timing attack. Applications that don't want it can call
10958 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10959 They would be ill-advised to do so in most cases.
10961 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10963 * Change RSA blinding code so that it works when the PRNG is not
10964 seeded (in this case, the secret RSA exponent is abused as
10965 an unpredictable seed -- if it is not unpredictable, there
10966 is no point in blinding anyway). Make RSA blinding thread-safe
10967 by remembering the creator's thread ID in rsa->blinding and
10968 having all other threads use local one-time blinding factors
10969 (this requires more computation than sharing rsa->blinding, but
10970 avoids excessive locking; and if an RSA object is not shared
10971 between threads, blinding will still be very fast).
10975 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10976 ENGINE as defaults for all supported algorithms irrespective of
10977 the 'flags' parameter. 'flags' is now honoured, so applications
10978 should make sure they are passing it correctly.
10982 * Target "mingw" now allows native Windows code to be generated in
10983 the Cygwin environment as well as with the MinGW compiler.
10987 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10989 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10990 via timing by performing a MAC computation even if incorrect
10991 block cipher padding has been found. This is a countermeasure
10992 against active attacks where the attacker has to distinguish
10993 between bad padding and a MAC verification error. [CVE-2003-0078][]
10995 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10996 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10997 Martin Vuagnoux (EPFL, Ilion)*
10999 * Make the no-err option work as intended. The intention with no-err
11000 is not to have the whole error stack handling routines removed from
11001 libcrypto, it's only intended to remove all the function name and
11002 reason texts, thereby removing some of the footprint that may not
11003 be interesting if those errors aren't displayed anyway.
11005 NOTE: it's still possible for any application or module to have its
11006 own set of error texts inserted. The routines are there, just not
11007 used by default when no-err is given.
11011 * Add support for FreeBSD on IA64.
11013 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11015 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11016 Kerberos function mit_des_cbc_cksum(). Before this change,
11017 the value returned by DES_cbc_cksum() was like the one from
11018 mit_des_cbc_cksum(), except the bytes were swapped.
11020 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11022 * Allow an application to disable the automatic SSL chain building.
11023 Before this a rather primitive chain build was always performed in
11024 ssl3_output_cert_chain(): an application had no way to send the
11025 correct chain if the automatic operation produced an incorrect result.
11027 Now the chain builder is disabled if either:
11029 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11031 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11033 The reasoning behind this is that an application would not want the
11034 auto chain building to take place if extra chain certificates are
11035 present and it might also want a means of sending no additional
11036 certificates (for example the chain has two certificates and the
11041 * Add the possibility to build without the ENGINE framework.
11043 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11045 * Under Win32 gmtime() can return NULL: check return value in
11046 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11050 * DSA routines: under certain error conditions uninitialized BN objects
11051 could be freed. Solution: make sure initialization is performed early
11052 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11053 Nils Larsch <nla@trustcenter.de> via PR#459)
11057 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11058 checked on reconnect on the client side, therefore session resumption
11059 could still fail with a "ssl session id is different" error. This
11060 behaviour is masked when SSL_OP_ALL is used due to
11061 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11062 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11063 followup to PR #377.
11067 * IA-32 assembler support enhancements: unified ELF targets, support
11068 for SCO/Caldera platforms, fix for Cygwin shared build.
11072 * Add support for FreeBSD on sparc64. As a consequence, support for
11073 FreeBSD on non-x86 processors is separate from x86 processors on
11074 the config script, much like the NetBSD support.
11076 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11078 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11080 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11083 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11084 code (06) was taken as the first octet of the session ID and the last
11085 octet was ignored consequently. As a result SSLv2 client side session
11086 caching could not have worked due to the session ID mismatch between
11088 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11093 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11094 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11099 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11100 seems that in spite of existing for more than a year, many application
11101 author have done nothing to provide the necessary callbacks, which
11102 means that this particular engine will not work properly anywhere.
11103 This is a very unfortunate situation which forces us, in the name
11104 of usability, to give the hw_ncipher.c a static lock, which is part
11106 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11107 appear in 0.9.8 or later. We EXPECT application authors to have
11108 dealt properly with this when 0.9.8 is released (unless we actually
11109 make such changes in the libcrypto locking code that changes will
11110 have to be made anyway).
11114 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11115 octets have been read, EOF or an error occurs. Without this change
11116 some truncated ASN1 structures will not produce an error.
11120 * Disable Heimdal support, since it hasn't been fully implemented.
11121 Still give the possibility to force the use of Heimdal, but with
11122 warnings and a request that patches get sent to openssl-dev.
11126 * Add the VC-CE target, introduce the WINCE sysname, and add
11127 INSTALL.WCE and appropriate conditionals to make it build.
11129 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11131 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11132 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11133 edit numbers of the version.
11135 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11137 * Introduce safe string copy and catenation functions
11138 (BUF_strlcpy() and BUF_strlcat()).
11140 *Ben Laurie (CHATS) and Richard Levitte*
11142 * Avoid using fixed-size buffers for one-line DNs.
11144 *Ben Laurie (CHATS)*
11146 * Add BUF_MEM_grow_clean() to avoid information leakage when
11147 resizing buffers containing secrets, and use where appropriate.
11149 *Ben Laurie (CHATS)*
11151 * Avoid using fixed size buffers for configuration file location.
11153 *Ben Laurie (CHATS)*
11155 * Avoid filename truncation for various CA files.
11157 *Ben Laurie (CHATS)*
11159 * Use sizeof in preference to magic numbers.
11161 *Ben Laurie (CHATS)*
11163 * Avoid filename truncation in cert requests.
11165 *Ben Laurie (CHATS)*
11167 * Add assertions to check for (supposedly impossible) buffer
11170 *Ben Laurie (CHATS)*
11172 * Don't cache truncated DNS entries in the local cache (this could
11173 potentially lead to a spoofing attack).
11175 *Ben Laurie (CHATS)*
11177 * Fix various buffers to be large enough for hex/decimal
11178 representations in a platform independent manner.
11180 *Ben Laurie (CHATS)*
11182 * Add CRYPTO_realloc_clean() to avoid information leakage when
11183 resizing buffers containing secrets, and use where appropriate.
11185 *Ben Laurie (CHATS)*
11187 * Add BIO_indent() to avoid much slightly worrying code to do
11190 *Ben Laurie (CHATS)*
11192 * Convert sprintf()/BIO_puts() to BIO_printf().
11194 *Ben Laurie (CHATS)*
11196 * buffer_gets() could terminate with the buffer only half
11199 *Ben Laurie (CHATS)*
11201 * Add assertions to prevent user-supplied crypto functions from
11202 overflowing internal buffers by having large block sizes, etc.
11204 *Ben Laurie (CHATS)*
11206 * New OPENSSL_assert() macro (similar to assert(), but enabled
11209 *Ben Laurie (CHATS)*
11211 * Eliminate unused copy of key in RC4.
11213 *Ben Laurie (CHATS)*
11215 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11217 *Ben Laurie (CHATS)*
11219 * Fix off-by-one error in EGD path.
11221 *Ben Laurie (CHATS)*
11223 * If RANDFILE path is too long, ignore instead of truncating.
11225 *Ben Laurie (CHATS)*
11227 * Eliminate unused and incorrectly sized X.509 structure
11230 *Ben Laurie (CHATS)*
11232 * Eliminate unused and dangerous function knumber().
11234 *Ben Laurie (CHATS)*
11236 * Eliminate unused and dangerous structure, KSSL_ERR.
11238 *Ben Laurie (CHATS)*
11240 * Protect against overlong session ID context length in an encoded
11241 session object. Since these are local, this does not appear to be
11244 *Ben Laurie (CHATS)*
11246 * Change from security patch (see 0.9.6e below) that did not affect
11247 the 0.9.6 release series:
11249 Remote buffer overflow in SSL3 protocol - an attacker could
11250 supply an oversized master key in Kerberos-enabled versions.
11253 *Ben Laurie (CHATS)*
11255 * Change the SSL kerb5 codes to match RFC 2712.
11259 * Make -nameopt work fully for req and add -reqopt switch.
11261 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11263 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11265 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11267 * Make sure tests can be performed even if the corresponding algorithms
11268 have been removed entirely. This was also the last step to make
11269 OpenSSL compilable with DJGPP under all reasonable conditions.
11271 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11273 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11274 to allow version independent disabling of normally unselected ciphers,
11275 which may be activated as a side-effect of selecting a single cipher.
11277 (E.g., cipher list string "RSA" enables ciphersuites that are left
11278 out of "ALL" because they do not provide symmetric encryption.
11279 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11281 *Lutz Jaenicke, Bodo Moeller*
11283 * Add appropriate support for separate platform-dependent build
11284 directories. The recommended way to make a platform-dependent
11285 build directory is the following (tested on Linux), maybe with
11288 # Place yourself outside of the OpenSSL source tree. In
11289 # this example, the environment variable OPENSSL_SOURCE
11290 # is assumed to contain the absolute OpenSSL source directory.
11291 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11292 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11293 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11294 mkdir -p `dirname $F`
11295 ln -s $OPENSSL_SOURCE/$F $F
11298 To be absolutely sure not to disturb the source tree, a "make clean"
11299 is a good thing. If it isn't successful, don't worry about it,
11300 it probably means the source directory is very clean.
11304 * Make sure any ENGINE control commands make local copies of string
11305 pointers passed to them whenever necessary. Otherwise it is possible
11306 the caller may have overwritten (or deallocated) the original string
11307 data when a later ENGINE operation tries to use the stored values.
11309 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11311 * Improve diagnostics in file reading and command-line digests.
11313 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11315 * Add AES modes CFB and OFB to the object database. Correct an
11316 error in AES-CFB decryption.
11320 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11321 allows existing EVP_CIPHER_CTX structures to be reused after
11322 calling `EVP_*Final()`. This behaviour is used by encryption
11323 BIOs and some applications. This has the side effect that
11324 applications must explicitly clean up cipher contexts with
11325 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11329 * Check the values of dna and dnb in bn_mul_recursive before calling
11330 bn_mul_comba (a non zero value means the a or b arrays do not contain
11331 n2 elements) and fallback to bn_mul_normal if either is not zero.
11335 * Fix escaping of non-ASCII characters when using the -subj option
11336 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11340 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11341 form for "surname", serialNumber has no short form.
11342 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11343 therefore remove "mail" short name for "internet 7".
11344 The OID for unique identifiers in X509 certificates is
11345 x500UniqueIdentifier, not uniqueIdentifier.
11346 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11350 * Add an "init" command to the ENGINE config module and auto initialize
11351 ENGINEs. Without any "init" command the ENGINE will be initialized
11352 after all ctrl commands have been executed on it. If init=1 the
11353 ENGINE is initialized at that point (ctrls before that point are run
11354 on the uninitialized ENGINE and after on the initialized one). If
11355 init=0 then the ENGINE will not be initialized at all.
11359 * Fix the 'app_verify_callback' interface so that the user-defined
11360 argument is actually passed to the callback: In the
11361 SSL_CTX_set_cert_verify_callback() prototype, the callback
11362 declaration has been changed from
11365 int (*cb)(X509_STORE_CTX *,void *);
11366 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11367 i=s->ctx->app_verify_callback(&ctx)
11368 has been changed into
11369 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11371 To update applications using SSL_CTX_set_cert_verify_callback(),
11372 a dummy argument can be added to their callback functions.
11374 *D. K. Smetters <smetters@parc.xerox.com>*
11376 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11378 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11380 * Add and OPENSSL_LOAD_CONF define which will cause
11381 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11382 This allows older applications to transparently support certain
11383 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11384 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11385 load the config file and OPENSSL_add_all_algorithms_conf() which will
11386 always load it have also been added.
11390 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11391 Adjust NIDs and EVP layer.
11393 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11395 * Config modules support in openssl utility.
11397 Most commands now load modules from the config file,
11398 though in a few (such as version) this isn't done
11399 because it couldn't be used for anything.
11401 In the case of ca and req the config file used is
11402 the same as the utility itself: that is the -config
11403 command line option can be used to specify an
11408 * Move default behaviour from OPENSSL_config(). If appname is NULL
11409 use "openssl_conf" if filename is NULL use default openssl config file.
11413 * Add an argument to OPENSSL_config() to allow the use of an alternative
11414 config section name. Add a new flag to tolerate a missing config file
11415 and move code to CONF_modules_load_file().
11419 * Support for crypto accelerator cards from Accelerated Encryption
11420 Processing, www.aep.ie. (Use engine 'aep')
11421 The support was copied from 0.9.6c [engine] and adapted/corrected
11422 to work with the new engine framework.
11424 *AEP Inc. and Richard Levitte*
11426 * Support for SureWare crypto accelerator cards from Baltimore
11427 Technologies. (Use engine 'sureware')
11428 The support was copied from 0.9.6c [engine] and adapted
11429 to work with the new engine framework.
11433 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11434 make the newer ENGINE framework commands for the CHIL engine work.
11436 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11438 * Make it possible to produce shared libraries on ReliantUNIX.
11440 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11442 * Add the configuration target debug-linux-ppro.
11443 Make 'openssl rsa' use the general key loading routines
11444 implemented in apps.c, and make those routines able to
11445 handle the key format FORMAT_NETSCAPE and the variant
11448 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11450 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11452 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11454 * Add -keyform to rsautl, and document -engine.
11456 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11458 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11459 BIO_R_NO_SUCH_FILE error code rather than the generic
11460 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11464 * Add new functions
11465 ERR_peek_last_error
11466 ERR_peek_last_error_line
11467 ERR_peek_last_error_line_data.
11468 These are similar to
11470 ERR_peek_error_line
11471 ERR_peek_error_line_data,
11472 but report on the latest error recorded rather than the first one
11473 still in the error queue.
11475 *Ben Laurie, Bodo Moeller*
11477 * default_algorithms option in ENGINE config module. This allows things
11479 default_algorithms = ALL
11480 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11484 * Preliminary ENGINE config module.
11488 * New experimental application configuration code.
11492 * Change the AES code to follow the same name structure as all other
11493 symmetric ciphers, and behave the same way. Move everything to
11494 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11496 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11498 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11500 *Ben Laurie and Theo de Raadt*
11502 * Add option to output public keys in req command.
11504 *Massimiliano Pala madwolf@openca.org*
11506 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11507 (up to about 10% better than before for P-192 and P-224).
11511 * New functions/macros
11513 SSL_CTX_set_msg_callback(ctx, cb)
11514 SSL_CTX_set_msg_callback_arg(ctx, arg)
11515 SSL_set_msg_callback(ssl, cb)
11516 SSL_set_msg_callback_arg(ssl, arg)
11518 to request calling a callback function
11520 void cb(int write_p, int version, int content_type,
11521 const void *buf, size_t len, SSL *ssl, void *arg)
11523 whenever a protocol message has been completely received
11524 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11525 protocol version according to which the SSL library interprets
11526 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11527 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11528 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11529 specification (change_cipher_spec(20), alert(21), handshake(22)).
11530 'buf' and 'len' point to the actual message, 'ssl' to the
11531 SSL object, and 'arg' is the application-defined value set by
11532 SSL[_CTX]_set_msg_callback_arg().
11534 'openssl s_client' and 'openssl s_server' have new '-msg' options
11535 to enable a callback that displays all protocol messages.
11539 * Change the shared library support so shared libraries are built as
11540 soon as the corresponding static library is finished, and thereby get
11541 openssl and the test programs linked against the shared library.
11542 This still only happens when the keyword "shard" has been given to
11543 the configuration scripts.
11545 NOTE: shared library support is still an experimental thing, and
11546 backward binary compatibility is still not guaranteed.
11548 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11550 * Add support for Subject Information Access extension.
11552 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11554 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11555 additional bytes when new memory had to be allocated, not just
11556 when reusing an existing buffer.
11560 * New command line and configuration option 'utf8' for the req command.
11561 This allows field values to be specified as UTF8 strings.
11565 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11566 runs for the former and machine-readable output for the latter.
11570 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11571 of the e-mail address in the DN (i.e., it will go into a certificate
11572 extension only). The new configuration file option 'email_in_dn = no'
11573 has the same effect.
11575 *Massimiliano Pala madwolf@openca.org*
11577 * Change all functions with names starting with `des_` to be starting
11578 with `DES_` instead. Add wrappers that are compatible with libdes,
11579 but are named `_ossl_old_des_*`. Finally, add macros that map the
11580 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11581 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11582 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11585 Since we provide two compatibility mappings, the user needs to
11586 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11587 compatibility is desired. The default (i.e., when that macro
11588 isn't defined) is OpenSSL 0.9.6c compatibility.
11590 There are also macros that enable and disable the support of old
11591 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11592 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11593 are defined, the default will apply: to support the old des routines.
11595 In either case, one must include openssl/des.h to get the correct
11596 definitions. Do not try to just include openssl/des_old.h, that
11599 NOTE: This is a major break of an old API into a new one. Software
11600 authors are encouraged to switch to the `DES_` style functions. Some
11601 time in the future, des_old.h and the libdes compatibility functions
11602 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11603 default), and then completely removed.
11607 * Test for certificates which contain unsupported critical extensions.
11608 If such a certificate is found during a verify operation it is
11609 rejected by default: this behaviour can be overridden by either
11610 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11611 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11612 X509_supported_extension() has also been added which returns 1 if a
11613 particular extension is supported.
11617 * Modify the behaviour of EVP cipher functions in similar way to digests
11618 to retain compatibility with existing code.
11622 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11623 compatibility with existing code. In particular the 'ctx' parameter does
11624 not have to be to be initialized before the call to EVP_DigestInit() and
11625 it is tidied up after a call to EVP_DigestFinal(). New function
11626 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11627 EVP_MD_CTX_copy() changed to not require the destination to be
11628 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11629 requires the destination to be valid.
11631 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11632 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11636 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11637 so that complete 'Handshake' protocol structures are kept in memory
11638 instead of overwriting 'msg_type' and 'length' with 'body' data.
11642 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11644 *Massimo Santin via Richard Levitte*
11646 * Major restructuring to the underlying ENGINE code. This includes
11647 reduction of linker bloat, separation of pure "ENGINE" manipulation
11648 (initialisation, etc) from functionality dealing with implementations
11649 of specific crypto interfaces. This change also introduces integrated
11650 support for symmetric ciphers and digest implementations - so ENGINEs
11651 can now accelerate these by providing EVP_CIPHER and EVP_MD
11652 implementations of their own. This is detailed in
11653 [crypto/engine/README.md](crypto/engine/README.md)
11654 as it couldn't be adequately described here. However, there are a few
11655 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11656 were changed in the original introduction of ENGINE code have now
11657 reverted back - the hooking from this code to ENGINE is now a good
11658 deal more passive and at run-time, operations deal directly with
11659 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11660 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11661 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11662 they were not being used by the framework as there is no concept of a
11663 BIGNUM_METHOD and they could not be generalised to the new
11664 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11665 ENGINE_cpy() has been removed as it cannot be consistently defined in
11670 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11674 * Change mkdef.pl to sort symbols that get the same entry number,
11675 and make sure the automatically generated functions `ERR_load_*`
11676 become part of libeay.num as well.
11680 * New function SSL_renegotiate_pending(). This returns true once
11681 renegotiation has been requested (either SSL_renegotiate() call
11682 or HelloRequest/ClientHello received from the peer) and becomes
11683 false once a handshake has been completed.
11684 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11685 sends a HelloRequest, but does not ensure that a handshake takes
11686 place. SSL_renegotiate_pending() is useful for checking if the
11687 client has followed the request.)
11691 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11692 By default, clients may request session resumption even during
11693 renegotiation (if session ID contexts permit); with this option,
11694 session resumption is possible only in the first handshake.
11696 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11697 more bits available for options that should not be part of
11698 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11702 * Add some demos for certificate and certificate request creation.
11706 * Make maximum certificate chain size accepted from the peer application
11707 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11708 "Douglas E. Engert" <deengert@anl.gov>.
11712 * Add support for shared libraries for Unixware-7
11713 (Boyd Lynn Gerber <gerberb@zenez.com>).
11717 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11718 be done prior to destruction. Use this to unload error strings from
11719 ENGINEs that load their own error strings. NB: This adds two new API
11720 functions to "get" and "set" this destroy handler in an ENGINE.
11724 * Alter all existing ENGINE implementations (except "openssl" and
11725 "openbsd") to dynamically instantiate their own error strings. This
11726 makes them more flexible to be built both as statically-linked ENGINEs
11727 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11728 Also, add stub code to each that makes building them as self-contained
11729 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11733 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11734 implementations into applications that are completely implemented in
11735 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11736 commands that can be used to configure what shared-library to load and
11737 to control aspects of the way it is handled. Also, made an update to
11738 the [README-Engine.md](README-Engine.md) file
11739 that brings its information up-to-date and
11740 provides some information and instructions on the "dynamic" ENGINE
11741 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11745 * Make it possible to unload ranges of ERR strings with a new
11746 "ERR_unload_strings" function.
11750 * Add a copy() function to EVP_MD.
11754 * Make EVP_MD routines take a context pointer instead of just the
11755 md_data void pointer.
11759 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11760 that the digest can only process a single chunk of data
11761 (typically because it is provided by a piece of
11762 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11763 is only going to provide a single chunk of data, and hence the
11764 framework needn't accumulate the data for oneshot drivers.
11768 * As with "ERR", make it possible to replace the underlying "ex_data"
11769 functions. This change also alters the storage and management of global
11770 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11771 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11772 index counters. The API functions that use this state have been changed
11773 to take a "class_index" rather than pointers to the class's local STACK
11774 and counter, and there is now an API function to dynamically create new
11775 classes. This centralisation allows us to (a) plug a lot of the
11776 thread-safety problems that existed, and (b) makes it possible to clean
11777 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11778 such data would previously have always leaked in application code and
11779 workarounds were in place to make the memory debugging turn a blind eye
11780 to it. Application code that doesn't use this new function will still
11781 leak as before, but their memory debugging output will announce it now
11782 rather than letting it slide.
11784 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11785 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11786 has a return value to indicate success or failure.
11790 * Make it possible to replace the underlying "ERR" functions such that the
11791 global state (2 LHASH tables and 2 locks) is only used by the "default"
11792 implementation. This change also adds two functions to "get" and "set"
11793 the implementation prior to it being automatically set the first time
11794 any other ERR function takes place. Ie. an application can call "get",
11795 pass the return value to a module it has just loaded, and that module
11796 can call its own "set" function using that value. This means the
11797 module's "ERR" operations will use (and modify) the error state in the
11798 application and not in its own statically linked copy of OpenSSL code.
11802 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11803 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11804 the operation, and provides a more encapsulated way for external code
11805 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11806 to use these functions rather than manually incrementing the counts.
11808 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11812 * Add EVP test program.
11816 * Add symmetric cipher support to ENGINE. Expect the API to change!
11820 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11821 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11822 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11823 These allow a CRL to be built without having to access X509_CRL fields
11824 directly. Modify 'ca' application to use new functions.
11828 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11829 bug workarounds. Rollback attack detection is a security feature.
11830 The problem will only arise on OpenSSL servers when TLSv1 is not
11831 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11832 Software authors not wanting to support TLSv1 will have special reasons
11833 for their choice and can explicitly enable this option.
11835 *Bodo Moeller, Lutz Jaenicke*
11837 * Rationalise EVP so it can be extended: don't include a union of
11838 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11839 (similar to those existing for EVP_CIPHER_CTX).
11844 EVP_MD_CTX_init(&md); /* new function call */
11845 EVP_DigestInit(&md, EVP_sha1());
11846 EVP_DigestUpdate(&md, in, len);
11847 EVP_DigestFinal(&md, out, NULL);
11848 EVP_MD_CTX_cleanup(&md); /* new function call */
11852 * Make DES key schedule conform to the usual scheme, as well as
11853 correcting its structure. This means that calls to DES functions
11854 now have to pass a pointer to a des_key_schedule instead of a
11855 plain des_key_schedule (which was actually always a pointer
11858 des_key_schedule ks;
11860 des_set_key_checked(..., &ks);
11861 des_ncbc_encrypt(..., &ks, ...);
11863 (Note that a later change renames 'des_...' into 'DES_...'.)
11867 * Initial reduction of linker bloat: the use of some functions, such as
11868 PEM causes large amounts of unused functions to be linked in due to
11869 poor organisation. For example pem_all.c contains every PEM function
11870 which has a knock on effect of linking in large amounts of (unused)
11871 ASN1 code. Grouping together similar functions and splitting unrelated
11872 functions prevents this.
11876 * Cleanup of EVP macros.
11880 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11881 correct `_ecb suffix`.
11885 * Add initial OCSP responder support to ocsp application. The
11886 revocation information is handled using the text based index
11887 use by the ca application. The responder can either handle
11888 requests generated internally, supplied in files (for example
11889 via a CGI script) or using an internal minimal server.
11893 * Add configuration choices to get zlib compression for TLS.
11897 * Changes to Kerberos SSL for RFC 2712 compliance:
11898 1. Implemented real KerberosWrapper, instead of just using
11899 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11900 2. Implemented optional authenticator field of KerberosWrapper.
11902 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11903 and authenticator structs; see crypto/krb5/.
11905 Generalized Kerberos calls to support multiple Kerberos libraries.
11906 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11907 via Richard Levitte*
11909 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11910 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11911 values for each of the key sizes rather than having just
11912 parameters (and 'speed' generating keys each time).
11916 * Speed up EVP routines.
11919 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11920 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11921 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11922 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11924 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11925 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11926 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11929 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11931 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11935 * Added the OS2-EMX target.
11937 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11939 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11940 to support NCONF routines in extension code. New function CONF_set_nconf()
11941 to allow functions which take an NCONF to also handle the old LHASH
11942 structure: this means that the old CONF compatible routines can be
11943 retained (in particular wrt extensions) without having to duplicate the
11944 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11948 * Enhance the general user interface with mechanisms for inner control
11949 and with possibilities to have yes/no kind of prompts.
11953 * Change all calls to low level digest routines in the library and
11954 applications to use EVP. Add missing calls to HMAC_cleanup() and
11955 don't assume HMAC_CTX can be copied using memcpy().
11957 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11959 * Add the possibility to control engines through control names but with
11960 arbitrary arguments instead of just a string.
11961 Change the key loaders to take a UI_METHOD instead of a callback
11962 function pointer. NOTE: this breaks binary compatibility with earlier
11963 versions of OpenSSL [engine].
11964 Adapt the nCipher code for these new conditions and add a card insertion
11969 * Enhance the general user interface with mechanisms to better support
11970 dialog box interfaces, application-defined prompts, the possibility
11971 to use defaults (for example default passwords from somewhere else)
11972 and interrupts/cancellations.
11976 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11977 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11981 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11982 tidy up some unnecessarily weird code in 'sk_new()').
11984 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11986 * Change the key loading routines for ENGINEs to use the same kind
11987 callback (pem_password_cb) as all other routines that need this
11992 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11993 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11994 than this minimum value is recommended.
11998 * New random seeder for OpenVMS, using the system process statistics
11999 that are easily reachable.
12003 * Windows apparently can't transparently handle global
12004 variables defined in DLLs. Initialisations such as:
12006 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12008 won't compile. This is used by the any applications that need to
12009 declare their own ASN1 modules. This was fixed by adding the option
12010 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12011 needed for static libraries under Win32.
12015 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12016 setting of purpose and trust fields. New X509_STORE trust and
12017 purpose functions and tidy up setting in other SSL functions.
12021 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12022 structure. These are inherited by X509_STORE_CTX when it is
12023 initialised. This allows various defaults to be set in the
12024 X509_STORE structure (such as flags for CRL checking and custom
12025 purpose or trust settings) for functions which only use X509_STORE_CTX
12026 internally such as S/MIME.
12028 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12029 trust settings if they are not set in X509_STORE. This allows X509_STORE
12030 purposes and trust (in S/MIME for example) to override any set by default.
12032 Add command line options for CRL checking to smime, s_client and s_server
12037 * Initial CRL based revocation checking. If the CRL checking flag(s)
12038 are set then the CRL is looked up in the X509_STORE structure and
12039 its validity and signature checked, then if the certificate is found
12040 in the CRL the verify fails with a revoked error.
12042 Various new CRL related callbacks added to X509_STORE_CTX structure.
12044 Command line options added to 'verify' application to support this.
12046 This needs some additional work, such as being able to handle multiple
12047 CRLs with different times, extension based lookup (rather than just
12048 by subject name) and ultimately more complete V2 CRL extension
12053 * Add a general user interface API (crypto/ui/). This is designed
12054 to replace things like des_read_password and friends (backward
12055 compatibility functions using this new API are provided).
12056 The purpose is to remove prompting functions from the DES code
12057 section as well as provide for prompting through dialog boxes in
12058 a window system and the like.
12062 * Add "ex_data" support to ENGINE so implementations can add state at a
12063 per-structure level rather than having to store it globally.
12067 * Make it possible for ENGINE structures to be copied when retrieved by
12068 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12069 This causes the "original" ENGINE structure to act like a template,
12070 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12071 operational state can be localised to each ENGINE structure, despite the
12072 fact they all share the same "methods". New ENGINE structures returned in
12073 this case have no functional references and the return value is the single
12074 structural reference. This matches the single structural reference returned
12075 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12080 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12081 needs to match any other type at all we need to manually clear the
12086 * Changes to the "openssl engine" utility to include;
12087 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12088 about an ENGINE's available control commands.
12089 - executing control commands from command line arguments using the
12090 '-pre' and '-post' switches. '-post' is only used if '-t' is
12091 specified and the ENGINE is successfully initialised. The syntax for
12092 the individual commands are colon-separated, for example;
12093 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12097 * New dynamic control command support for ENGINEs. ENGINEs can now
12098 declare their own commands (numbers), names (strings), descriptions,
12099 and input types for run-time discovery by calling applications. A
12100 subset of these commands are implicitly classed as "executable"
12101 depending on their input type, and only these can be invoked through
12102 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12103 can be based on user input, config files, etc). The distinction is
12104 that "executable" commands cannot return anything other than a boolean
12105 result and can only support numeric or string input, whereas some
12106 discoverable commands may only be for direct use through
12107 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12108 pointers, or other custom uses. The "executable" commands are to
12109 support parameterisations of ENGINE behaviour that can be
12110 unambiguously defined by ENGINEs and used consistently across any
12111 OpenSSL-based application. Commands have been added to all the
12112 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12113 control over shared-library paths without source code alterations.
12117 * Changed all ENGINE implementations to dynamically allocate their
12118 ENGINEs rather than declaring them statically. Apart from this being
12119 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12120 this also allows the implementations to compile without using the
12121 internal engine_int.h header.
12125 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12126 'const' value. Any code that should be able to modify a RAND_METHOD
12127 should already have non-const pointers to it (ie. they should only
12128 modify their own ones).
12132 * Made a variety of little tweaks to the ENGINE code.
12133 - "atalla" and "ubsec" string definitions were moved from header files
12134 to C code. "nuron" string definitions were placed in variables
12135 rather than hard-coded - allowing parameterisation of these values
12136 later on via ctrl() commands.
12137 - Removed unused "#if 0"'d code.
12138 - Fixed engine list iteration code so it uses ENGINE_free() to release
12139 structural references.
12140 - Constified the RAND_METHOD element of ENGINE structures.
12141 - Constified various get/set functions as appropriate and added
12142 missing functions (including a catch-all ENGINE_cpy that duplicates
12143 all ENGINE values onto a new ENGINE except reference counts/state).
12144 - Removed NULL parameter checks in get/set functions. Setting a method
12145 or function to NULL is a way of cancelling out a previously set
12146 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12147 and doesn't justify the extra error symbols and code.
12148 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12149 flags from engine_int.h to engine.h.
12150 - Changed prototypes for ENGINE handler functions (init(), finish(),
12151 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12155 * Implement binary inversion algorithm for BN_mod_inverse in addition
12156 to the algorithm using long division. The binary algorithm can be
12157 used only if the modulus is odd. On 32-bit systems, it is faster
12158 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12159 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12160 up to 450 bits. In 64-bit environments, the binary algorithm
12161 appears to be advantageous for much longer moduli; here we use it
12162 for moduli up to 2048 bits.
12166 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12167 could not support the combine flag in choice fields.
12171 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12172 extensions from a certificate request to the certificate.
12176 * Allow multiple 'certopt' and 'nameopt' options to be separated
12177 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12178 file: this allows the display of the certificate about to be
12179 signed to be customised, to allow certain fields to be included
12180 or excluded and extension details. The old system didn't display
12181 multicharacter strings properly, omitted fields not in the policy
12182 and couldn't display additional details such as extensions.
12186 * Function EC_POINTs_mul for multiple scalar multiplication
12187 of an arbitrary number of elliptic curve points
12188 \sum scalars[i]*points[i],
12189 optionally including the generator defined for the EC_GROUP:
12190 scalar*generator + \sum scalars[i]*points[i].
12192 EC_POINT_mul is a simple wrapper function for the typical case
12193 that the point list has just one item (besides the optional
12198 * First EC_METHODs for curves over GF(p):
12200 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12201 operations and provides various method functions that can also
12202 operate with faster implementations of modular arithmetic.
12204 EC_GFp_mont_method() reuses most functions that are part of
12205 EC_GFp_simple_method, but uses Montgomery arithmetic.
12207 *Bodo Moeller; point addition and point doubling
12208 implementation directly derived from source code provided by
12209 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12211 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12212 crypto/ec/ec_lib.c):
12214 Curves are EC_GROUP objects (with an optional group generator)
12215 based on EC_METHODs that are built into the library.
12217 Points are EC_POINT objects based on EC_GROUP objects.
12219 Most of the framework would be able to handle curves over arbitrary
12220 finite fields, but as there are no obvious types for fields other
12221 than GF(p), some functions are limited to that for now.
12225 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12226 that the file contains a complete HTTP response.
12230 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12231 change the def and num file printf format specifier from "%-40sXXX"
12232 to "%-39s XXX". The latter will always guarantee a space after the
12233 field while the former will cause them to run together if the field
12234 is 40 of more characters long.
12238 * Constify the cipher and digest 'method' functions and structures
12239 and modify related functions to take constant EVP_MD and EVP_CIPHER
12244 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12245 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12249 * Modify `EVP_Digest*()` routines so they now return values. Although the
12250 internal software routines can never fail additional hardware versions
12255 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12257 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12258 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12261 ERR_R_NESTED_ASN1_ERROR
12263 ERR_R_MISSING_ASN1_EOS
12264 were 4 .. 9, conflicting with
12265 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12267 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12268 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12270 Add new error code 'ERR_R_INTERNAL_ERROR'.
12274 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12279 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12280 sets the subject name for a new request or supersedes the
12281 subject name in a given request. Formats that can be parsed are
12282 'CN=Some Name, OU=myOU, C=IT'
12284 'CN=Some Name/OU=myOU/C=IT'.
12286 Add options '-batch' and '-verbose' to 'openssl req'.
12288 *Massimiliano Pala <madwolf@hackmasters.net>*
12290 * Introduce the possibility to access global variables through
12291 functions on platform were that's the best way to handle exporting
12292 global variables in shared libraries. To enable this functionality,
12293 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12294 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12295 is normally done by Configure or something similar).
12297 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12298 in the source file (foo.c) like this:
12300 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12301 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12303 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12304 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12306 OPENSSL_DECLARE_GLOBAL(int,foo);
12307 #define foo OPENSSL_GLOBAL_REF(foo)
12308 OPENSSL_DECLARE_GLOBAL(double,bar);
12309 #define bar OPENSSL_GLOBAL_REF(bar)
12311 The #defines are very important, and therefore so is including the
12312 header file everywhere where the defined globals are used.
12314 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12315 of ASN.1 items, but that structure is a bit different.
12317 The largest change is in util/mkdef.pl which has been enhanced with
12318 better and easier to understand logic to choose which symbols should
12319 go into the Windows .def files as well as a number of fixes and code
12320 cleanup (among others, algorithm keywords are now sorted
12321 lexicographically to avoid constant rewrites).
12325 * In BN_div() keep a copy of the sign of 'num' before writing the
12326 result to 'rm' because if rm==num the value will be overwritten
12327 and produce the wrong result if 'num' is negative: this caused
12328 problems with BN_mod() and BN_nnmod().
12332 * Function OCSP_request_verify(). This checks the signature on an
12333 OCSP request and verifies the signer certificate. The signer
12334 certificate is just checked for a generic purpose and OCSP request
12339 * Add OCSP_check_validity() function to check the validity of OCSP
12340 responses. OCSP responses are prepared in real time and may only
12341 be a few seconds old. Simply checking that the current time lies
12342 between thisUpdate and nextUpdate max reject otherwise valid responses
12343 caused by either OCSP responder or client clock inaccuracy. Instead
12344 we allow thisUpdate and nextUpdate to fall within a certain period of
12345 the current time. The age of the response can also optionally be
12346 checked. Two new options -validity_period and -status_age added to
12351 * If signature or public key algorithm is unrecognized print out its
12352 OID rather that just UNKNOWN.
12356 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12357 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12358 ID to be generated from the issuer certificate alone which can then be
12359 passed to OCSP_id_issuer_cmp().
12363 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12364 ASN1 modules to export functions returning ASN1_ITEM pointers
12365 instead of the ASN1_ITEM structures themselves. This adds several
12366 new macros which allow the underlying ASN1 function/structure to
12367 be accessed transparently. As a result code should not use ASN1_ITEM
12368 references directly (such as &X509_it) but instead use the relevant
12369 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12370 use of the new ASN1 code on platforms where exporting structures
12371 is problematical (for example in shared libraries) but exporting
12372 functions returning pointers to structures is not.
12376 * Add support for overriding the generation of SSL/TLS session IDs.
12377 These callbacks can be registered either in an SSL_CTX or per SSL.
12378 The purpose of this is to allow applications to control, if they wish,
12379 the arbitrary values chosen for use as session IDs, particularly as it
12380 can be useful for session caching in multiple-server environments. A
12381 command-line switch for testing this (and any client code that wishes
12382 to use such a feature) has been added to "s_server".
12384 *Geoff Thorpe, Lutz Jaenicke*
12386 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12387 of the form `#if defined(...) || defined(...) || ...` and
12388 `#if !defined(...) && !defined(...) && ...`. This also avoids
12389 the growing number of special cases it was previously handling.
12393 * Make all configuration macros available for application by making
12394 sure they are available in opensslconf.h, by giving them names starting
12395 with `OPENSSL_` to avoid conflicts with other packages and by making
12396 sure e_os2.h will cover all platform-specific cases together with
12398 Additionally, it is now possible to define configuration/platform-
12399 specific names (called "system identities"). In the C code, these
12400 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12401 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12402 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12407 * New option -set_serial to 'req' and 'x509' this allows the serial
12408 number to use to be specified on the command line. Previously self
12409 signed certificates were hard coded with serial number 0 and the
12410 CA options of 'x509' had to use a serial number in a file which was
12415 * New options to 'ca' utility to support V2 CRL entry extensions.
12416 Currently CRL reason, invalidity date and hold instruction are
12417 supported. Add new CRL extensions to V3 code and some new objects.
12421 * New function EVP_CIPHER_CTX_set_padding() this is used to
12422 disable standard block padding (aka PKCS#5 padding) in the EVP
12423 API, which was previously mandatory. This means that the data is
12424 not padded in any way and so the total length much be a multiple
12425 of the block size, otherwise an error occurs.
12429 * Initial (incomplete) OCSP SSL support.
12433 * New function OCSP_parse_url(). This splits up a URL into its host,
12434 port and path components: primarily to parse OCSP URLs. New -url
12435 option to ocsp utility.
12439 * New nonce behavior. The return value of OCSP_check_nonce() now
12440 reflects the various checks performed. Applications can decide
12441 whether to tolerate certain situations such as an absent nonce
12442 in a response when one was present in a request: the ocsp application
12443 just prints out a warning. New function OCSP_add1_basic_nonce()
12444 this is to allow responders to include a nonce in a response even if
12445 the request is nonce-less.
12449 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12450 skipped when using openssl x509 multiple times on a single input file,
12451 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12455 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12456 set string type: to handle setting ASN1_TIME structures. Fix ca
12457 utility to correctly initialize revocation date of CRLs.
12461 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12462 the clients preferred ciphersuites and rather use its own preferences.
12463 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12464 Internet Explorer by ensuring unchanged hash method during stepup.
12465 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12469 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12470 to aes and add a new 'exist' option to print out symbols that don't
12475 * Additional options to ocsp utility to allow flags to be set and
12476 additional certificates supplied.
12480 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12481 OCSP client a number of certificate to only verify the response
12486 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12487 handle the new API. Currently only ECB, CBC modes supported. Add new
12490 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12491 Encryption Standard (AES) Ciphersuites for Transport Layer
12492 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12493 not enabled by default and were not part of the "ALL" ciphersuite
12494 alias because they were not yet official; they could be
12495 explicitly requested by specifying the "AESdraft" ciphersuite
12496 group alias. In the final release of OpenSSL 0.9.7, the group
12497 alias is called "AES" and is part of "ALL".)
12499 *Ben Laurie, Steve Henson, Bodo Moeller*
12501 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12502 request to response.
12506 * Functions for OCSP responders. OCSP_request_onereq_count(),
12507 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12508 extract information from a certificate request. OCSP_response_create()
12509 creates a response and optionally adds a basic response structure.
12510 OCSP_basic_add1_status() adds a complete single response to a basic
12511 response and returns the OCSP_SINGLERESP structure just added (to allow
12512 extensions to be included for example). OCSP_basic_add1_cert() adds a
12513 certificate to a basic response and OCSP_basic_sign() signs a basic
12514 response with various flags. New helper functions ASN1_TIME_check()
12515 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12516 (converts ASN1_TIME to GeneralizedTime).
12520 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12521 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12522 structure from a certificate. X509_pubkey_digest() digests the public_key
12523 contents: this is used in various key identifiers.
12527 * Make sk_sort() tolerate a NULL argument.
12529 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12531 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12532 passed by the function are trusted implicitly. If any of them signed the
12533 response then it is assumed to be valid and is not verified.
12537 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12538 to data. This was previously part of the PKCS7 ASN1 code. This
12539 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12540 *Steve Henson, reported by Kenneth R. Robinette
12541 <support@securenetterm.com>*
12543 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12544 routines: without these tracing memory leaks is very painful.
12545 Fix leaks in PKCS12 and PKCS7 routines.
12549 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12550 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12551 effectively meant GeneralizedTime would never be used. Now it
12552 is initialised to -1 but X509_time_adj() now has to check the value
12553 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12554 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12555 *Steve Henson, reported by Kenneth R. Robinette
12556 <support@securenetterm.com>*
12558 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12559 result in a zero length in the ASN1_INTEGER structure which was
12560 not consistent with the structure when d2i_ASN1_INTEGER() was used
12561 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12562 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12563 where it did not print out a minus for negative ASN1_INTEGER.
12567 * Add summary printout to ocsp utility. The various functions which
12568 convert status values to strings have been renamed to:
12569 OCSP_response_status_str(), OCSP_cert_status_str() and
12570 OCSP_crl_reason_str() and are no longer static. New options
12571 to verify nonce values and to disable verification. OCSP response
12572 printout format cleaned up.
12576 * Add additional OCSP certificate checks. These are those specified
12577 in RFC2560. This consists of two separate checks: the CA of the
12578 certificate being checked must either be the OCSP signer certificate
12579 or the issuer of the OCSP signer certificate. In the latter case the
12580 OCSP signer certificate must contain the OCSP signing extended key
12581 usage. This check is performed by attempting to match the OCSP
12582 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12583 in the OCSP_CERTID structures of the response.
12587 * Initial OCSP certificate verification added to OCSP_basic_verify()
12588 and related routines. This uses the standard OpenSSL certificate
12589 verify routines to perform initial checks (just CA validity) and
12590 to obtain the certificate chain. Then additional checks will be
12591 performed on the chain. Currently the root CA is checked to see
12592 if it is explicitly trusted for OCSP signing. This is used to set
12593 a root CA as a global signing root: that is any certificate that
12594 chains to that CA is an acceptable OCSP signing certificate.
12598 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12599 extensions from a separate configuration file.
12600 As when reading extensions from the main configuration file,
12601 the '-extensions ...' option may be used for specifying the
12604 *Massimiliano Pala <madwolf@comune.modena.it>*
12606 * New OCSP utility. Allows OCSP requests to be generated or
12607 read. The request can be sent to a responder and the output
12608 parsed, outputted or printed in text form. Not complete yet:
12609 still needs to check the OCSP response validity.
12613 * New subcommands for 'openssl ca':
12614 `openssl ca -status <serial>` prints the status of the cert with
12615 the given serial number (according to the index file).
12616 `openssl ca -updatedb` updates the expiry status of certificates
12619 *Massimiliano Pala <madwolf@comune.modena.it>*
12621 * New '-newreq-nodes' command option to CA.pl. This is like
12622 '-newreq', but calls 'openssl req' with the '-nodes' option
12623 so that the resulting key is not encrypted.
12625 *Damien Miller <djm@mindrot.org>*
12627 * New configuration for the GNU Hurd.
12629 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12631 * Initial code to implement OCSP basic response verify. This
12632 is currently incomplete. Currently just finds the signer's
12633 certificate and verifies the signature on the response.
12637 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12638 value of OPENSSLDIR. This is available via the new '-d' option
12639 to 'openssl version', and is also included in 'openssl version -a'.
12643 * Allowing defining memory allocation callbacks that will be given
12644 file name and line number information in additional arguments
12645 (a `const char*` and an int). The basic functionality remains, as
12646 well as the original possibility to just replace malloc(),
12647 realloc() and free() by functions that do not know about these
12648 additional arguments. To register and find out the current
12649 settings for extended allocation functions, the following
12650 functions are provided:
12652 CRYPTO_set_mem_ex_functions
12653 CRYPTO_set_locked_mem_ex_functions
12654 CRYPTO_get_mem_ex_functions
12655 CRYPTO_get_locked_mem_ex_functions
12657 These work the same way as CRYPTO_set_mem_functions and friends.
12658 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12659 extended allocation function is enabled.
12660 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12661 a conventional allocation function is enabled.
12663 *Richard Levitte, Bodo Moeller*
12665 * Finish off removing the remaining LHASH function pointer casts.
12666 There should no longer be any prototype-casting required when using
12667 the LHASH abstraction, and any casts that remain are "bugs". See
12668 the callback types and macros at the head of lhash.h for details
12669 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12673 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12674 If /dev/[u]random devices are not available or do not return enough
12675 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12677 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12678 /etc/entropy will be queried once each in this sequence, querying stops
12679 when enough entropy was collected without querying more sockets.
12683 * Change the Unix RAND_poll() variant to be able to poll several
12684 random devices, as specified by DEVRANDOM, until a sufficient amount
12685 of data has been collected. We spend at most 10 ms on each file
12686 (select timeout) and read in non-blocking mode. DEVRANDOM now
12687 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12688 (previously it was just the string "/dev/urandom"), so on typical
12689 platforms the 10 ms delay will never occur.
12690 Also separate out the Unix variant to its own file, rand_unix.c.
12691 For VMS, there's a currently-empty rand_vms.c.
12695 * Move OCSP client related routines to ocsp_cl.c. These
12696 provide utility functions which an application needing
12697 to issue a request to an OCSP responder and analyse the
12698 response will typically need: as opposed to those which an
12699 OCSP responder itself would need which will be added later.
12701 OCSP_request_sign() signs an OCSP request with an API similar
12702 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12703 response. OCSP_response_get1_basic() extracts basic response
12704 from response. OCSP_resp_find_status(): finds and extracts status
12705 information from an OCSP_CERTID structure (which will be created
12706 when the request structure is built). These are built from lower
12707 level functions which work on OCSP_SINGLERESP structures but
12708 won't normally be used unless the application wishes to examine
12709 extensions in the OCSP response for example.
12711 Replace nonce routines with a pair of functions.
12712 OCSP_request_add1_nonce() adds a nonce value and optionally
12713 generates a random value. OCSP_check_nonce() checks the
12714 validity of the nonce in an OCSP response.
12718 * Change function OCSP_request_add() to OCSP_request_add0_id().
12719 This doesn't copy the supplied OCSP_CERTID and avoids the
12720 need to free up the newly created id. Change return type
12721 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12722 This can then be used to add extensions to the request.
12723 Deleted OCSP_request_new(), since most of its functionality
12724 is now in OCSP_REQUEST_new() (and the case insensitive name
12725 clash) apart from the ability to set the request name which
12726 will be added elsewhere.
12730 * Update OCSP API. Remove obsolete extensions argument from
12731 various functions. Extensions are now handled using the new
12732 OCSP extension code. New simple OCSP HTTP function which
12733 can be used to send requests and parse the response.
12737 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12738 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12739 uses the special reorder version of SET OF to sort the attributes
12740 and reorder them to match the encoded order. This resolves a long
12741 standing problem: a verify on a PKCS7 structure just after signing
12742 it used to fail because the attribute order did not match the
12743 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12744 it uses the received order. This is necessary to tolerate some broken
12745 software that does not order SET OF. This is handled by encoding
12746 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12747 to produce the required SET OF.
12751 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12752 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12753 files to get correct declarations of the ASN.1 item variables.
12757 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12758 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12759 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12760 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12761 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12762 ASN1_ITEM and no wrapper functions.
12766 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12767 replace the old function pointer based I/O routines. Change most of
12768 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12772 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12773 lines, recognize more "algorithms" that can be deselected, and make
12774 it complain about algorithm deselection that isn't recognised.
12778 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12779 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12780 to use new functions. Add NO_ASN1_OLD which can be set to remove
12781 some old style ASN1 functions: this can be used to determine if old
12782 code will still work when these eventually go away.
12786 * New extension functions for OCSP structures, these follow the
12787 same conventions as certificates and CRLs.
12791 * New function X509V3_add1_i2d(). This automatically encodes and
12792 adds an extension. Its behaviour can be customised with various
12793 flags to append, replace or delete. Various wrappers added for
12794 certificates and CRLs.
12798 * Fix to avoid calling the underlying ASN1 print routine when
12799 an extension cannot be parsed. Correct a typo in the
12800 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12804 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12805 entries for variables.
12809 * Add functionality to apps/openssl.c for detecting locking
12810 problems: As the program is single-threaded, all we have
12811 to do is register a locking callback using an array for
12812 storing which locks are currently held by the program.
12816 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12817 SSL_get_ex_data_X509_STORE_idx(), which is used in
12818 ssl_verify_cert_chain() and thus can be called at any time
12819 during TLS/SSL handshakes so that thread-safety is essential.
12820 Unfortunately, the ex_data design is not at all suited
12821 for multi-threaded use, so it probably should be abolished.
12825 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12827 *Broadcom, tweaked and integrated by Geoff Thorpe*
12829 * Move common extension printing code to new function
12830 X509V3_print_extensions(). Reorganise OCSP print routines and
12831 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12835 * New function X509_signature_print() to remove duplication in some
12840 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12841 set (this was treated exactly the same as SET OF previously). This
12842 is used to reorder the STACK representing the structure to match the
12843 encoding. This will be used to get round a problem where a PKCS7
12844 structure which was signed could not be verified because the STACK
12845 order did not reflect the encoded order.
12849 * Reimplement the OCSP ASN1 module using the new code.
12853 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12854 for its ASN1 operations. The old style function pointers still exist
12855 for now but they will eventually go away.
12859 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12860 completely replaces the old ASN1 functionality with a table driven
12861 encoder and decoder which interprets an ASN1_ITEM structure describing
12862 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12863 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12864 has also been converted to the new form.
12868 * Change BN_mod_exp_recp so that negative moduli are tolerated
12869 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12870 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12871 for negative moduli.
12875 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12876 of not touching the result's sign bit.
12880 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12885 * Changed the LHASH code to use prototypes for callbacks, and created
12886 macros to declare and implement thin (optionally static) functions
12887 that provide type-safety and avoid function pointer casting for the
12888 type-specific callbacks.
12892 * Added Kerberos Cipher Suites to be used with TLS, as written in
12894 *Veers Staats <staatsvr@asc.hpc.mil>,
12895 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12897 * Reformat the FAQ so the different questions and answers can be divided
12898 in sections depending on the subject.
12902 * Have the zlib compression code load ZLIB.DLL dynamically under
12907 * New function BN_mod_sqrt for computing square roots modulo a prime
12908 (using the probabilistic Tonelli-Shanks algorithm unless
12909 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12910 be handled deterministically).
12912 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12914 * Make BN_mod_inverse faster by explicitly handling small quotients
12915 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12916 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12920 * New function BN_kronecker.
12924 * Fix BN_gcd so that it works on negative inputs; the result is
12925 positive unless both parameters are zero.
12926 Previously something reasonably close to an infinite loop was
12927 possible because numbers could be growing instead of shrinking
12928 in the implementation of Euclid's algorithm.
12932 * Fix BN_is_word() and BN_is_one() macros to take into account the
12933 sign of the number in question.
12935 Fix BN_is_word(a,w) to work correctly for w == 0.
12937 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12938 because its test if the absolute value of 'a' equals 'w'.
12939 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12940 it exists mostly for use in the implementations of BN_is_zero(),
12941 BN_is_one(), and BN_is_word().
12945 * New function BN_swap.
12949 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12950 the exponentiation functions are more likely to produce reasonable
12951 results on negative inputs.
12955 * Change BN_mod_mul so that the result is always non-negative.
12956 Previously, it could be negative if one of the factors was negative;
12957 I don't think anyone really wanted that behaviour.
12961 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
12962 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
12963 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
12964 and add new functions:
12973 BN_mod_lshift1_quick
12975 BN_mod_lshift_quick
12977 These functions always generate non-negative results.
12979 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
12980 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
12982 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
12983 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
12984 be reduced modulo `m`.
12986 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12989 The following entry accidentally appeared in the CHANGES file
12990 distributed with OpenSSL 0.9.7. The modifications described in
12991 it do *not* apply to OpenSSL 0.9.7.
12993 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12994 was actually never needed) and in BN_mul(). The removal in BN_mul()
12995 required a small change in bn_mul_part_recursive() and the addition
12996 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12997 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12998 bn_sub_words() and bn_add_words() except they take arrays with
13004 * In 'openssl passwd', verify passwords read from the terminal
13005 unless the '-salt' option is used (which usually means that
13006 verification would just waste user's time since the resulting
13007 hash is going to be compared with some given password hash)
13008 or the new '-noverify' option is used.
13010 This is an incompatible change, but it does not affect
13011 non-interactive use of 'openssl passwd' (passwords on the command
13012 line, '-stdin' option, '-in ...' option) and thus should not
13013 cause any problems.
13017 * Remove all references to RSAref, since there's no more need for it.
13021 * Make DSO load along a path given through an environment variable
13022 (SHLIB_PATH) with shl_load().
13026 * Constify the ENGINE code as a result of BIGNUM constification.
13027 Also constify the RSA code and most things related to it. In a
13028 few places, most notable in the depth of the ASN.1 code, ugly
13029 casts back to non-const were required (to be solved at a later
13034 * Make it so the openssl application has all engines loaded by default.
13038 * Constify the BIGNUM routines a little more.
13042 * Add the following functions:
13044 ENGINE_load_cswift()
13046 ENGINE_load_atalla()
13047 ENGINE_load_nuron()
13048 ENGINE_load_builtin_engines()
13050 That way, an application can itself choose if external engines that
13051 are built-in in OpenSSL shall ever be used or not. The benefit is
13052 that applications won't have to be linked with libdl or other dso
13053 libraries unless it's really needed.
13055 Changed 'openssl engine' to load all engines on demand.
13056 Changed the engine header files to avoid the duplication of some
13057 declarations (they differed!).
13061 * 'openssl engine' can now list capabilities.
13065 * Better error reporting in 'openssl engine'.
13069 * Never call load_dh_param(NULL) in s_server.
13073 * Add engine application. It can currently list engines by name and
13074 identity, and test if they are actually available.
13078 * Improve RPM specification file by forcing symbolic linking and making
13079 sure the installed documentation is also owned by root.root.
13081 *Damien Miller <djm@mindrot.org>*
13083 * Give the OpenSSL applications more possibilities to make use of
13084 keys (public as well as private) handled by engines.
13088 * Add OCSP code that comes from CertCo.
13092 * Add VMS support for the Rijndael code.
13096 * Added untested support for Nuron crypto accelerator.
13100 * Add support for external cryptographic devices. This code was
13101 previously distributed separately as the "engine" branch.
13103 *Geoff Thorpe, Richard Levitte*
13105 * Rework the filename-translation in the DSO code. It is now possible to
13106 have far greater control over how a "name" is turned into a filename
13107 depending on the operating environment and any oddities about the
13108 different shared library filenames on each system.
13112 * Support threads on FreeBSD-elf in Configure.
13116 * Fix for SHA1 assembly problem with MASM: it produces
13117 warnings about corrupt line number information when assembling
13118 with debugging information. This is caused by the overlapping
13121 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13124 NCONF_get_number() has no error checking at all. As a replacement,
13125 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13126 promoted strongly. The old NCONF_get_number is kept around for
13127 binary backward compatibility.
13128 Make it possible for methods to load from something other than a BIO,
13129 by providing a function pointer that is given a name instead of a BIO.
13130 For example, this could be used to load configuration data from an
13135 * Fix for non blocking accept BIOs. Added new I/O special reason
13136 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13137 with non blocking I/O was not possible because no retry code was
13138 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13143 * Added the beginnings of Rijndael support.
13147 * Fix for bug in DirectoryString mask setting. Add support for
13148 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13149 to allow certificate printing to more controllable, additional
13150 'certopt' option to 'x509' to allow new printing options to be
13155 * Clean old EAY MD5 hack from e_os.h.
13159 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13161 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13162 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13164 *Joe Orton, Steve Henson*
13166 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13168 * Fix additional bug revealed by the NISCC test suite:
13170 Stop bug triggering large recursion when presented with
13171 certain ASN.1 tags [CVE-2003-0851][]
13175 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13177 * Fix various bugs revealed by running the NISCC test suite:
13179 Stop out of bounds reads in the ASN1 code when presented with
13180 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13182 If verify callback ignores invalid public key errors don't try to check
13183 certificate signature with the NULL public key.
13187 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13188 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13193 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13194 extra data after the compression methods not only for TLS 1.0
13195 but also for SSL 3.0 (as required by the specification).
13197 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13199 * Change X509_certificate_type() to mark the key as exported/exportable
13200 when it's 512 *bits* long, not 512 bytes.
13204 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13206 * Countermeasure against the Klima-Pokorny-Rosa extension of
13207 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13208 a protocol version number mismatch like a decryption error
13209 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13213 * Turn on RSA blinding by default in the default implementation
13214 to avoid a timing attack. Applications that don't want it can call
13215 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13216 They would be ill-advised to do so in most cases.
13218 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13220 * Change RSA blinding code so that it works when the PRNG is not
13221 seeded (in this case, the secret RSA exponent is abused as
13222 an unpredictable seed -- if it is not unpredictable, there
13223 is no point in blinding anyway). Make RSA blinding thread-safe
13224 by remembering the creator's thread ID in rsa->blinding and
13225 having all other threads use local one-time blinding factors
13226 (this requires more computation than sharing rsa->blinding, but
13227 avoids excessive locking; and if an RSA object is not shared
13228 between threads, blinding will still be very fast).
13232 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13234 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13235 via timing by performing a MAC computation even if incorrect
13236 block cipher padding has been found. This is a countermeasure
13237 against active attacks where the attacker has to distinguish
13238 between bad padding and a MAC verification error. [CVE-2003-0078][]
13240 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13241 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13242 Martin Vuagnoux (EPFL, Ilion)*
13244 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13246 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13247 memory from its contents. This is done with a counter that will
13248 place alternating values in each byte. This can be used to solve
13249 two issues: 1) the removal of calls to memset() by highly optimizing
13250 compilers, and 2) cleansing with other values than 0, since those can
13251 be read through on certain media, for example a swap space on disk.
13255 * Bugfix: client side session caching did not work with external caching,
13256 because the session->cipher setting was not restored when reloading
13257 from the external cache. This problem was masked, when
13258 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13259 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13263 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13264 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13266 *Zeev Lieber <zeev-l@yahoo.com>*
13268 * Undo an undocumented change introduced in 0.9.6e which caused
13269 repeated calls to OpenSSL_add_all_ciphers() and
13270 OpenSSL_add_all_digests() to be ignored, even after calling
13275 * Change the default configuration reader to deal with last line not
13276 being properly terminated.
13280 * Change X509_NAME_cmp() so it applies the special rules on handling
13281 DN values that are of type PrintableString, as well as RDNs of type
13282 emailAddress where the value has the type ia5String.
13284 *stefank@valicert.com via Richard Levitte*
13286 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13287 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13288 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13289 the bitwise-OR of the two for use by the majority of applications
13290 wanting this behaviour, and update the docs. The documented
13291 behaviour and actual behaviour were inconsistent and had been
13292 changing anyway, so this is more a bug-fix than a behavioural
13295 *Geoff Thorpe, diagnosed by Nadav Har'El*
13297 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13298 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13302 * Fix initialization code race conditions in
13303 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13304 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13305 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13306 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13307 ssl2_get_cipher_by_char(),
13308 ssl3_get_cipher_by_char().
13310 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13312 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13313 the cached sessions are flushed, as the remove_cb() might use ex_data
13314 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13315 (see [openssl.org #212]).
13317 *Geoff Thorpe, Lutz Jaenicke*
13319 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13320 length, instead of the encoding length to d2i_ASN1_OBJECT.
13324 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13326 * [In 0.9.6g-engine release:]
13327 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13329 *Lynn Gazis <lgazis@rainbow.com>*
13331 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13333 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13334 and get fix the header length calculation.
13335 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13336 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13338 * Use proper error handling instead of 'assertions' in buffer
13339 overflow checks added in 0.9.6e. This prevents DoS (the
13340 assertions could call abort()).
13342 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13344 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13346 * Add various sanity checks to asn1_get_length() to reject
13347 the ASN1 length bytes if they exceed sizeof(long), will appear
13348 negative or the content length exceeds the length of the
13351 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13353 * Fix cipher selection routines: ciphers without encryption had no flags
13354 for the cipher strength set and where therefore not handled correctly
13355 by the selection routines (PR #130).
13359 * Fix EVP_dsa_sha macro.
13364 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13365 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13366 that was added in OpenSSL 0.9.6d.
13368 As the countermeasure turned out to be incompatible with some
13369 broken SSL implementations, the new option is part of SSL_OP_ALL.
13370 SSL_OP_ALL is usually employed when compatibility with weird SSL
13371 implementations is desired (e.g. '-bugs' option to 's_client' and
13372 's_server'), so the new option is automatically set in many
13377 * Changes in security patch:
13379 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13380 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13381 Air Force Materiel Command, USAF, under agreement number
13384 * Add various sanity checks to asn1_get_length() to reject
13385 the ASN1 length bytes if they exceed sizeof(long), will appear
13386 negative or the content length exceeds the length of the
13387 supplied buffer. [CVE-2002-0659][]
13389 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13391 * Assertions for various potential buffer overflows, not known to
13392 happen in practice.
13394 *Ben Laurie (CHATS)*
13396 * Various temporary buffers to hold ASCII versions of integers were
13397 too small for 64 bit platforms. [CVE-2002-0655][]
13398 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13400 * Remote buffer overflow in SSL3 protocol - an attacker could
13401 supply an oversized session ID to a client. [CVE-2002-0656][]
13403 *Ben Laurie (CHATS)*
13405 * Remote buffer overflow in SSL2 protocol - an attacker could
13406 supply an oversized client master key. [CVE-2002-0656][]
13408 *Ben Laurie (CHATS)*
13410 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13412 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13413 encoded as NULL) with id-dsa-with-sha1.
13415 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13417 * Check various `X509_...()` return values in apps/req.c.
13419 *Nils Larsch <nla@trustcenter.de>*
13421 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13422 an end-of-file condition would erroneously be flagged, when the CRLF
13423 was just at the end of a processed block. The bug was discovered when
13424 processing data through a buffering memory BIO handing the data to a
13425 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13426 <ptsekov@syntrex.com> and Nedelcho Stanev.
13430 * Implement a countermeasure against a vulnerability recently found
13431 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13432 before application data chunks to avoid the use of known IVs
13433 with data potentially chosen by the attacker.
13437 * Fix length checks in ssl3_get_client_hello().
13441 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13442 to prevent ssl3_read_internal() from incorrectly assuming that
13443 ssl3_read_bytes() found application data while handshake
13444 processing was enabled when in fact s->s3->in_read_app_data was
13445 merely automatically cleared during the initial handshake.
13447 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13449 * Fix object definitions for Private and Enterprise: they were not
13450 recognized in their shortname (=lowercase) representation. Extend
13451 obj_dat.pl to issue an error when using undefined keywords instead
13452 of silently ignoring the problem (Svenning Sorensen
13453 <sss@sss.dnsalias.net>).
13457 * Fix DH_generate_parameters() so that it works for 'non-standard'
13458 generators, i.e. generators other than 2 and 5. (Previously, the
13459 code did not properly initialise the 'add' and 'rem' values to
13460 BN_generate_prime().)
13462 In the new general case, we do not insist that 'generator' is
13463 actually a primitive root: This requirement is rather pointless;
13464 a generator of the order-q subgroup is just as good, if not
13469 * Map new X509 verification errors to alerts. Discovered and submitted by
13470 Tom Wu <tom@arcot.com>.
13474 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13475 returning non-zero before the data has been completely received
13476 when using non-blocking I/O.
13478 *Bodo Moeller; problem pointed out by John Hughes*
13480 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13482 *Ben Laurie, Lutz Jaenicke*
13484 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13485 Yoram Zahavi <YoramZ@gilian.com>).
13489 * Add information about CygWin 1.3 and on, and preserve proper
13490 configuration for the versions before that.
13492 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13494 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13495 check whether we deal with a copy of a session and do not delete from
13496 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13497 <izhar@checkpoint.com>.
13501 * Do not store session data into the internal session cache, if it
13502 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13503 flag is set). Proposed by Aslam <aslam@funk.com>.
13507 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13512 * [In 0.9.6d-engine release:]
13513 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13515 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13517 * Add the configuration target linux-s390x.
13519 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13521 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13522 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13523 variable as an indication that a ClientHello message has been
13524 received. As the flag value will be lost between multiple
13525 invocations of ssl3_accept when using non-blocking I/O, the
13526 function may not be aware that a handshake has actually taken
13527 place, thus preventing a new session from being added to the
13530 To avoid this problem, we now set s->new_session to 2 instead of
13531 using a local variable.
13533 *Lutz Jaenicke, Bodo Moeller*
13535 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13536 if the SSL_R_LENGTH_MISMATCH error is detected.
13538 *Geoff Thorpe, Bodo Moeller*
13540 * New 'shared_ldflag' column in Configure platform table.
13544 * Fix EVP_CIPHER_mode macro.
13546 *"Dan S. Camper" <dan@bti.net>*
13548 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13549 type, we must throw them away by setting rr->length to 0.
13551 *D P Chang <dpc@qualys.com>*
13553 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13555 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13556 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13557 worked incorrectly for those cases where range = `10..._2` and
13558 `3*range` is two bits longer than range.)
13562 * Only add signing time to PKCS7 structures if it is not already
13567 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13568 OBJ_ld_ce should be OBJ_id_ce.
13569 Also some ip-pda OIDs in crypto/objects/objects.txt were
13570 incorrect (cf. RFC 3039).
13572 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13574 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13575 returns early because it has nothing to do.
13577 *Andy Schneider <andy.schneider@bjss.co.uk>*
13579 * [In 0.9.6c-engine release:]
13580 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13582 *Andy Schneider <andy.schneider@bjss.co.uk>*
13584 * [In 0.9.6c-engine release:]
13585 Add support for Cryptographic Appliance's keyserver technology.
13586 (Use engine 'keyclient')
13588 *Cryptographic Appliances and Geoff Thorpe*
13590 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13591 is called via tools/c89.sh because arguments have to be
13592 rearranged (all '-L' options must appear before the first object
13595 *Richard Shapiro <rshapiro@abinitio.com>*
13597 * [In 0.9.6c-engine release:]
13598 Add support for Broadcom crypto accelerator cards, backported
13601 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13603 * [In 0.9.6c-engine release:]
13604 Add support for SureWare crypto accelerator cards from
13605 Baltimore Technologies. (Use engine 'sureware')
13607 *Baltimore Technologies and Mark Cox*
13609 * [In 0.9.6c-engine release:]
13610 Add support for crypto accelerator cards from Accelerated
13611 Encryption Processing, www.aep.ie. (Use engine 'aep')
13613 *AEP Inc. and Mark Cox*
13615 * Add a configuration entry for gcc on UnixWare.
13617 *Gary Benson <gbenson@redhat.com>*
13619 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13620 messages are stored in a single piece (fixed-length part and
13621 variable-length part combined) and fix various bugs found on the way.
13625 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13626 instead. BIO_gethostbyname() does not know what timeouts are
13627 appropriate, so entries would stay in cache even when they have
13629 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13631 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13632 faced with a pathologically small ClientHello fragment that does
13633 not contain client_version: Instead of aborting with an error,
13634 simply choose the highest available protocol version (i.e.,
13635 TLS 1.0 unless it is disabled). In practice, ClientHello
13636 messages are never sent like this, but this change gives us
13637 strictly correct behaviour at least for TLS.
13641 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13642 never resets s->method to s->ctx->method when called from within
13643 one of the SSL handshake functions.
13645 *Bodo Moeller; problem pointed out by Niko Baric*
13647 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13648 (sent using the client's version number) if client_version is
13649 smaller than the protocol version in use. Also change
13650 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13651 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13652 the client will at least see that alert.
13656 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13661 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13662 client receives HelloRequest while in a handshake.
13664 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13666 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13667 should end in 'break', not 'goto end' which circumvents various
13668 cleanups done in state SSL_ST_OK. But session related stuff
13669 must be disabled for SSL_ST_OK in the case that we just sent a
13672 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13673 before just sending a HelloRequest.
13675 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13677 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13678 reveal whether illegal block cipher padding was found or a MAC
13679 verification error occurred. (Neither SSLerr() codes nor alerts
13680 are directly visible to potential attackers, but the information
13681 may leak via logfiles.)
13683 Similar changes are not required for the SSL 2.0 implementation
13684 because the number of padding bytes is sent in clear for SSL 2.0,
13685 and the extra bytes are just ignored. However ssl/s2_pkt.c
13686 failed to verify that the purported number of padding bytes is in
13691 * Add OpenUNIX-8 support including shared libraries
13692 (Boyd Lynn Gerber <gerberb@zenez.com>).
13696 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13697 'wristwatch attack' using huge encoding parameters (cf.
13698 James H. Manger's CRYPTO 2001 paper). Note that the
13699 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13700 encoding parameters and hence was not vulnerable.
13704 * BN_sqr() bug fix.
13706 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13708 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13709 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13710 followed by modular reduction.
13712 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13714 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13715 equivalent based on BN_pseudo_rand() instead of BN_rand().
13719 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13720 This function was broken, as the check for a new client hello message
13721 to handle SGC did not allow these large messages.
13722 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13726 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13730 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13731 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13735 * Rework the configuration and shared library support for Tru64 Unix.
13736 The configuration part makes use of modern compiler features and
13737 still retains old compiler behavior for those that run older versions
13738 of the OS. The shared library support part includes a variant that
13739 uses the RPATH feature, and is available through the special
13740 configuration target "alpha-cc-rpath", which will never be selected
13743 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13745 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13746 with the same message size as in ssl3_get_certificate_request().
13747 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13748 messages might inadvertently be reject as too long.
13750 *Petr Lampa <lampa@fee.vutbr.cz>*
13752 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13756 * Modified SSL library such that the verify_callback that has been set
13757 specifically for an SSL object with SSL_set_verify() is actually being
13758 used. Before the change, a verify_callback set with this function was
13759 ignored and the verify_callback() set in the SSL_CTX at the time of
13760 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13761 to allow the necessary settings.
13765 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13766 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13767 done automatically (in contradiction to the requirements of the C
13768 standard). This made problems when used from OpenSSH.
13772 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13773 dh->length and always used
13775 BN_rand_range(priv_key, dh->p).
13777 BN_rand_range() is not necessary for Diffie-Hellman, and this
13778 specific range makes Diffie-Hellman unnecessarily inefficient if
13779 dh->length (recommended exponent length) is much smaller than the
13780 length of dh->p. We could use BN_rand_range() if the order of
13781 the subgroup was stored in the DH structure, but we only have
13786 BN_rand(priv_key, l, ...)
13788 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13795 RSA_eay_public_encrypt
13796 RSA_eay_private_decrypt
13797 RSA_eay_private_encrypt (signing)
13798 RSA_eay_public_decrypt (signature verification)
13800 (default implementations for RSA_public_encrypt,
13801 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13802 always reject numbers >= n.
13806 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13807 to synchronize access to 'locking_thread'. This is necessary on
13808 systems where access to 'locking_thread' (an 'unsigned long'
13809 variable) is not atomic.
13813 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13814 *before* setting the 'crypto_lock_rand' flag. The previous code had
13815 a race condition if 0 is a valid thread ID.
13817 *Travis Vitek <vitek@roguewave.com>*
13819 * Add support for shared libraries under Irix.
13821 *Albert Chin-A-Young <china@thewrittenword.com>*
13823 * Add configuration option to build on Linux on both big-endian and
13824 little-endian MIPS.
13826 *Ralf Baechle <ralf@uni-koblenz.de>*
13828 * Add the possibility to create shared libraries on HP-UX.
13832 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13834 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13835 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13836 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13837 PRNG state recovery was possible based on the output of
13838 one PRNG request appropriately sized to gain knowledge on
13839 'md' followed by enough consecutive 1-byte PRNG requests
13840 to traverse all of 'state'.
13842 1. When updating 'md_local' (the current thread's copy of 'md')
13843 during PRNG output generation, hash all of the previous
13844 'md_local' value, not just the half used for PRNG output.
13846 2. Make the number of bytes from 'state' included into the hash
13847 independent from the number of PRNG bytes requested.
13849 The first measure alone would be sufficient to avoid
13850 Markku-Juhani's attack. (Actually it had never occurred
13851 to me that the half of 'md_local' used for chaining was the
13852 half from which PRNG output bytes were taken -- I had always
13853 assumed that the secret half would be used.) The second
13854 measure makes sure that additional data from 'state' is never
13855 mixed into 'md_local' in small portions; this heuristically
13856 further strengthens the PRNG.
13860 * Fix crypto/bn/asm/mips3.s.
13864 * When only the key is given to "enc", the IV is undefined. Print out
13865 an error message in this case.
13869 * Handle special case when X509_NAME is empty in X509 printing routines.
13873 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13874 positive and less than q.
13878 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13879 used: it isn't thread safe and the add_lock_callback should handle
13882 *Paul Rose <Paul.Rose@bridge.com>*
13884 * Verify that incoming data obeys the block size in
13885 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13891 *Ulf Möller, Bodo Möller*
13893 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13894 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13895 when fixing the server behaviour for backwards-compatible 'client
13896 hello' messages. (Note that the attack is impractical against
13897 SSL 3.0 and TLS 1.0 anyway because length and version checking
13898 means that the probability of guessing a valid ciphertext is
13899 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13902 Before 0.9.5, the countermeasure (hide the error by generating a
13903 random 'decryption result') did not work properly because
13904 ERR_clear_error() was missing, meaning that SSL_get_error() would
13905 detect the supposedly ignored error.
13907 Both problems are now fixed.
13911 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13912 (previously it was 1024).
13916 * Fix for compatibility mode trust settings: ignore trust settings
13917 unless some valid trust or reject settings are present.
13921 * Fix for blowfish EVP: its a variable length cipher.
13925 * Fix various bugs related to DSA S/MIME verification. Handle missing
13926 parameters in DSA public key structures and return an error in the
13927 DSA routines if parameters are absent.
13931 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13932 in the current directory if neither $RANDFILE nor $HOME was set.
13933 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13934 caused some confusion to Windows users who haven't defined $HOME.
13935 Thus RAND_file_name() is changed again: e_os.h can define a
13936 DEFAULT_HOME, which will be used if $HOME is not set.
13937 For Windows, we use "C:"; on other platforms, we still require
13938 environment variables.
13940 * Move 'if (!initialized) RAND_poll()' into regions protected by
13941 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13942 having multiple threads call RAND_poll() concurrently.
13946 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13947 combination of a flag and a thread ID variable.
13948 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13949 flag), *other* threads can enter ssleay_add_bytes without obeying
13950 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13951 that they do not hold after the first thread unsets add_do_not_lock).
13955 * Change bctest again: '-x' expressions are not available in all
13956 versions of 'test'.
13960 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13962 * Fix a couple of memory leaks in PKCS7_dataDecode()
13964 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13966 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13967 the default extension for executables, if any. Also, make the perl
13968 scripts that use symlink() to test if it really exists and use "cp"
13969 if it doesn't. All this made OpenSSL compilable and installable in
13974 * Fix for asn1_GetSequence() for indefinite length constructed data.
13975 If SEQUENCE is length is indefinite just set c->slen to the total
13976 amount of data available.
13978 *Steve Henson, reported by shige@FreeBSD.org*
13980 *This change does not apply to 0.9.7.*
13982 * Change bctest to avoid here-documents inside command substitution
13983 (workaround for FreeBSD /bin/sh bug).
13984 For compatibility with Ultrix, avoid shell functions (introduced
13985 in the bctest version that searches along $PATH).
13989 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13990 with des_encrypt() defined on some operating systems, like Solaris
13995 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13996 On the Importance of Eliminating Errors in Cryptographic
13997 Computations, J. Cryptology 14 (2001) 2, 101-119,
13998 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14002 * MIPS assembler BIGNUM division bug fix.
14006 * Disabled incorrect Alpha assembler code.
14010 * Fix PKCS#7 decode routines so they correctly update the length
14011 after reading an EOC for the EXPLICIT tag.
14015 *This change does not apply to 0.9.7.*
14017 * Fix bug in PKCS#12 key generation routines. This was triggered
14018 if a 3DES key was generated with a 0 initial byte. Include
14019 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14020 (but broken) behaviour.
14024 * Enhance bctest to search for a working bc along $PATH and print
14027 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14029 * Fix memory leaks in err.c: free err_data string if necessary;
14030 don't write to the wrong index in ERR_set_error_data.
14034 * Implement ssl23_peek (analogous to ssl23_read), which previously
14039 * Replace rdtsc with `_emit` statements for VC++ version 5.
14041 *Jeremy Cooper <jeremy@baymoo.org>*
14043 * Make it possible to reuse SSLv2 sessions.
14047 * In copy_email() check for >= 0 as a return value for
14048 X509_NAME_get_index_by_NID() since 0 is a valid index.
14050 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14052 * Avoid coredump with unsupported or invalid public keys by checking if
14053 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14054 PKCS7_verify() fails with non detached data.
14058 * Don't use getenv in library functions when run as setuid/setgid.
14059 New function OPENSSL_issetugid().
14063 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14064 due to incorrect handling of multi-threading:
14066 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14068 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14070 3. Count how many times MemCheck_off() has been called so that
14071 nested use can be treated correctly. This also avoids
14072 inband-signalling in the previous code (which relied on the
14073 assumption that thread ID 0 is impossible).
14077 * Add "-rand" option also to s_client and s_server.
14081 * Fix CPU detection on Irix 6.x.
14082 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14083 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14085 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14090 *This change does not apply to 0.9.7.*
14092 * Use the cached encoding of an X509_NAME structure rather than
14093 copying it. This is apparently the reason for the libsafe "errors"
14094 but the code is actually correct.
14098 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14099 Bleichenbacher's DSA attack.
14100 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14101 to be set and top=0 forces the highest bit to be set; top=-1 is new
14102 and leaves the highest bit random.
14104 *Ulf Moeller, Bodo Moeller*
14106 * In the `NCONF_...`-based implementations for `CONF_...` queries
14107 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14108 a temporary CONF structure with the data component set to NULL
14109 (which gives segmentation faults in lh_retrieve).
14110 Instead, use NULL for the CONF pointer in CONF_get_string and
14111 CONF_get_number (which may use environment variables) and directly
14112 return NULL from CONF_get_section.
14116 * Fix potential buffer overrun for EBCDIC.
14120 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14121 keyUsage if basicConstraints absent for a CA.
14125 * Make SMIME_write_PKCS7() write mail header values with a format that
14126 is more generally accepted (no spaces before the semicolon), since
14127 some programs can't parse those values properly otherwise. Also make
14128 sure BIO's that break lines after each write do not create invalid
14133 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14134 macros previously used would not encode an empty SEQUENCE OF
14135 and break the signature.
14139 *This change does not apply to 0.9.7.*
14141 * Zero the premaster secret after deriving the master secret in
14146 * Add some EVP_add_digest_alias registrations (as found in
14147 OpenSSL_add_all_digests()) to SSL_library_init()
14148 aka OpenSSL_add_ssl_algorithms(). This provides improved
14149 compatibility with peers using X.509 certificates
14150 with unconventional AlgorithmIdentifier OIDs.
14154 * Fix for Irix with NO_ASM.
14156 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14158 * ./config script fixes.
14160 *Ulf Moeller, Richard Levitte*
14162 * Fix 'openssl passwd -1'.
14166 * Change PKCS12_key_gen_asc() so it can cope with non null
14167 terminated strings whose length is passed in the passlen
14168 parameter, for example from PEM callbacks. This was done
14169 by adding an extra length parameter to asc2uni().
14171 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14173 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14174 call failed, free the DSA structure.
14178 * Fix to uni2asc() to cope with zero length Unicode strings.
14179 These are present in some PKCS#12 files.
14183 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14184 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14185 when writing a 32767 byte record.
14187 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14189 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14190 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14192 (RSA objects have a reference count access to which is protected
14193 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14194 so they are meant to be shared between threads.)
14195 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14196 "Reddie, Steven" <Steven.Reddie@ca.com>*
14198 * Fix a deadlock in CRYPTO_mem_leaks().
14202 * Use better test patterns in bntest.
14206 * rand_win.c fix for Borland C.
14210 * BN_rshift bugfix for n == 0.
14214 * Add a 'bctest' script that checks for some known 'bc' bugs
14215 so that 'make test' does not abort just because 'bc' is broken.
14219 * Store verify_result within SSL_SESSION also for client side to
14220 avoid potential security hole. (Re-used sessions on the client side
14221 always resulted in verify_result==X509_V_OK, not using the original
14222 result of the server certificate verification.)
14226 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14227 SSL3_RT_APPLICATION_DATA, return 0.
14228 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14233 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14234 releases, have been re-implemented by renaming the previous
14235 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14236 and ssl3_read_internal, respectively, and adding 'peek' parameters
14237 to them. The new ssl[23]_{read,peek} functions are calls to
14238 ssl[23]_read_internal with the 'peek' flag set appropriately.
14239 A 'peek' parameter has also been added to ssl3_read_bytes, which
14240 does the actual work for ssl3_read_internal.
14244 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14245 the method-specific "init()" handler. Also clean up ex_data after
14246 calling the method-specific "finish()" handler. Previously, this was
14247 happening the other way round.
14251 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14252 The previous value, 12, was not always sufficient for BN_mod_exp().
14256 * Make sure that shared libraries get the internal name engine with
14257 the full version number and not just 0. This should mark the
14258 shared libraries as not backward compatible. Of course, this should
14259 be changed again when we can guarantee backward binary compatibility.
14263 * Fix typo in get_cert_by_subject() in by_dir.c
14265 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14267 * Rework the system to generate shared libraries:
14269 - Make note of the expected extension for the shared libraries and
14270 if there is a need for symbolic links from for example libcrypto.so.0
14271 to libcrypto.so.0.9.7. There is extended info in Configure for
14274 - Make as few rebuilds of the shared libraries as possible.
14276 - Still avoid linking the OpenSSL programs with the shared libraries.
14278 - When installing, install the shared libraries separately from the
14283 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14285 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14286 and not in SSL_clear because the latter is also used by the
14287 accept/connect functions; previously, the settings made by
14288 SSL_set_read_ahead would be lost during the handshake.
14290 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14292 * Correct util/mkdef.pl to be selective about disabled algorithms.
14293 Previously, it would create entries for disabled algorithms no
14298 * Added several new manual pages for SSL_* function.
14302 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14304 * In ssl23_get_client_hello, generate an error message when faced
14305 with an initial SSL 3.0/TLS record that is too small to contain the
14306 first two bytes of the ClientHello message, i.e. client_version.
14307 (Note that this is a pathologic case that probably has never happened
14308 in real life.) The previous approach was to use the version number
14309 from the record header as a substitute; but our protocol choice
14310 should not depend on that one because it is not authenticated
14311 by the Finished messages.
14315 * More robust randomness gathering functions for Windows.
14317 *Jeffrey Altman <jaltman@columbia.edu>*
14319 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14320 not set then we don't setup the error code for issuer check errors
14321 to avoid possibly overwriting other errors which the callback does
14322 handle. If an application does set the flag then we assume it knows
14323 what it is doing and can handle the new informational codes
14328 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14329 a general "ANY" type, as such it should be able to decode anything
14330 including tagged types. However it didn't check the class so it would
14331 wrongly interpret tagged types in the same way as their universal
14332 counterpart and unknown types were just rejected. Changed so that the
14333 tagged and unknown types are handled in the same way as a SEQUENCE:
14334 that is the encoding is stored intact. There is also a new type
14335 "V_ASN1_OTHER" which is used when the class is not universal, in this
14336 case we have no idea what the actual type is so we just lump them all
14341 * On VMS, stdout may very well lead to a file that is written to
14342 in a record-oriented fashion. That means that every write() will
14343 write a separate record, which will be read separately by the
14344 programs trying to read from it. This can be very confusing.
14346 The solution is to put a BIO filter in the way that will buffer
14347 text until a linefeed is reached, and then write everything a
14348 line at a time, so every record written will be an actual line,
14349 not chunks of lines and not (usually doesn't happen, but I've
14350 seen it once) several lines in one record. BIO_f_linebuffer() is
14353 Currently, it's a VMS-only method, because that's where it has
14354 been tested well enough.
14358 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14359 it can return incorrect results.
14360 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14361 but it was in 0.9.6-beta[12].)
14365 * Disable the check for content being present when verifying detached
14366 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14367 include zero length content when signing messages.
14371 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14372 BIO_ctrl (for BIO pairs).
14376 * Add DSO method for VMS.
14380 * Bug fix: Montgomery multiplication could produce results with the
14385 * Add RPM specification openssl.spec and modify it to build three
14386 packages. The default package contains applications, application
14387 documentation and run-time libraries. The devel package contains
14388 include files, static libraries and function documentation. The
14389 doc package contains the contents of the doc directory. The original
14390 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14394 * Add a large number of documentation files for many SSL routines.
14396 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14398 * Add a configuration entry for Sony News 4.
14400 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14402 * Don't set the two most significant bits to one when generating a
14403 random number < q in the DSA library.
14407 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14408 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14409 the underlying transport is blocking) if a handshake took place.
14410 (The default behaviour is needed by applications such as s_client
14411 and s_server that use select() to determine when to use SSL_read;
14412 but for applications that know in advance when to expect data, it
14413 just makes things more complicated.)
14417 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14422 * Add a few more EBCDIC conditionals that make `req` and `x509`
14423 work better on such systems.
14425 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14427 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14428 Update PKCS12_parse() so it copies the friendlyName and the
14429 keyid to the certificates aux info.
14433 * Fix bug in PKCS7_verify() which caused an infinite loop
14434 if there was more than one signature.
14436 *Sven Uszpelkat <su@celocom.de>*
14438 * Major change in util/mkdef.pl to include extra information
14439 about each symbol, as well as presenting variables as well
14440 as functions. This change means that there's n more need
14441 to rebuild the .num files when some algorithms are excluded.
14445 * Allow the verify time to be set by an application,
14446 rather than always using the current time.
14450 * Phase 2 verify code reorganisation. The certificate
14451 verify code now looks up an issuer certificate by a
14452 number of criteria: subject name, authority key id
14453 and key usage. It also verifies self signed certificates
14454 by the same criteria. The main comparison function is
14455 X509_check_issued() which performs these checks.
14457 Lot of changes were necessary in order to support this
14458 without completely rewriting the lookup code.
14460 Authority and subject key identifier are now cached.
14462 The LHASH 'certs' is X509_STORE has now been replaced
14463 by a STACK_OF(X509_OBJECT). This is mainly because an
14464 LHASH can't store or retrieve multiple objects with
14465 the same hash value.
14467 As a result various functions (which were all internal
14468 use only) have changed to handle the new X509_STORE
14469 structure. This will break anything that messed round
14470 with X509_STORE internally.
14472 The functions X509_STORE_add_cert() now checks for an
14473 exact match, rather than just subject name.
14475 The X509_STORE API doesn't directly support the retrieval
14476 of multiple certificates matching a given criteria, however
14477 this can be worked round by performing a lookup first
14478 (which will fill the cache with candidate certificates)
14479 and then examining the cache for matches. This is probably
14480 the best we can do without throwing out X509_LOOKUP
14481 entirely (maybe later...).
14483 The X509_VERIFY_CTX structure has been enhanced considerably.
14485 All certificate lookup operations now go via a get_issuer()
14486 callback. Although this currently uses an X509_STORE it
14487 can be replaced by custom lookups. This is a simple way
14488 to bypass the X509_STORE hackery necessary to make this
14489 work and makes it possible to use more efficient techniques
14490 in future. A very simple version which uses a simple
14491 STACK for its trusted certificate store is also provided
14492 using X509_STORE_CTX_trusted_stack().
14494 The verify_cb() and verify() callbacks now have equivalents
14495 in the X509_STORE_CTX structure.
14497 X509_STORE_CTX also has a 'flags' field which can be used
14498 to customise the verify behaviour.
14502 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14503 excludes S/MIME capabilities.
14507 * When a certificate request is read in keep a copy of the
14508 original encoding of the signed data and use it when outputting
14509 again. Signatures then use the original encoding rather than
14510 a decoded, encoded version which may cause problems if the
14511 request is improperly encoded.
14515 * For consistency with other BIO_puts implementations, call
14516 buffer_write(b, ...) directly in buffer_puts instead of calling
14519 In BIO_puts, increment b->num_write as in BIO_write.
14521 *Peter.Sylvester@EdelWeb.fr*
14523 * Fix BN_mul_word for the case where the word is 0. (We have to use
14524 BN_zero, we may not return a BIGNUM with an array consisting of
14525 words set to zero.)
14529 * Avoid calling abort() from within the library when problems are
14530 detected, except if preprocessor symbols have been defined
14531 (such as REF_CHECK, BN_DEBUG etc.).
14535 * New openssl application 'rsautl'. This utility can be
14536 used for low level RSA operations. DER public key
14537 BIO/fp routines also added.
14541 * New Configure entry and patches for compiling on QNX 4.
14543 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14545 * A demo state-machine implementation was sponsored by
14546 Nuron (<http://www.nuron.com/>) and is now available in
14547 demos/state_machine.
14551 * New options added to the 'dgst' utility for signature
14552 generation and verification.
14556 * Unrecognized PKCS#7 content types are now handled via a
14557 catch all ASN1_TYPE structure. This allows unsupported
14558 types to be stored as a "blob" and an application can
14559 encode and decode it manually.
14563 * Fix various signed/unsigned issues to make a_strex.c
14564 compile under VC++.
14566 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14568 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14569 length if passed a buffer. ASN1_INTEGER_to_BN failed
14570 if passed a NULL BN and its argument was negative.
14572 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14574 * Modification to PKCS#7 encoding routines to output definite
14575 length encoding. Since currently the whole structures are in
14576 memory there's not real point in using indefinite length
14577 constructed encoding. However if OpenSSL is compiled with
14578 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14582 * Added BIO_vprintf() and BIO_vsnprintf().
14586 * Added more prefixes to parse for in the strings written
14587 through a logging bio, to cover all the levels that are available
14588 through syslog. The prefixes are now:
14590 PANIC, EMERG, EMR => LOG_EMERG
14591 ALERT, ALR => LOG_ALERT
14592 CRIT, CRI => LOG_CRIT
14593 ERROR, ERR => LOG_ERR
14594 WARNING, WARN, WAR => LOG_WARNING
14595 NOTICE, NOTE, NOT => LOG_NOTICE
14596 INFO, INF => LOG_INFO
14597 DEBUG, DBG => LOG_DEBUG
14599 and as before, if none of those prefixes are present at the
14600 beginning of the string, LOG_ERR is chosen.
14602 On Win32, the `LOG_*` levels are mapped according to this:
14604 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14605 LOG_WARNING => EVENTLOG_WARNING_TYPE
14606 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14610 * Made it possible to reconfigure with just the configuration
14611 argument "reconf" or "reconfigure". The command line arguments
14612 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14613 and are retrieved from there when reconfiguring.
14619 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14621 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14625 * The obj_dat.pl script was messing up the sorting of object
14626 names. The reason was that it compared the quoted version
14627 of strings as a result "OCSP" > "OCSP Signing" because
14628 " > SPACE. Changed script to store unquoted versions of
14629 names and add quotes on output. It was also omitting some
14630 names from the lookup table if they were given a default
14631 value (that is if SN is missing it is given the same
14632 value as LN and vice versa), these are now added on the
14633 grounds that if an object has a name we should be able to
14634 look it up. Finally added warning output when duplicate
14635 short or long names are found.
14639 * Changes needed for Tandem NSK.
14641 *Scott Uroff <scott@xypro.com>*
14643 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14644 RSA_padding_check_SSLv23(), special padding was never detected
14645 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14646 version rollback attacks was not effective.
14648 In s23_clnt.c, don't use special rollback-attack detection padding
14649 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14650 client; similarly, in s23_srvr.c, don't do the rollback check if
14651 SSL 2.0 is the only protocol enabled in the server.
14655 * Make it possible to get hexdumps of unprintable data with 'openssl
14656 asn1parse'. By implication, the functions ASN1_parse_dump() and
14657 BIO_dump_indent() are added.
14661 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14662 these print out strings and name structures based on various
14663 flags including RFC2253 support and proper handling of
14664 multibyte characters. Added options to the 'x509' utility
14665 to allow the various flags to be set.
14669 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14670 Also change the functions X509_cmp_current_time() and
14671 X509_gmtime_adj() work with an ASN1_TIME structure,
14672 this will enable certificates using GeneralizedTime in validity
14673 dates to be checked.
14677 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14678 negative public key encodings) on by default,
14679 NO_NEG_PUBKEY_BUG can be set to disable it.
14683 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14684 content octets. An i2c_ASN1_OBJECT is unnecessary because
14685 the encoding can be trivially obtained from the structure.
14689 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14690 not read locks (`CRYPTO_r_[un]lock`).
14694 * A first attempt at creating official support for shared
14695 libraries through configuration. I've kept it so the
14696 default is static libraries only, and the OpenSSL programs
14697 are always statically linked for now, but there are
14698 preparations for dynamic linking in place.
14699 This has been tested on Linux and Tru64.
14703 * Randomness polling function for Win9x, as described in:
14704 Peter Gutmann, Software Generation of Practically Strong
14709 * Fix so PRNG is seeded in req if using an already existing
14714 * New options to smime application. -inform and -outform
14715 allow alternative formats for the S/MIME message including
14716 PEM and DER. The -content option allows the content to be
14717 specified separately. This should allow things like Netscape
14718 form signing output easier to verify.
14722 * Fix the ASN1 encoding of tags using the 'long form'.
14726 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14727 STRING types. These convert content octets to and from the
14728 underlying type. The actual tag and length octets are
14729 already assumed to have been read in and checked. These
14730 are needed because all other string types have virtually
14731 identical handling apart from the tag. By having versions
14732 of the ASN1 functions that just operate on content octets
14733 IMPLICIT tagging can be handled properly. It also allows
14734 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14735 and ASN1_INTEGER are identical apart from the tag.
14739 * Change the handling of OID objects as follows:
14741 - New object identifiers are inserted in objects.txt, following
14742 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14743 - objects.pl is used to process obj_mac.num and create a new
14745 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14748 This is currently kind of a hack, and the perl code in objects.pl
14749 isn't very elegant, but it works as I intended. The simplest way
14750 to check that it worked correctly is to look in obj_dat.h and
14751 check the array nid_objs and make sure the objects haven't moved
14752 around (this is important!). Additions are OK, as well as
14753 consistent name changes.
14757 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14761 * Addition of the command line parameter '-rand file' to 'openssl req'.
14762 The given file adds to whatever has already been seeded into the
14763 random pool through the RANDFILE configuration file option or
14764 environment variable, or the default random state file.
14768 * mkstack.pl now sorts each macro group into lexical order.
14769 Previously the output order depended on the order the files
14770 appeared in the directory, resulting in needless rewriting
14775 * Patches to make OpenSSL compile under Win32 again. Mostly
14776 work arounds for the VC++ problem that it treats func() as
14777 func(void). Also stripped out the parts of mkdef.pl that
14778 added extra typesafe functions: these no longer exist.
14782 * Reorganisation of the stack code. The macros are now all
14783 collected in safestack.h . Each macro is defined in terms of
14784 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14785 DEBUG_SAFESTACK is now handled in terms of function casts,
14786 this has the advantage of retaining type safety without the
14787 use of additional functions. If DEBUG_SAFESTACK is not defined
14788 then the non typesafe macros are used instead. Also modified the
14789 mkstack.pl script to handle the new form. Needs testing to see
14790 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14791 the default if no major problems. Similar behaviour for ASN1_SET_OF
14792 and PKCS12_STACK_OF.
14796 * When some versions of IIS use the 'NET' form of private key the
14797 key derivation algorithm is different. Normally MD5(password) is
14798 used as a 128 bit RC4 key. In the modified case
14799 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14800 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14801 as the old Netscape_RSA functions except they have an additional
14802 'sgckey' parameter which uses the modified algorithm. Also added
14803 an -sgckey command line option to the rsa utility. Thanks to
14804 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14805 algorithm to openssl-dev.
14809 * The evp_local.h macros were using 'c.##kname' which resulted in
14810 invalid expansion on some systems (SCO 5.0.5 for example).
14811 Corrected to 'c.kname'.
14813 *Phillip Porch <root@theporch.com>*
14815 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14816 a STACK of email addresses from a certificate or request, these look
14817 in the subject name and the subject alternative name extensions and
14818 omit any duplicate addresses.
14822 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14823 This makes DSA verification about 2 % faster.
14827 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14828 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14829 plus overhead for 1024 bit moduli).
14830 This makes exponentiations about 0.5 % faster for 1024 bit
14831 exponents (as measured by "openssl speed rsa2048").
14835 * Rename memory handling macros to avoid conflicts with other
14837 Malloc => OPENSSL_malloc
14838 Malloc_locked => OPENSSL_malloc_locked
14839 Realloc => OPENSSL_realloc
14840 Free => OPENSSL_free
14844 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14845 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14849 * CygWin32 support.
14851 *John Jarvie <jjarvie@newsguy.com>*
14853 * The type-safe stack code has been rejigged. It is now only compiled
14854 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14855 by default all type-specific stack functions are "#define"d back to
14856 standard stack functions. This results in more streamlined output
14857 but retains the type-safety checking possibilities of the original
14862 * The STACK code has been cleaned up, and certain type declarations
14863 that didn't make a lot of sense have been brought in line. This has
14864 also involved a cleanup of sorts in safestack.h to more correctly
14865 map type-safe stack functions onto their plain stack counterparts.
14866 This work has also resulted in a variety of "const"ifications of
14867 lots of the code, especially `_cmp` operations which should normally
14868 be prototyped with "const" parameters anyway.
14872 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14873 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14874 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14875 where all of 'md' is used each time the PRNG is used, but 'state'
14876 is used only indexed by a cyclic counter. As entropy may not be
14877 well distributed from the beginning, 'md' is important as a
14878 chaining variable. However, the output function chains only half
14879 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14880 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14881 in all of 'state' being rewritten, with the new values depending
14882 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14886 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14887 the handshake is continued after ssl_verify_cert_chain();
14888 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14889 can lead to 'unexplainable' connection aborts later.
14891 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14893 * Major EVP API cipher revision.
14894 Add hooks for extra EVP features. This allows various cipher
14895 parameters to be set in the EVP interface. Support added for variable
14896 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14897 setting of RC2 and RC5 parameters.
14899 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14902 Remove lots of duplicated code from the EVP library. For example *every*
14903 cipher init() function handles the 'iv' in the same way according to the
14904 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14905 for CFB and OFB modes they zero ctx->num.
14907 New functionality allows removal of S/MIME code RC2 hack.
14909 Most of the routines have the same form and so can be declared in terms
14912 By shifting this to the top level EVP_CipherInit() it can be removed from
14913 all individual ciphers. If the cipher wants to handle IVs or keys
14914 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14917 Change lots of functions like EVP_EncryptUpdate() to now return a
14918 value: although software versions of the algorithms cannot fail
14919 any installed hardware versions can.
14923 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14924 this option is set, tolerate broken clients that send the negotiated
14925 protocol version number instead of the requested protocol version
14930 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14931 i.e. non-zero for export ciphersuites, zero otherwise.
14932 Previous versions had this flag inverted, inconsistent with
14933 rsa_tmp_cb (..._TMP_RSA_CB).
14935 *Bodo Moeller; problem reported by Amit Chopra*
14937 * Add missing DSA library text string. Work around for some IIS
14938 key files with invalid SEQUENCE encoding.
14942 * Add a document (doc/standards.txt) that list all kinds of standards
14943 and so on that are implemented in OpenSSL.
14947 * Enhance c_rehash script. Old version would mishandle certificates
14948 with the same subject name hash and wouldn't handle CRLs at all.
14949 Added -fingerprint option to crl utility, to support new c_rehash
14954 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14958 * Fix for SSL server purpose checking. Server checking was
14959 rejecting certificates which had extended key usage present
14960 but no ssl client purpose.
14962 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14964 * Make PKCS#12 code work with no password. The PKCS#12 spec
14965 is a little unclear about how a blank password is handled.
14966 Since the password in encoded as a BMPString with terminating
14967 double NULL a zero length password would end up as just the
14968 double NULL. However no password at all is different and is
14969 handled differently in the PKCS#12 key generation code. NS
14970 treats a blank password as zero length. MSIE treats it as no
14971 password on export: but it will try both on import. We now do
14972 the same: PKCS12_parse() tries zero length and no password if
14973 the password is set to "" or NULL (NULL is now a valid password:
14974 it wasn't before) as does the pkcs12 application.
14978 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14979 perror when PEM_read_bio_X509_REQ fails, the error message must
14980 be obtained from the error queue.
14984 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14985 it in ERR_remove_state if appropriate, and change ERR_get_state
14986 accordingly to avoid race conditions (this is necessary because
14987 thread_hash is no longer constant once set).
14991 * Bugfix for linux-elf makefile.one.
14995 * RSA_get_default_method() will now cause a default
14996 RSA_METHOD to be chosen if one doesn't exist already.
14997 Previously this was only set during a call to RSA_new()
14998 or RSA_new_method(NULL) meaning it was possible for
14999 RSA_get_default_method() to return NULL.
15003 * Added native name translation to the existing DSO code
15004 that will convert (if the flag to do so is set) filenames
15005 that are sufficiently small and have no path information
15006 into a canonical native form. Eg. "blah" converted to
15007 "libblah.so" or "blah.dll" etc.
15011 * New function ERR_error_string_n(e, buf, len) which is like
15012 ERR_error_string(e, buf), but writes at most 'len' bytes
15013 including the 0 terminator. For ERR_error_string_n, 'buf'
15016 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15018 * CONF library reworked to become more general. A new CONF
15019 configuration file reader "class" is implemented as well as a
15020 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15021 old `CONF_*` functions are still there, but are reimplemented to
15022 work in terms of the new functions. Also, a set of functions
15023 to handle the internal storage of the configuration data is
15024 provided to make it easier to write new configuration file
15025 reader "classes" (I can definitely see something reading a
15026 configuration file in XML format, for example), called `_CONF_*`,
15027 or "the configuration storage API"...
15029 The new configuration file reading functions are:
15031 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15032 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15034 NCONF_default, NCONF_WIN32
15036 NCONF_dump_fp, NCONF_dump_bio
15038 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15039 NCONF_new creates a new CONF object. This works in the same way
15040 as other interfaces in OpenSSL, like the BIO interface.
15041 `NCONF_dump_*` dump the internal storage of the configuration file,
15042 which is useful for debugging. All other functions take the same
15043 arguments as the old `CONF_*` functions with the exception of the
15044 first that must be a `CONF *` instead of a `LHASH *`.
15046 To make it easier to use the new classes with the old `CONF_*` functions,
15047 the function CONF_set_default_method is provided.
15051 * Add '-tls1' option to 'openssl ciphers', which was already
15052 mentioned in the documentation but had not been implemented.
15053 (This option is not yet really useful because even the additional
15054 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15058 * Initial DSO code added into libcrypto for letting OpenSSL (and
15059 OpenSSL-based applications) load shared libraries and bind to
15060 them in a portable way.
15062 *Geoff Thorpe, with contributions from Richard Levitte*
15064 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15066 * Make sure _lrotl and _lrotr are only used with MSVC.
15068 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15069 (the default implementation of RAND_status).
15071 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15072 to '-clrext' (= clear extensions), as intended and documented.
15073 *Bodo Moeller; inconsistency pointed out by Michael Attili
15074 <attili@amaxo.com>*
15076 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15077 was larger than the MD block size.
15079 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15081 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15082 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15083 using the passed key: if the passed key was a private key the result
15084 of X509_print(), for example, would be to print out all the private key
15089 * des_quad_cksum() byte order bug fix.
15090 *Ulf Möller, using the problem description in krb4-0.9.7, where
15091 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15093 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15096 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15098 * For easily testing in shell scripts whether some command
15099 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15100 returns with exit code 0 iff no command of the given name is available.
15101 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15102 the output goes to stdout and nothing is printed to stderr.
15103 Additional arguments are always ignored.
15105 Since for each cipher there is a command of the same name,
15106 the 'no-cipher' compilation switches can be tested this way.
15108 ('openssl no-XXX' is not able to detect pseudo-commands such
15109 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15113 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15117 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15118 is set; it will be thrown away anyway because each handshake creates
15120 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15121 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15122 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15123 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15127 * New s_client option -ign_eof: EOF at stdin is ignored, and
15128 'Q' and 'R' lose their special meanings (quit/renegotiate).
15129 This is part of what -quiet does; unlike -quiet, -ign_eof
15130 does not suppress any output.
15134 * Add compatibility options to the purpose and trust code. The
15135 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15136 accepts a certificate or CA, this was the previous behaviour,
15137 with all the associated security issues.
15139 X509_TRUST_COMPAT is the old trust behaviour: only and
15140 automatically trust self signed roots in certificate store. A
15141 new trust setting X509_TRUST_DEFAULT is used to specify that
15142 a purpose has no associated trust setting and it should instead
15143 use the value in the default purpose.
15147 * Fix the PKCS#8 DSA private key code so it decodes keys again
15148 and fix a memory leak.
15152 * In util/mkerr.pl (which implements 'make errors'), preserve
15153 reason strings from the previous version of the .c file, as
15154 the default to have only downcase letters (and digits) in
15155 automatically generated reasons codes is not always appropriate.
15159 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15160 using strerror. Previously, ERR_reason_error_string() returned
15161 library names as reason strings for SYSerr; but SYSerr is a special
15162 case where small numbers are errno values, not library numbers.
15166 * Add '-dsaparam' option to 'openssl dhparam' application. This
15167 converts DSA parameters into DH parameters. (When creating parameters,
15168 DSA_generate_parameters is used.)
15172 * Include 'length' (recommended exponent length) in C code generated
15173 by 'openssl dhparam -C'.
15177 * The second argument to set_label in perlasm was already being used
15178 so couldn't be used as a "file scope" flag. Moved to third argument
15183 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15184 instead of RAND_bytes for encryption IVs and salts.
15188 * Include RAND_status() into RAND_METHOD instead of implementing
15189 it only for md_rand.c Otherwise replacing the PRNG by calling
15190 RAND_set_rand_method would be impossible.
15194 * Don't let DSA_generate_key() enter an infinite loop if the random
15195 number generation fails.
15199 * New 'rand' application for creating pseudo-random output.
15203 * Added configuration support for Linux/IA64
15205 *Rolf Haberrecker <rolf@suse.de>*
15207 * Assembler module support for Mingw32.
15211 * Shared library support for HPUX (in shlib/).
15213 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15215 * Shared library support for Solaris gcc.
15217 *Lutz Behnke <behnke@trustcenter.de>*
15219 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15221 * PKCS7_encrypt() was adding text MIME headers twice because they
15222 were added manually and by SMIME_crlf_copy().
15226 * In bntest.c don't call BN_rand with zero bits argument.
15228 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15230 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15231 case was implemented. This caused BN_div_recp() to fail occasionally.
15235 * Add an optional second argument to the set_label() in the perl
15236 assembly language builder. If this argument exists and is set
15237 to 1 it signals that the assembler should use a symbol whose
15238 scope is the entire file, not just the current function. This
15239 is needed with MASM which uses the format label:: for this scope.
15241 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15243 * Change the ASN1 types so they are typedefs by default. Before
15244 almost all types were #define'd to ASN1_STRING which was causing
15245 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15250 * Change names of new functions to the new get1/get0 naming
15251 convention: After 'get1', the caller owns a reference count
15252 and has to call `..._free`; 'get0' returns a pointer to some
15253 data structure without incrementing reference counters.
15254 (Some of the existing 'get' functions increment a reference
15255 counter, some don't.)
15256 Similarly, 'set1' and 'add1' functions increase reference
15257 counters or duplicate objects.
15261 * Allow for the possibility of temp RSA key generation failure:
15262 the code used to assume it always worked and crashed on failure.
15266 * Fix potential buffer overrun problem in BIO_printf().
15267 *Ulf Möller, using public domain code by Patrick Powell; problem
15268 pointed out by David Sacerdote <das33@cornell.edu>*
15270 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15271 RAND_egd() and RAND_status(). In the command line application,
15272 the EGD socket can be specified like a seed file using RANDFILE
15277 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15278 Some CAs (e.g. Verisign) distribute certificates in this form.
15282 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15283 list to exclude them. This means that no special compilation option
15284 is needed to use anonymous DH: it just needs to be included in the
15289 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15290 EVP_MD_type. The old functionality is available in a new macro called
15291 EVP_MD_md(). Change code that uses it and update docs.
15295 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15296 where the `void *` argument is replaced by a function pointer argument.
15297 Previously `void *` was abused to point to functions, which works on
15298 many platforms, but is not correct. As these functions are usually
15299 called by macros defined in OpenSSL header files, most source code
15300 should work without changes.
15304 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15305 sections with information on -D... compiler switches used for
15306 compiling the library so that applications can see them. To enable
15307 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15308 must be defined. E.g.,
15309 #define OPENSSL_ALGORITHM_DEFINES
15310 #include <openssl/opensslconf.h>
15311 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15313 *Richard Levitte, Ulf and Bodo Möller*
15315 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15320 * Change the 'other' type in certificate aux info to a STACK_OF
15321 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15322 the required ASN1 format: arbitrary types determined by an OID.
15326 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15327 argument to 'req'. This is not because the function is newer or
15328 better than others it just uses the work 'NEW' in the certificate
15329 request header lines. Some software needs this.
15333 * Reorganise password command line arguments: now passwords can be
15334 obtained from various sources. Delete the PEM_cb function and make
15335 it the default behaviour: i.e. if the callback is NULL and the
15336 usrdata argument is not NULL interpret it as a null terminated pass
15337 phrase. If usrdata and the callback are NULL then the pass phrase
15338 is prompted for as usual.
15342 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15343 the support is automatically enabled. The resulting binaries will
15344 autodetect the card and use it if present.
15346 *Ben Laurie and Compaq Inc.*
15348 * Work around for Netscape hang bug. This sends certificate request
15349 and server done in one record. Since this is perfectly legal in the
15350 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15351 the bugs/SSLv3 entry for more info.
15355 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15359 * Add -rand argument to smime and pkcs12 applications and read/write
15364 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15368 * Add command line password options to the remaining applications.
15372 * Bug fix for BN_div_recp() for numerators with an even number of
15377 * More tests in bntest.c, and changed test_bn output.
15381 * ./config recognizes MacOS X now.
15385 * Bug fix for BN_div() when the first words of num and divisor are
15386 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15390 * Add support for various broken PKCS#8 formats, and command line
15391 options to produce them.
15395 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15396 get temporary BIGNUMs from a BN_CTX.
15400 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15405 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15406 include a #define from the old name to the new. The original intent
15407 was that statically linked binaries could for example just call
15408 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15409 link with digests. This never worked because SSLeay_add_all_digests()
15410 and SSLeay_add_all_ciphers() were in the same source file so calling
15411 one would link with the other. They are now in separate source files.
15415 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15419 * Use a less unusual form of the Miller-Rabin primality test (it used
15420 a binary algorithm for exponentiation integrated into the Miller-Rabin
15421 loop, our standard modexp algorithms are faster).
15425 * Support for the EBCDIC character set completed.
15427 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15429 * Source code cleanups: use const where appropriate, eliminate casts,
15430 use `void *` instead of `char *` in lhash.
15434 * Bugfix: ssl3_send_server_key_exchange was not restartable
15435 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15436 this the server could overwrite ephemeral keys that the client
15441 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15442 using 50 iterations of the Rabin-Miller test.
15444 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15445 iterations of the Rabin-Miller test as required by the appendix
15446 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15447 As BN_is_prime_fasttest includes trial division, DSA parameter
15448 generation becomes much faster.
15450 This implies a change for the callback functions in DSA_is_prime
15451 and DSA_generate_parameters: The callback function is called once
15452 for each positive witness in the Rabin-Miller test, not just
15453 occasionally in the inner loop; and the parameters to the
15454 callback function now provide an iteration count for the outer
15455 loop rather than for the current invocation of the inner loop.
15456 DSA_generate_parameters additionally can call the callback
15457 function with an 'iteration count' of -1, meaning that a
15458 candidate has passed the trial division test (when q is generated
15459 from an application-provided seed, trial division is skipped).
15463 * New function BN_is_prime_fasttest that optionally does trial
15464 division before starting the Rabin-Miller test and has
15465 an additional BN_CTX * argument (whereas BN_is_prime always
15466 has to allocate at least one BN_CTX).
15467 'callback(1, -1, cb_arg)' is called when a number has passed the
15468 trial division stage.
15472 * Fix for bug in CRL encoding. The validity dates weren't being handled
15477 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15481 * New function BN_pseudo_rand().
15485 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15486 bignum version of BN_from_montgomery() with the working code from
15487 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15492 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15493 made it impossible to use the same SSL_SESSION data structure in
15494 SSL2 clients in multiple threads.
15498 * The return value of RAND_load_file() no longer counts bytes obtained
15499 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15500 to seed the PRNG (previously an explicit byte count was required).
15502 *Ulf Möller, Bodo Möller*
15504 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15505 used `char *` instead of `void *` and had casts all over the place.
15509 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15513 * Retain source code compatibility for BN_prime_checks macro:
15514 BN_is_prime(..., BN_prime_checks, ...) now uses
15515 BN_prime_checks_for_size to determine the appropriate number of
15516 Rabin-Miller iterations.
15520 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15521 DH_CHECK_P_NOT_SAFE_PRIME.
15522 (Check if this is true? OpenPGP calls them "strong".)
15526 * Merge the functionality of "dh" and "gendh" programs into a new program
15527 "dhparam". The old programs are retained for now but will handle DH keys
15528 (instead of parameters) in future.
15532 * Make the ciphers, s_server and s_client programs check the return values
15533 when a new cipher list is set.
15537 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15538 ciphers. Before when the 56bit ciphers were enabled the sorting was
15541 The syntax for the cipher sorting has been extended to support sorting by
15542 cipher-strength (using the strength_bits hard coded in the tables).
15543 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15545 Fix a bug in the cipher-command parser: when supplying a cipher command
15546 string with an "undefined" symbol (neither command nor alphanumeric
15547 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15548 an error is flagged.
15550 Due to the strength-sorting extension, the code of the
15551 ssl_create_cipher_list() function was completely rearranged. I hope that
15552 the readability was also increased :-)
15554 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15556 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15557 for the first serial number and places 2 in the serial number file. This
15558 avoids problems when the root CA is created with serial number zero and
15559 the first user certificate has the same issuer name and serial number
15564 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15565 the new code. Add documentation for this stuff.
15569 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15570 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15571 structures and behave in an analogous way to the X509v3 functions:
15572 they shouldn't be called directly but wrapper functions should be used
15575 So we also now have some wrapper functions that call the X509at functions
15576 when passed certificate requests. (TO DO: similar things can be done with
15577 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15578 things. Some of these need some d2i or i2d and print functionality
15579 because they handle more complex structures.)
15583 * Add missing #ifndefs that caused missing symbols when building libssl
15584 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15585 NO_RSA in `ssl/s2*.c`.
15587 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15589 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15590 has a return value which indicates the quality of the random data
15591 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15592 error queue. New function RAND_pseudo_bytes() generates output that is
15593 guaranteed to be unique but not unpredictable. RAND_add is like
15594 RAND_seed, but takes an extra argument for an entropy estimate
15595 (RAND_seed always assumes full entropy).
15599 * Do more iterations of Rabin-Miller probable prime test (specifically,
15600 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15601 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15602 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15603 false-positive rate of at most 2^-80 for random input.
15607 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15611 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15612 in the 0.9.5 release), this returns the chain
15613 from an X509_CTX structure with a dup of the stack and all
15614 the X509 reference counts upped: so the stack will exist
15615 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15618 Also make SSL_SESSION_print() print out the verify return
15623 * Add manpage for the pkcs12 command. Also change the default
15624 behaviour so MAC iteration counts are used unless the new
15625 -nomaciter option is used. This improves file security and
15626 only older versions of MSIE (4.0 for example) need it.
15630 * Honor the no-xxx Configure options when creating .DEF files.
15634 * Add PKCS#10 attributes to field table: challengePassword,
15635 unstructuredName and unstructuredAddress. These are taken from
15636 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15637 international characters are used.
15639 More changes to X509_ATTRIBUTE code: allow the setting of types
15640 based on strings. Remove the 'loc' parameter when adding
15641 attributes because these will be a SET OF encoding which is sorted
15646 * Initial changes to the 'req' utility to allow request generation
15647 automation. This will allow an application to just generate a template
15648 file containing all the field values and have req construct the
15651 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15652 used all over the place including certificate requests and PKCS#7
15653 structures. They are currently handled manually where necessary with
15654 some primitive wrappers for PKCS#7. The new functions behave in a
15655 manner analogous to the X509 extension functions: they allow
15656 attributes to be looked up by NID and added.
15658 Later something similar to the X509V3 code would be desirable to
15659 automatically handle the encoding, decoding and printing of the
15660 more complex types. The string types like challengePassword can
15661 be handled by the string table functions.
15663 Also modified the multi byte string table handling. Now there is
15664 a 'global mask' which masks out certain types. The table itself
15665 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15666 is useful when for example there is only one permissible type
15667 (as in countryName) and using the mask might result in no valid
15672 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15673 SSL_get_peer_finished to allow applications to obtain the latest
15674 Finished messages sent to the peer or expected from the peer,
15675 respectively. (SSL_get_peer_finished is usually the Finished message
15676 actually received from the peer, otherwise the protocol will be aborted.)
15678 As the Finished message are message digests of the complete handshake
15679 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15680 be used for external authentication procedures when the authentication
15681 provided by SSL/TLS is not desired or is not enough.
15685 * Enhanced support for Alpha Linux is added. Now ./config checks if
15686 the host supports BWX extension and if Compaq C is present on the
15687 $PATH. Just exploiting of the BWX extension results in 20-30%
15688 performance kick for some algorithms, e.g. DES and RC4 to mention
15689 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15694 * Add support for MS "fast SGC". This is arguably a violation of the
15695 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15696 weak crypto and after checking the certificate is SGC a second one
15697 with strong crypto. MS SGC stops the first handshake after receiving
15698 the server certificate message and sends a second client hello. Since
15699 a server will typically do all the time consuming operations before
15700 expecting any further messages from the client (server key exchange
15701 is the most expensive) there is little difference between the two.
15703 To get OpenSSL to support MS SGC we have to permit a second client
15704 hello message after we have sent server done. In addition we have to
15705 reset the MAC if we do get this second client hello.
15709 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15710 if a DER encoded private key is RSA or DSA traditional format. Changed
15711 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15712 format DER encoded private key. Newer code should use PKCS#8 format which
15713 has the key type encoded in the ASN1 structure. Added DER private key
15714 support to pkcs8 application.
15718 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15719 ciphersuites has been selected (as required by the SSL 3/TLS 1
15720 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15721 is set, we interpret this as a request to violate the specification
15722 (the worst that can happen is a handshake failure, and 'correct'
15723 behaviour would result in a handshake failure anyway).
15727 * In SSL_CTX_add_session, take into account that there might be multiple
15728 SSL_SESSION structures with the same session ID (e.g. when two threads
15729 concurrently obtain them from an external cache).
15730 The internal cache can handle only one SSL_SESSION with a given ID,
15731 so if there's a conflict, we now throw out the old one to achieve
15736 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15737 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15738 some routines that use cipher OIDs: some ciphers do not have OIDs
15739 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15744 * Simplify the trust setting structure and code. Now we just have
15745 two sequences of OIDs for trusted and rejected settings. These will
15746 typically have values the same as the extended key usage extension
15747 and any application specific purposes.
15749 The trust checking code now has a default behaviour: it will just
15750 check for an object with the same NID as the passed id. Functions can
15751 be provided to override either the default behaviour or the behaviour
15752 for a given id. SSL client, server and email already have functions
15753 in place for compatibility: they check the NID and also return "trusted"
15754 if the certificate is self signed.
15758 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15759 traditional format into an EVP_PKEY structure.
15763 * Add a password callback function PEM_cb() which either prompts for
15764 a password if usr_data is NULL or otherwise assumes it is a null
15765 terminated password. Allow passwords to be passed on command line
15766 environment or config files in a few more utilities.
15770 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15771 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15772 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15773 Update documentation.
15777 * Support for ASN1 "NULL" type. This could be handled before by using
15778 ASN1_TYPE but there wasn't any function that would try to read a NULL
15779 and produce an error if it couldn't. For compatibility we also have
15780 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15781 don't allocate anything because they don't need to.
15785 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15788 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15790 * Rebuild of the memory allocation routines used by OpenSSL code and
15791 possibly others as well. The purpose is to make an interface that
15792 provide hooks so anyone can build a separate set of allocation and
15793 deallocation routines to be used by OpenSSL, for example memory
15794 pool implementations, or something else, which was previously hard
15795 since Malloc(), Realloc() and Free() were defined as macros having
15796 the values malloc, realloc and free, respectively (except for Win32
15797 compilations). The same is provided for memory debugging code.
15798 OpenSSL already comes with functionality to find memory leaks, but
15799 this gives people a chance to debug other memory problems.
15801 With these changes, a new set of functions and macros have appeared:
15803 CRYPTO_set_mem_debug_functions() [F]
15804 CRYPTO_get_mem_debug_functions() [F]
15805 CRYPTO_dbg_set_options() [F]
15806 CRYPTO_dbg_get_options() [F]
15807 CRYPTO_malloc_debug_init() [M]
15809 The memory debug functions are NULL by default, unless the library
15810 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15811 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15812 gives the standard debugging functions that come with OpenSSL) or
15813 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15814 provided by the library user) must be used. When the standard
15815 debugging functions are used, CRYPTO_dbg_set_options can be used to
15816 request additional information:
15817 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15818 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15820 Also, things like CRYPTO_set_mem_functions will always give the
15821 expected result (the new set of functions is used for allocation
15822 and deallocation) at all times, regardless of platform and compiler
15825 To finish it up, some functions that were never use in any other
15826 way than through macros have a new API and new semantic:
15828 CRYPTO_dbg_malloc()
15829 CRYPTO_dbg_realloc()
15832 All macros of value have retained their old syntax.
15834 *Richard Levitte and Bodo Moeller*
15836 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15837 ordering of SMIMECapabilities wasn't in "strength order" and there
15838 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15843 * Some ASN1 types with illegal zero length encoding (INTEGER,
15844 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15846 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15848 * Merge in my S/MIME library for OpenSSL. This provides a simple
15849 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15850 functionality to handle multipart/signed properly) and a utility
15851 called 'smime' to call all this stuff. This is based on code I
15852 originally wrote for Celo who have kindly allowed it to be
15853 included in OpenSSL.
15857 * Add variants des_set_key_checked and des_set_key_unchecked of
15858 des_set_key (aka des_key_sched). Global variable des_check_key
15859 decides which of these is called by des_set_key; this way
15860 des_check_key behaves as it always did, but applications and
15861 the library itself, which was buggy for des_check_key == 1,
15862 have a cleaner way to pick the version they need.
15866 * New function PKCS12_newpass() which changes the password of a
15871 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15872 dynamic mix. In both cases the ids can be used as an index into the
15873 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15874 functions so they accept a list of the field values and the
15875 application doesn't need to directly manipulate the X509_TRUST
15880 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15885 * Modify the way the V3 extension code looks up extensions. This now
15886 works in a similar way to the object code: we have some "standard"
15887 extensions in a static table which is searched with OBJ_bsearch()
15888 and the application can add dynamic ones if needed. The file
15889 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15890 updated whenever a new extension is added to the core code and kept
15891 in ext_nid order. There is a simple program 'tabtest.c' which checks
15892 this. New extensions are not added too often so this file can readily
15893 be maintained manually.
15895 There are two big advantages in doing things this way. The extensions
15896 can be looked up immediately and no longer need to be "added" using
15897 X509V3_add_standard_extensions(): this function now does nothing.
15898 Side note: I get *lots* of email saying the extension code doesn't
15899 work because people forget to call this function.
15900 Also no dynamic allocation is done unless new extensions are added:
15901 so if we don't add custom extensions there is no need to call
15902 X509V3_EXT_cleanup().
15906 * Modify enc utility's salting as follows: make salting the default. Add a
15907 magic header, so unsalted files fail gracefully instead of just decrypting
15908 to garbage. This is because not salting is a big security hole, so people
15909 should be discouraged from doing it.
15913 * Fixes and enhancements to the 'x509' utility. It allowed a message
15914 digest to be passed on the command line but it only used this
15915 parameter when signing a certificate. Modified so all relevant
15916 operations are affected by the digest parameter including the
15917 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15918 DSA key was used because it didn't fix the digest.
15922 * Initial certificate chain verify code. Currently tests the untrusted
15923 certificates for consistency with the verify purpose (which is set
15924 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15926 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15927 this is because it will reject chains with invalid extensions whereas
15928 every previous version of OpenSSL and SSLeay made no checks at all.
15930 Trust code: checks the root CA for the relevant trust settings. Trust
15931 settings have an initial value consistent with the verify purpose: e.g.
15932 if the verify purpose is for SSL client use it expects the CA to be
15933 trusted for SSL client use. However the default value can be changed to
15934 permit custom trust settings: one example of this would be to only trust
15935 certificates from a specific "secure" set of CAs.
15937 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15938 which should be used for version portability: especially since the
15939 verify structure is likely to change more often now.
15941 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15942 to set them. If not set then assume SSL clients will verify SSL servers
15945 Two new options to the verify program: -untrusted allows a set of
15946 untrusted certificates to be passed in and -purpose which sets the
15947 intended purpose of the certificate. If a purpose is set then the
15948 new chain verify code is used to check extension consistency.
15952 * Support for the authority information access extension.
15956 * Modify RSA and DSA PEM read routines to transparently handle
15957 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15958 public keys in a format compatible with certificate
15959 SubjectPublicKeyInfo structures. Unfortunately there were already
15960 functions called *_PublicKey_* which used various odd formats so
15961 these are retained for compatibility: however the DSA variants were
15962 never in a public release so they have been deleted. Changed dsa/rsa
15963 utilities to handle the new format: note no releases ever handled public
15964 keys so we should be OK.
15966 The primary motivation for this change is to avoid the same fiasco
15967 that dogs private keys: there are several incompatible private key
15968 formats some of which are standard and some OpenSSL specific and
15969 require various evil hacks to allow partial transparent handling and
15970 even then it doesn't work with DER formats. Given the option anything
15971 other than PKCS#8 should be dumped: but the other formats have to
15972 stay in the name of compatibility.
15974 With public keys and the benefit of hindsight one standard format
15975 is used which works with EVP_PKEY, RSA or DSA structures: though
15976 it clearly returns an error if you try to read the wrong kind of key.
15978 Added a -pubkey option to the 'x509' utility to output the public key.
15979 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15980 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15981 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15982 that do the same as the `EVP_PKEY_assign_*()` except they up the
15983 reference count of the added key (they don't "swallow" the
15988 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15989 CRLs would fail if the file contained no certificates or no CRLs:
15990 added a new function to read in both types and return the number
15991 read: this means that if none are read it will be an error. The
15992 DER versions of the certificate and CRL reader would always fail
15993 because it isn't possible to mix certificates and CRLs in DER format
15994 without choking one or the other routine. Changed this to just read
15995 a certificate: this is the best we can do. Also modified the code
15996 in apps/verify.c to take notice of return codes: it was previously
15997 attempting to read in certificates from NULL pointers and ignoring
15998 any errors: this is one reason why the cert and CRL reader seemed
15999 to work. It doesn't check return codes from the default certificate
16000 routines: these may well fail if the certificates aren't installed.
16004 * Code to support otherName option in GeneralName.
16008 * First update to verify code. Change the verify utility
16009 so it warns if it is passed a self signed certificate:
16010 for consistency with the normal behaviour. X509_verify
16011 has been modified to it will now verify a self signed
16012 certificate if *exactly* the same certificate appears
16013 in the store: it was previously impossible to trust a
16014 single self signed certificate. This means that:
16015 openssl verify ss.pem
16016 now gives a warning about a self signed certificate but
16017 openssl verify -CAfile ss.pem ss.pem
16022 * For servers, store verify_result in SSL_SESSION data structure
16023 (and add it to external session representation).
16024 This is needed when client certificate verifications fails,
16025 but an application-provided verification callback (set by
16026 SSL_CTX_set_cert_verify_callback) allows accepting the session
16027 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16028 but returns 1): When the session is reused, we have to set
16029 ssl->verify_result to the appropriate error code to avoid
16032 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16034 * Fix a bug in the new PKCS#7 code: it didn't consider the
16035 case in PKCS7_dataInit() where the signed PKCS7 structure
16036 didn't contain any existing data because it was being created.
16038 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16040 * Add a salt to the key derivation routines in enc.c. This
16041 forms the first 8 bytes of the encrypted file. Also add a
16042 -S option to allow a salt to be input on the command line.
16046 * New function X509_cmp(). Oddly enough there wasn't a function
16047 to compare two certificates. We do this by working out the SHA1
16048 hash and comparing that. X509_cmp() will be needed by the trust
16053 * SSL_get1_session() is like SSL_get_session(), but increments
16054 the reference count in the SSL_SESSION returned.
16056 *Geoff Thorpe <geoff@eu.c2.net>*
16058 * Fix for 'req': it was adding a null to request attributes.
16059 Also change the X509_LOOKUP and X509_INFO code to handle
16060 certificate auxiliary information.
16064 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16069 * Add the possibility to add extra information to the memory leak
16070 detecting output, to form tracebacks, showing from where each
16071 allocation was originated: CRYPTO_push_info("constant string") adds
16072 the string plus current file name and line number to a per-thread
16073 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16074 is like calling CYRPTO_pop_info() until the stack is empty.
16075 Also updated memory leak detection code to be multi-thread-safe.
16079 * Add options -text and -noout to pkcs7 utility and delete the
16080 encryption options which never did anything. Update docs.
16084 * Add options to some of the utilities to allow the pass phrase
16085 to be included on either the command line (not recommended on
16086 OSes like Unix) or read from the environment. Update the
16087 manpages and fix a few bugs.
16091 * Add a few manpages for some of the openssl commands.
16095 * Fix the -revoke option in ca. It was freeing up memory twice,
16096 leaking and not finding already revoked certificates.
16100 * Extensive changes to support certificate auxiliary information.
16101 This involves the use of X509_CERT_AUX structure and X509_AUX
16102 functions. An X509_AUX function such as PEM_read_X509_AUX()
16103 can still read in a certificate file in the usual way but it
16104 will also read in any additional "auxiliary information". By
16105 doing things this way a fair degree of compatibility can be
16106 retained: existing certificates can have this information added
16107 using the new 'x509' options.
16109 Current auxiliary information includes an "alias" and some trust
16110 settings. The trust settings will ultimately be used in enhanced
16111 certificate chain verification routines: currently a certificate
16112 can only be trusted if it is self signed and then it is trusted
16117 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16118 The problem was that one of the replacement routines had not been working
16119 since SSLeay releases. For now the offending routine has been replaced
16120 with non-optimised assembler. Even so, this now gives around 95%
16121 performance improvement for 1024 bit RSA signs.
16125 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16126 handling. Most clients have the effective key size in bits equal to
16127 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16128 A few however don't do this and instead use the size of the decrypted key
16129 to determine the RC2 key length and the AlgorithmIdentifier to determine
16130 the effective key length. In this case the effective key length can still
16131 be 40 bits but the key length can be 168 bits for example. This is fixed
16132 by manually forcing an RC2 key into the EVP_PKEY structure because the
16133 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16134 the key length and effective key length are equal.
16138 * Add a bunch of functions that should simplify the creation of
16139 X509_NAME structures. Now you should be able to do:
16140 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16141 and have it automatically work out the correct field type and fill in
16142 the structures. The more adventurous can try:
16143 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16144 and it will (hopefully) work out the correct multibyte encoding.
16148 * Change the 'req' utility to use the new field handling and multibyte
16149 copy routines. Before the DN field creation was handled in an ad hoc
16150 way in req, ca, and x509 which was rather broken and didn't support
16151 BMPStrings or UTF8Strings. Since some software doesn't implement
16152 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16153 using the dirstring_type option. See the new comment in the default
16154 openssl.cnf for more info.
16158 * Make crypto/rand/md_rand.c more robust:
16159 - Assure unique random numbers after fork().
16160 - Make sure that concurrent threads access the global counter and
16161 md serializably so that we never lose entropy in them
16162 or use exactly the same state in multiple threads.
16163 Access to the large state is not always serializable because
16164 the additional locking could be a performance killer, and
16165 md should be large enough anyway.
16169 * New file apps/app_rand.c with commonly needed functionality
16170 for handling the random seed file.
16172 Use the random seed file in some applications that previously did not:
16174 dsaparam -genkey (which also ignored its '-rand' option),
16177 x509 (when signing).
16178 Except on systems with /dev/urandom, it is crucial to have a random
16179 seed file at least for key creation, DSA signing, and for DH exchanges;
16180 for RSA signatures we could do without one.
16182 gendh and gendsa (unlike genrsa) used to read only the first byte
16183 of each file listed in the '-rand' option. The function as previously
16184 found in genrsa is now in app_rand.c and is used by all programs
16185 that support '-rand'.
16189 * In RAND_write_file, use mode 0600 for creating files;
16190 don't just chmod when it may be too late.
16194 * Report an error from X509_STORE_load_locations
16195 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16199 * New function ASN1_mbstring_copy() this copies a string in either
16200 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16201 into an ASN1_STRING type. A mask of permissible types is passed
16202 and it chooses the "minimal" type to use or an error if not type
16207 * Add function equivalents to the various macros in asn1.h. The old
16208 macros are retained with an `M_` prefix. Code inside the library can
16209 use the `M_` macros. External code (including the openssl utility)
16210 should *NOT* in order to be "shared library friendly".
16214 * Add various functions that can check a certificate's extensions
16215 to see if it usable for various purposes such as SSL client,
16216 server or S/MIME and CAs of these types. This is currently
16217 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16218 verification. Also added a -purpose flag to x509 utility to
16219 print out all the purposes.
16223 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16228 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16229 for, obtain and decode and extension and obtain its critical flag.
16230 This allows all the necessary extension code to be handled in a
16231 single function call.
16235 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16236 platforms. See crypto/rc4/rc4_enc.c for further details.
16240 * New -noout option to asn1parse. This causes no output to be produced
16241 its main use is when combined with -strparse and -out to extract data
16242 from a file (which may not be in ASN.1 format).
16246 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16247 when producing the local key id.
16249 *Richard Levitte <levitte@stacken.kth.se>*
16251 * New option -dhparam in s_server. This allows a DH parameter file to be
16252 stated explicitly. If it is not stated then it tries the first server
16253 certificate file. The previous behaviour hard coded the filename
16258 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16259 a public key to be input or output. For example:
16260 openssl rsa -in key.pem -pubout -out pubkey.pem
16261 Also added necessary DSA public key functions to handle this.
16265 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16266 in the message. This was handled by allowing
16267 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16269 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16271 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16272 to the end of the strings whereas this didn't. This would cause problems
16273 if strings read with d2i_ASN1_bytes() were later modified.
16275 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16277 * Fix for base64 decode bug. When a base64 bio reads only one line of
16278 data and it contains EOF it will end up returning an error. This is
16279 caused by input 46 bytes long. The cause is due to the way base64
16280 BIOs find the start of base64 encoded data. They do this by trying a
16281 trial decode on each line until they find one that works. When they
16282 do a flag is set and it starts again knowing it can pass all the
16283 data directly through the decoder. Unfortunately it doesn't reset
16284 the context it uses. This means that if EOF is reached an attempt
16285 is made to pass two EOFs through the context and this causes the
16286 resulting error. This can also cause other problems as well. As is
16287 usual with these problems it takes *ages* to find and the fix is
16288 trivial: move one line.
16290 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16292 * Ugly workaround to get s_client and s_server working under Windows. The
16293 old code wouldn't work because it needed to select() on sockets and the
16294 tty (for keypresses and to see if data could be written). Win32 only
16295 supports select() on sockets so we select() with a 1s timeout on the
16296 sockets and then see if any characters are waiting to be read, if none
16297 are present then we retry, we also assume we can always write data to
16298 the tty. This isn't nice because the code then blocks until we've
16299 received a complete line of data and it is effectively polling the
16300 keyboard at 1s intervals: however it's quite a bit better than not
16301 working at all :-) A dedicated Windows application might handle this
16302 with an event loop for example.
16306 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16307 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16308 will be called when RSA_sign() and RSA_verify() are used. This is useful
16309 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16310 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16311 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16312 This necessitated the support of an extra signature type NID_md5_sha1
16313 for SSL signatures and modifications to the SSL library to use it instead
16314 of calling RSA_public_decrypt() and RSA_private_encrypt().
16318 * Add new -verify -CAfile and -CApath options to the crl program, these
16319 will lookup a CRL issuers certificate and verify the signature in a
16320 similar way to the verify program. Tidy up the crl program so it
16321 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16322 less strict. It will now permit CRL extensions even if it is not
16323 a V2 CRL: this will allow it to tolerate some broken CRLs.
16327 * Initialize all non-automatic variables each time one of the openssl
16328 sub-programs is started (this is necessary as they may be started
16329 multiple times from the "OpenSSL>" prompt).
16331 *Lennart Bang, Bodo Moeller*
16333 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16334 removing all other RSA functionality (this is what NO_RSA does). This
16335 is so (for example) those in the US can disable those operations covered
16336 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16341 * Non-copying interface to BIO pairs.
16342 (still largely untested)
16346 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16347 ASCII string. This was handled independently in various places before.
16351 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16352 UTF8 strings a character at a time.
16356 * Use client_version from client hello to select the protocol
16357 (s23_srvr.c) and for RSA client key exchange verification
16358 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16362 * Add various utility functions to handle SPKACs, these were previously
16363 handled by poking round in the structure internals. Added new function
16364 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16365 print, verify and generate SPKACs. Based on an original idea from
16366 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16370 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16374 * Allow the config file extension section to be overwritten on the
16375 command line. Based on an original idea from Massimiliano Pala
16376 <madwolf@comune.modena.it>. The new option is called -extensions
16377 and can be applied to ca, req and x509. Also -reqexts to override
16378 the request extensions in req and -crlexts to override the crl extensions
16383 * Add new feature to the SPKAC handling in ca. Now you can include
16384 the same field multiple times by preceding it by "XXXX." for example:
16387 this is the same syntax as used in the req config file.
16391 * Allow certificate extensions to be added to certificate requests. These
16392 are specified in a 'req_extensions' option of the req section of the
16393 config file. They can be printed out with the -text option to req but
16394 are otherwise ignored at present.
16398 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16399 data read consists of only the final block it would not decrypted because
16400 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16401 A misplaced 'break' also meant the decrypted final block might not be
16402 copied until the next read.
16406 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16407 a few extra parameters to the DH structure: these will be useful if
16408 for example we want the value of 'q' or implement X9.42 DH.
16412 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16413 provides hooks that allow the default DSA functions or functions on a
16414 "per key" basis to be replaced. This allows hardware acceleration and
16415 hardware key storage to be handled without major modification to the
16416 library. Also added low level modexp hooks and CRYPTO_EX structure and
16417 associated functions.
16421 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16422 as "read only": it can't be written to and the buffer it points to will
16423 not be freed. Reading from a read only BIO is much more efficient than
16424 a normal memory BIO. This was added because there are several times when
16425 an area of memory needs to be read from a BIO. The previous method was
16426 to create a memory BIO and write the data to it, this results in two
16427 copies of the data and an O(n^2) reading algorithm. There is a new
16428 function BIO_new_mem_buf() which creates a read only memory BIO from
16429 an area of memory. Also modified the PKCS#7 routines to use read only
16434 * Bugfix: ssl23_get_client_hello did not work properly when called in
16435 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16436 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16437 but a retry condition occurred while trying to read the rest.
16441 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16442 NID_pkcs7_encrypted by default: this was wrong since this should almost
16443 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16444 the encrypted data type: this is a more sensible place to put it and it
16445 allows the PKCS#12 code to be tidied up that duplicated this
16450 * Changed obj_dat.pl script so it takes its input and output files on
16451 the command line. This should avoid shell escape redirection problems
16456 * Initial support for certificate extension requests, these are included
16457 in things like Xenroll certificate requests. Included functions to allow
16458 extensions to be obtained and added.
16462 * -crlf option to s_client and s_server for sending newlines as
16463 CRLF (as required by many protocols).
16467 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16469 * Install libRSAglue.a when OpenSSL is built with RSAref.
16471 *Ralf S. Engelschall*
16473 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16475 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16477 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16482 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16483 DH parameters/keys (q is lost during that conversion, but the resulting
16484 DH parameters contain its length).
16486 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16487 much faster than DH_generate_parameters (which creates parameters
16488 where `p = 2*q + 1`), and also the smaller q makes DH computations
16489 much more efficient (160-bit exponentiation instead of 1024-bit
16490 exponentiation); so this provides a convenient way to support DHE
16491 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16492 utter importance to use
16493 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16495 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16496 when such DH parameters are used, because otherwise small subgroup
16497 attacks may become possible!
16501 * Avoid memory leak in i2d_DHparams.
16505 * Allow the -k option to be used more than once in the enc program:
16506 this allows the same encrypted message to be read by multiple recipients.
16510 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16511 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16512 it will always use the numerical form of the OID, even if it has a short
16517 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16518 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16519 otherwise bn_mod_exp was called. In the case of hardware keys for example
16520 no private key components need be present and it might store extra data
16521 in the RSA structure, which cannot be accessed from bn_mod_exp.
16522 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16523 private key operations.
16527 * Added support for SPARC Linux.
16531 * pem_password_cb function type incompatibly changed from
16532 typedef int pem_password_cb(char *buf, int size, int rwflag);
16534 ....(char *buf, int size, int rwflag, void *userdata);
16535 so that applications can pass data to their callbacks:
16536 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16537 additional void * argument, which is just handed through whenever
16538 the password callback is called.
16540 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16542 New function SSL_CTX_set_default_passwd_cb_userdata.
16544 Compatibility note: As many C implementations push function arguments
16545 onto the stack in reverse order, the new library version is likely to
16546 interoperate with programs that have been compiled with the old
16547 pem_password_cb definition (PEM_whatever takes some data that
16548 happens to be on the stack as its last argument, and the callback
16549 just ignores this garbage); but there is no guarantee whatsoever that
16552 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16553 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16554 problems not only on Windows, but also on some Unix platforms.
16555 To avoid problematic command lines, these definitions are now in an
16556 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16557 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16561 * MIPS III/IV assembler module is reimplemented.
16565 * More DES library cleanups: remove references to srand/rand and
16566 delete an unused file.
16570 * Add support for the free Netwide assembler (NASM) under Win32,
16571 since not many people have MASM (ml) and it can be hard to obtain.
16572 This is currently experimental but it seems to work OK and pass all
16573 the tests. Check out INSTALL.W32 for info.
16577 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16578 without temporary keys kept an extra copy of the server key,
16579 and connections with temporary keys did not free everything in case
16584 * New function RSA_check_key and new openssl rsa option -check
16585 for verifying the consistency of RSA keys.
16587 *Ulf Moeller, Bodo Moeller*
16589 * Various changes to make Win32 compile work:
16590 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16591 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16592 comparison" warnings.
16593 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16597 * Add a debugging option to PKCS#5 v2 key generation function: when
16598 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16599 derived keys are printed to stderr.
16603 * Copy the flags in ASN1_STRING_dup().
16605 *Roman E. Pavlov <pre@mo.msk.ru>*
16607 * The x509 application mishandled signing requests containing DSA
16608 keys when the signing key was also DSA and the parameters didn't match.
16610 It was supposed to omit the parameters when they matched the signing key:
16611 the verifying software was then supposed to automatically use the CA's
16612 parameters if they were absent from the end user certificate.
16614 Omitting parameters is no longer recommended. The test was also
16615 the wrong way round! This was probably due to unusual behaviour in
16616 EVP_cmp_parameters() which returns 1 if the parameters match.
16617 This meant that parameters were omitted when they *didn't* match and
16618 the certificate was useless. Certificates signed with 'ca' didn't have
16621 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16623 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16624 The interface is as follows:
16625 Applications can use
16626 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16627 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16628 "off" is now the default.
16629 The library internally uses
16630 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16631 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16632 to disable memory-checking temporarily.
16634 Some inconsistent states that previously were possible (and were
16635 even the default) are now avoided.
16637 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16638 with each memory chunk allocated; this is occasionally more helpful
16639 than just having a counter.
16641 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16643 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16648 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16649 which largely parallels "options", but is for changing API behaviour,
16650 whereas "options" are about protocol behaviour.
16651 Initial "mode" flags are:
16653 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16654 a single record has been written.
16655 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16656 retries use the same buffer location.
16657 (But all of the contents must be
16662 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16665 * Fix problems with no-hmac etc.
16667 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16669 * New functions RSA_get_default_method(), RSA_set_method() and
16670 RSA_get_method(). These allows replacement of RSA_METHODs without having
16671 to mess around with the internals of an RSA structure.
16675 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16676 Also really enable memory leak checks in openssl.c and in some
16679 *Chad C. Mulligan, Bodo Moeller*
16681 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16682 up the length of negative integers. This has now been simplified to just
16683 store the length when it is first determined and use it later, rather
16684 than trying to keep track of where data is copied and updating it to
16686 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16688 * Add a new function PKCS7_signatureVerify. This allows the verification
16689 of a PKCS#7 signature but with the signing certificate passed to the
16690 function itself. This contrasts with PKCS7_dataVerify which assumes the
16691 certificate is present in the PKCS#7 structure. This isn't always the
16692 case: certificates can be omitted from a PKCS#7 structure and be
16693 distributed by "out of band" means (such as a certificate database).
16697 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16698 function prototypes in pem.h, also change util/mkdef.pl to add the
16699 necessary function names.
16703 * mk1mf.pl (used by Windows builds) did not properly read the
16704 options set by Configure in the top level Makefile, and Configure
16705 was not even able to write more than one option correctly.
16706 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16710 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16711 file to be loaded from a BIO or FILE pointer. The BIO version will
16712 for example allow memory BIOs to contain config info.
16716 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16717 Whoever hopes to achieve shared-library compatibility across versions
16718 must use this, not the compile-time macro.
16719 (Exercise 0.9.4: Which is the minimum library version required by
16721 Note: All this applies only to multi-threaded programs, others don't
16726 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16727 through a BIO pair triggered the default case, i.e.
16728 SSLerr(...,SSL_R_UNKNOWN_STATE).
16732 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16733 can use the SSL library even if none of the specific BIOs is
16738 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16739 for the encoded length.
16741 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16743 * Add initial documentation of the X509V3 functions.
16747 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16748 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16749 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16750 secure PKCS#8 private key format with a high iteration count.
16754 * Fix determination of Perl interpreter: A perl or perl5
16755 *directory* in $PATH was also accepted as the interpreter.
16757 *Ralf S. Engelschall*
16759 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16760 wrong with it but it was very old and did things like calling
16761 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16762 unusual formatting.
16766 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16767 to use the new extension code.
16771 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16772 with macros. This should make it easier to change their form, add extra
16773 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16778 * Add to configuration table a new entry that can specify an alternative
16779 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16780 according to Mark Crispin <MRC@Panda.COM>.
16784 * DES CBC did not update the IV. Weird.
16788 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16789 Changing the behaviour of the former might break existing programs --
16790 where IV updating is needed, des_ncbc_encrypt can be used.
16793 * When bntest is run from "make test" it drives bc to check its
16794 calculations, as well as internally checking them. If an internal check
16795 fails, it needs to cause bc to give a non-zero result or make test carries
16796 on without noticing the failure. Fixed.
16800 * DES library cleanups.
16804 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16805 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16806 ciphers. NOTE: although the key derivation function has been verified
16807 against some published test vectors it has not been extensively tested
16808 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16813 * Instead of "mkdir -p", which is not fully portable, use new
16814 Perl script "util/mkdir-p.pl".
16818 * Rewrite the way password based encryption (PBE) is handled. It used to
16819 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16820 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16821 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16822 the 'parameter' field of the AlgorithmIdentifier is passed to the
16823 underlying key generation function so it must do its own ASN1 parsing.
16824 This has also changed the EVP_PBE_CipherInit() function which now has a
16825 'parameter' argument instead of literal salt and iteration count values
16826 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16830 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16831 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16832 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16833 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16834 value was just used as a "magic string" and not used directly its
16835 value doesn't matter.
16839 * Introduce some semblance of const correctness to BN. Shame C doesn't
16844 * "linux-sparc64" configuration (ultrapenguin).
16846 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16847 "linux-sparc" configuration.
16849 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16851 * config now generates no-xxx options for missing ciphers.
16855 * Support the EBCDIC character set (work in progress).
16856 File ebcdic.c not yet included because it has a different license.
16858 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16860 * Support BS2000/OSD-POSIX.
16862 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16864 * Make callbacks for key generation use `void *` instead of `char *`.
16868 * Make S/MIME samples compile (not yet tested).
16872 * Additional typesafe stacks.
16876 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16880 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16882 * New configuration variant "sco5-gcc".
16884 * Updated some demos.
16886 *Sean O Riordain, Wade Scholine*
16888 * Add missing BIO_free at exit of pkcs12 application.
16892 * Fix memory leak in conf.c.
16896 * Updates for Win32 to assembler version of MD5.
16900 * Set #! path to perl in apps/der_chop to where we found it
16901 instead of using a fixed path.
16905 * SHA library changes for irix64-mips4-cc.
16909 * Improvements for VMS support.
16913 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16915 * Bignum library bug fix. IRIX 6 passes "make test" now!
16916 This also avoids the problems with SC4.2 and unpatched SC5.
16918 *Andy Polyakov <appro@fy.chalmers.se>*
16920 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16921 These are required because of the typesafe stack would otherwise break
16922 existing code. If old code used a structure member which used to be STACK
16923 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16924 sk_num or sk_value it would produce an error because the num, data members
16925 are not present in STACK_OF. Now it just produces a warning. sk_set
16926 replaces the old method of assigning a value to sk_value
16927 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16928 that does this will no longer work (and should use sk_set instead) but
16929 this could be regarded as a "questionable" behaviour anyway.
16933 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16934 correctly handle encrypted S/MIME data.
16938 * Change type of various DES function arguments from des_cblock
16939 (which means, in function argument declarations, pointer to char)
16940 to des_cblock * (meaning pointer to array with 8 char elements),
16941 which allows the compiler to do more typechecking; it was like
16942 that back in SSLeay, but with lots of ugly casts.
16944 Introduce new type const_des_cblock.
16948 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16949 problems: find RecipientInfo structure that matches recipient certificate
16950 and initialise the ASN1 structures properly based on passed cipher.
16954 * Belatedly make the BN tests actually check the results.
16958 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16959 to and from BNs: it was completely broken. New compilation option
16960 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16961 key elements as negative integers.
16965 * Reorganize and speed up MD5.
16967 *Andy Polyakov <appro@fy.chalmers.se>*
16971 *Richard Levitte <richard@levitte.org>*
16973 * New option -out to asn1parse to allow the parsed structure to be
16974 output to a file. This is most useful when combined with the -strparse
16975 option to examine the output of things like OCTET STRINGS.
16979 * Make SSL library a little more fool-proof by not requiring any longer
16980 that `SSL_set_{accept,connect}_state` be called before
16981 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16982 in many applications because usually everything *appeared* to work as
16983 intended anyway -- now it really works as intended).
16987 * Move openssl.cnf out of lib/.
16991 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16992 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16993 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16995 *Ralf S. Engelschall*
16997 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16998 handle PKCS#7 enveloped data properly.
17000 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17002 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17003 copying pointers. The cert_st handling is changed by this in
17004 various ways (and thus what used to be known as ctx->default_cert
17005 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17006 any longer when s->cert does not give us what we need).
17007 ssl_cert_instantiate becomes obsolete by this change.
17008 As soon as we've got the new code right (possibly it already is?),
17009 we have solved a couple of bugs of the earlier code where s->cert
17010 was used as if it could not have been shared with other SSL structures.
17012 Note that using the SSL API in certain dirty ways now will result
17013 in different behaviour than observed with earlier library versions:
17014 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17015 does not influence s as it used to.
17017 In order to clean up things more thoroughly, inside SSL_SESSION
17018 we don't use CERT any longer, but a new structure SESS_CERT
17019 that holds per-session data (if available); currently, this is
17020 the peer's certificate chain and, for clients, the server's certificate
17021 and temporary key. CERT holds only those values that can have
17022 meaningful defaults in an SSL_CTX.
17026 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17027 from the internal representation. Various PKCS#7 fixes: remove some
17028 evil casts and set the enc_dig_alg field properly based on the signing
17033 * Allow PKCS#12 password to be set from the command line or the
17034 environment. Let 'ca' get its config file name from the environment
17035 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17040 * Allow certificate policies extension to use an IA5STRING for the
17041 organization field. This is contrary to the PKIX definition but
17042 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17047 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17048 without disallowing inline assembler and the like for non-pedantic builds.
17052 * Support Borland C++ builder.
17054 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17060 * SHA-1 cleanups and performance enhancements.
17062 *Andy Polyakov <appro@fy.chalmers.se>*
17064 * Sparc v8plus assembler for the bignum library.
17066 *Andy Polyakov <appro@fy.chalmers.se>*
17068 * Accept any -xxx and +xxx compiler options in Configure.
17072 * Update HPUX configuration.
17076 * Add missing `sk_<type>_unshift()` function to safestack.h
17078 *Ralf S. Engelschall*
17080 * New function SSL_CTX_use_certificate_chain_file that sets the
17081 "extra_cert"s in addition to the certificate. (This makes sense
17082 only for "PEM" format files, as chains as a whole are not
17087 * Support verify_depth from the SSL API.
17088 x509_vfy.c had what can be considered an off-by-one-error:
17089 Its depth (which was not part of the external interface)
17090 was actually counting the number of certificates in a chain;
17091 now it really counts the depth.
17095 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17096 instead of X509err, which often resulted in confusing error
17097 messages since the error codes are not globally unique
17098 (e.g. an alleged error in ssl3_accept when a certificate
17099 didn't match the private key).
17101 * New function SSL_CTX_set_session_id_context that allows to set a default
17102 value (so that you don't need SSL_set_session_id_context for each
17103 connection using the SSL_CTX).
17107 * OAEP decoding bug fix.
17111 * Support INSTALL_PREFIX for package builders, as proposed by
17116 * New Configure options "threads" and "no-threads". For systems
17117 where the proper compiler options are known (currently Solaris
17118 and Linux), "threads" is the default.
17122 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17126 * Install various scripts to $(OPENSSLDIR)/misc, not to
17127 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17128 such as /usr/local/bin.
17132 * "make linux-shared" to build shared libraries.
17134 *Niels Poppe <niels@netbox.org>*
17136 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17140 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17141 extension adding in x509 utility.
17145 * Remove NOPROTO sections and error code comments.
17149 * Partial rewrite of the DEF file generator to now parse the ANSI
17154 * New Configure options --prefix=DIR and --openssldir=DIR.
17158 * Complete rewrite of the error code script(s). It is all now handled
17159 by one script at the top level which handles error code gathering,
17160 header rewriting and C source file generation. It should be much better
17161 than the old method: it now uses a modified version of Ulf's parser to
17162 read the ANSI prototypes in all header files (thus the old K&R definitions
17163 aren't needed for error creation any more) and do a better job of
17164 translating function codes into names. The old 'ASN1 error code embedded
17165 in a comment' is no longer necessary and it doesn't use .err files which
17166 have now been deleted. Also the error code call doesn't have to appear all
17167 on one line (which resulted in some large lines...).
17171 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17175 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17176 0 (which usually indicates a closed connection), but continue reading.
17180 * Fix some race conditions.
17184 * Add support for CRL distribution points extension. Add Certificate
17185 Policies and CRL distribution points documentation.
17189 * Move the autogenerated header file parts to crypto/opensslconf.h.
17193 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17194 8 of keying material. Merlin has also confirmed interop with this fix
17195 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17197 *Merlin Hughes <merlin@baltimore.ie>*
17199 * Fix lots of warnings.
17201 *Richard Levitte <levitte@stacken.kth.se>*
17203 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17204 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17206 *Richard Levitte <levitte@stacken.kth.se>*
17208 * Fix problems with sizeof(long) == 8.
17210 *Andy Polyakov <appro@fy.chalmers.se>*
17212 * Change functions to ANSI C.
17216 * Fix typos in error codes.
17218 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17220 * Remove defunct assembler files from Configure.
17224 * SPARC v8 assembler BIGNUM implementation.
17226 *Andy Polyakov <appro@fy.chalmers.se>*
17228 * Support for Certificate Policies extension: both print and set.
17229 Various additions to support the r2i method this uses.
17233 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17234 return a const string when you are expecting an allocated buffer.
17238 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17239 types DirectoryString and DisplayText.
17243 * Add code to allow r2i extensions to access the configuration database,
17244 add an LHASH database driver and add several ctx helper functions.
17248 * Fix an evil bug in bn_expand2() which caused various BN functions to
17249 fail when they extended the size of a BIGNUM.
17253 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17254 support typesafe stack.
17258 * Fix typo in SSL_[gs]et_options().
17260 *Nils Frostberg <nils@medcom.se>*
17262 * Delete various functions and files that belonged to the (now obsolete)
17263 old X509V3 handling code.
17267 * New Configure option "rsaref".
17271 * Don't auto-generate pem.h.
17275 * Introduce type-safe ASN.1 SETs.
17279 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17281 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17283 * Introduce type-safe STACKs. This will almost certainly break lots of code
17284 that links with OpenSSL (well at least cause lots of warnings), but fear
17285 not: the conversion is trivial, and it eliminates loads of evil casts. A
17286 few STACKed things have been converted already. Feel free to convert more.
17287 In the fullness of time, I'll do away with the STACK type altogether.
17291 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17292 specified in `<certfile>` by updating the entry in the index.txt file.
17293 This way one no longer has to edit the index.txt file manually for
17294 revoking a certificate. The -revoke option does the gory details now.
17296 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17298 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17299 `-text` option at all and this way the `-noout -text` combination was
17300 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17302 *Ralf S. Engelschall*
17304 * Make sure a corresponding plain text error message exists for the
17305 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17306 verify callback function determined that a certificate was revoked.
17308 *Ralf S. Engelschall*
17310 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17311 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17312 all available ciphers including rc5, which was forgotten until now.
17313 In order to let the testing shell script know which algorithms
17314 are available, a new (up to now undocumented) command
17315 `openssl list-cipher-commands` is used.
17319 * Bugfix: s_client occasionally would sleep in select() when
17320 it should have checked SSL_pending() first.
17324 * New functions DSA_do_sign and DSA_do_verify to provide access to
17325 the raw DSA values prior to ASN.1 encoding.
17329 * Tweaks to Configure
17331 *Niels Poppe <niels@netbox.org>*
17333 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17338 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17342 * New config option to avoid instructions that are illegal on the 80386.
17343 The default code is faster, but requires at least a 486.
17347 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17348 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17349 same as SSL2_VERSION anyway.
17353 * New "-showcerts" option for s_client.
17357 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17358 application. Various cleanups and fixes.
17362 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17363 modify error routines to work internally. Add error codes and PBE init
17364 to library startup routines.
17368 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17369 packing functions to asn1 and evp. Changed function names and error
17370 codes along the way.
17374 * PKCS12 integration: and so it begins... First of several patches to
17375 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17376 objects to objects.h
17380 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17381 and display support for Thawte strong extranet extension.
17385 * Add LinuxPPC support.
17387 *Jeff Dubrule <igor@pobox.org>*
17389 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17390 bn_div_words in alpha.s.
17392 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17394 * Make sure the RSA OAEP test is skipped under -DRSAref because
17395 OAEP isn't supported when OpenSSL is built with RSAref.
17397 *Ulf Moeller <ulf@fitug.de>*
17399 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17400 so they no longer are missing under -DNOPROTO.
17402 *Soren S. Jorvang <soren@t.dk>*
17404 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17406 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17407 doesn't work when the session is reused. Coming soon!
17411 * Fix a security hole, that allows sessions to be reused in the wrong
17412 context thus bypassing client cert protection! All software that uses
17413 client certs and session caches in multiple contexts NEEDS PATCHING to
17414 allow session reuse! A fuller solution is in the works.
17416 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17418 * Some more source tree cleanups (removed obsolete files
17419 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17420 permission on "config" script to be executable) and a fix for the INSTALL
17423 *Ulf Moeller <ulf@fitug.de>*
17425 * Remove some legacy and erroneous uses of malloc, free instead of
17428 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17430 * Make rsa_oaep_test return non-zero on error.
17432 *Ulf Moeller <ulf@fitug.de>*
17434 * Add support for native Solaris shared libraries. Configure
17435 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17436 if someone would make that last step automatic.
17438 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17440 * ctx_size was not built with the right compiler during "make links". Fixed.
17444 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17445 except NULL ciphers". This means the default cipher list will no longer
17446 enable NULL ciphers. They need to be specifically enabled e.g. with
17447 the string "DEFAULT:eNULL".
17451 * Fix to RSA private encryption routines: if p < q then it would
17452 occasionally produce an invalid result. This will only happen with
17453 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17457 * Be less restrictive and allow also `perl util/perlpath.pl
17458 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17459 because this way one can also use an interpreter named `perl5` (which is
17460 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17461 installed as `perl`).
17463 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17465 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17467 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17469 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17470 advapi32.lib to Win32 build and change the pem test comparison
17471 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17472 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17473 and crypto/des/ede_cbcm_enc.c.
17477 * DES quad checksum was broken on big-endian architectures. Fixed.
17481 * Comment out two functions in bio.h that aren't implemented. Fix up the
17482 Win32 test batch file so it (might) work again. The Win32 test batch file
17483 is horrible: I feel ill....
17487 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17488 in e_os.h. Audit of header files to check ANSI and non ANSI
17489 sections: 10 functions were absent from non ANSI section and not exported
17490 from Windows DLLs. Fixed up libeay.num for new functions.
17494 * Make `openssl version` output lines consistent.
17496 *Ralf S. Engelschall*
17498 * Fix Win32 symbol export lists for BIO functions: Added
17499 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17500 to ms/libeay{16,32}.def.
17502 *Ralf S. Engelschall*
17504 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17505 fine under Unix and passes some trivial tests I've now added. But the
17506 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17507 added to make sure no one expects that this stuff really works in the
17508 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17509 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17512 *Ralf S. Engelschall*
17514 * Fix the generation of two part addresses in perl.
17516 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17518 * Add config entry for Linux on MIPS.
17520 *John Tobey <jtobey@channel1.com>*
17522 * Make links whenever Configure is run, unless we are on Windoze.
17526 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17527 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17532 * Add a useful kludge to allow package maintainers to specify compiler and
17533 other platforms details on the command line without having to patch the
17534 Configure script every time: One now can use
17535 `perl Configure <id>:<details>`,
17536 i.e. platform ids are allowed to have details appended
17537 to them (separated by colons). This is treated as there would be a static
17538 pre-configured entry in Configure's %table under key `<id>` with value
17539 `<details>` and `perl Configure <id>` is called. So, when you want to
17540 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17541 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17542 now, which overrides the FreeBSD-elf entry on-the-fly.
17544 *Ralf S. Engelschall*
17546 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17550 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17551 on the `perl Configure ...` command line. This way one can compile
17552 OpenSSL libraries with Position Independent Code (PIC) which is needed
17553 for linking it into DSOs.
17555 *Ralf S. Engelschall*
17557 * Remarkably, export ciphers were totally broken and no-one had noticed!
17562 * Cleaned up the LICENSE document: The official contact for any license
17563 questions now is the OpenSSL core team under openssl-core@openssl.org.
17564 And add a paragraph about the dual-license situation to make sure people
17565 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17566 to the OpenSSL toolkit.
17568 *Ralf S. Engelschall*
17570 * General source tree makefile cleanups: Made `making xxx in yyy...`
17571 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17572 Additionally cleaned up the `make links` target: Remove unnecessary
17573 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17574 to speed processing and no longer clutter the display with confusing
17575 stuff. Instead only the actually done links are displayed.
17577 *Ralf S. Engelschall*
17579 * Permit null encryption ciphersuites, used for authentication only. It used
17580 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17581 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17586 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17587 signed attributes when verifying signatures (this would break them),
17588 the detached data encoding was wrong and public keys obtained using
17589 X509_get_pubkey() weren't freed.
17593 * Add text documentation for the BUFFER functions. Also added a work around
17594 to a Win95 console bug. This was triggered by the password read stuff: the
17595 last character typed gets carried over to the next fread(). If you were
17596 generating a new cert request using 'req' for example then the last
17597 character of the passphrase would be CR which would then enter the first
17602 * Added the new 'Includes OpenSSL Cryptography Software' button as
17603 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17604 button and can be used by applications based on OpenSSL to show the
17605 relationship to the OpenSSL project.
17607 *Ralf S. Engelschall*
17609 * Remove confusing variables in function signatures in files
17610 ssl/ssl_lib.c and ssl/ssl.h.
17612 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17614 * Don't install bss_file.c under PREFIX/include/
17616 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17618 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17619 functions that return function pointers and has support for NT specific
17620 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17621 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17622 unsigned to signed types: this was killing the Win32 compile.
17626 * Add new certificate file to stack functions,
17627 SSL_add_dir_cert_subjects_to_stack() and
17628 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17629 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17630 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17631 This means that Apache-SSL and similar packages don't have to mess around
17632 to add as many CAs as they want to the preferred list.
17636 * Experiment with doxygen documentation. Currently only partially applied to
17638 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17639 openssl.doxy as the configuration file.
17643 * Get rid of remaining C++-style comments which strict C compilers hate.
17645 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17647 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17648 compiled in by default: it has problems with large keys.
17652 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17653 DH private keys and/or callback functions which directly correspond to
17654 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17655 is needed for applications which have to configure certificates on a
17656 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17658 For the RSA certificate situation is makes no difference, but
17659 for the DSA certificate situation this fixes the "no shared cipher"
17660 problem where the OpenSSL cipher selection procedure failed because the
17661 temporary keys were not overtaken from the context and the API provided
17662 no way to reconfigure them.
17663 The new functions now let applications reconfigure the stuff and they
17664 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17665 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17666 non-public-API function ssl_cert_instantiate() is used as a helper
17667 function and also to reduce code redundancy inside ssl_rsa.c.
17669 *Ralf S. Engelschall*
17671 * Move s_server -dcert and -dkey options out of the undocumented feature
17672 area because they are useful for the DSA situation and should be
17673 recognized by the users.
17675 *Ralf S. Engelschall*
17677 * Fix the cipher decision scheme for export ciphers: the export bits are
17678 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17679 SSL_EXP_MASK. So, the original variable has to be used instead of the
17680 already masked variable.
17682 *Richard Levitte <levitte@stacken.kth.se>*
17684 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17686 *Richard Levitte <levitte@stacken.kth.se>*
17688 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17689 from `int` to `unsigned int` because it is a length and initialized by
17690 EVP_DigestFinal() which expects an `unsigned int *`.
17692 *Richard Levitte <levitte@stacken.kth.se>*
17694 * Don't hard-code path to Perl interpreter on shebang line of Configure
17695 script. Instead use the usual Shell->Perl transition trick.
17697 *Ralf S. Engelschall*
17699 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17700 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17701 -noout -modulus` as it's already the case for `openssl rsa -noout
17702 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17703 currently the public key is printed (a decision which was already done by
17704 `openssl dsa -modulus` in the past) which serves a similar purpose.
17705 Additionally the NO_RSA no longer completely removes the whole -modulus
17706 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17709 *Ralf S. Engelschall*
17711 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17712 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17714 *Arne Ansper <arne@ats.cyber.ee>*
17716 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17717 to be added. Now both 'req' and 'ca' can use new objects defined in the
17722 * Add cool BIO that does syslog (or event log on NT).
17724 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17726 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17727 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17728 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17729 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17733 * Add preliminary config info for new extension code.
17737 * Make RSA_NO_PADDING really use no padding.
17739 *Ulf Moeller <ulf@fitug.de>*
17741 * Generate errors when private/public key check is done.
17745 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17746 for some CRL extensions and new objects added.
17750 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17751 key usage extension and fuller support for authority key id.
17755 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17756 padding method for RSA, which is recommended for new applications in PKCS
17757 #1 v2.0 (RFC 2437, October 1998).
17758 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17759 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17760 against Bleichbacher's attack on RSA.
17761 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17764 * Updates to the new SSL compression code
17766 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17768 * Fix so that the version number in the master secret, when passed
17769 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17770 (because the server will not accept higher), that the version number
17771 is 0x03,0x01, not 0x03,0x00
17773 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17775 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17776 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17777 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17781 * Support for RAW extensions where an arbitrary extension can be
17782 created by including its DER encoding. See apps/openssl.cnf for
17787 * Make sure latest Perl versions don't interpret some generated C array
17788 code as Perl array code in the crypto/err/err_genc.pl script.
17790 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17792 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17793 not many people have the assembler. Various Win32 compilation fixes and
17794 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17795 build instructions.
17799 * Modify configure script 'Configure' to automatically create crypto/date.h
17800 file under Win32 and also build pem.h from pem.org. New script
17801 util/mkfiles.pl to create the MINFO file on environments that can't do a
17802 'make files': perl util/mkfiles.pl >MINFO should work.
17806 * Major rework of DES function declarations, in the pursuit of correctness
17807 and purity. As a result, many evil casts evaporated, and some weirdness,
17808 too. You may find this causes warnings in your code. Zapping your evil
17809 casts will probably fix them. Mostly.
17813 * Fix for a typo in asn1.h. Bug fix to object creation script
17814 obj_dat.pl. It considered a zero in an object definition to mean
17815 "end of object": none of the objects in objects.h have any zeros
17816 so it wasn't spotted.
17818 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17820 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17821 Masking (CBCM). In the absence of test vectors, the best I have been able
17822 to do is check that the decrypt undoes the encrypt, so far. Send me test
17823 vectors if you have them.
17827 * Correct calculation of key length for export ciphers (too much space was
17828 allocated for null ciphers). This has not been tested!
17832 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17833 message is now correct (it understands "crypto" and "ssl" on its
17834 command line). There is also now an "update" option. This will update
17835 the util/ssleay.num and util/libeay.num files with any new functions.
17837 perl util/mkdef.pl crypto ssl update
17838 it will update them.
17842 * Overhauled the Perl interface:
17843 - ported BN stuff to OpenSSL's different BN library
17844 - made the perl/ source tree CVS-aware
17845 - renamed the package from SSLeay to OpenSSL (the files still contain
17846 their history because I've copied them in the repository)
17847 - removed obsolete files (the test scripts will be replaced
17848 by better Test::Harness variants in the future)
17850 *Ralf S. Engelschall*
17852 * First cut for a very conservative source tree cleanup:
17853 1. merge various obsolete readme texts into doc/ssleay.txt
17854 where we collect the old documents and readme texts.
17855 2. remove the first part of files where I'm already sure that we no
17856 longer need them because of three reasons: either they are just temporary
17857 files which were left by Eric or they are preserved original files where
17858 I've verified that the diff is also available in the CVS via "cvs diff
17859 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17860 the crypto/md/ stuff).
17862 *Ralf S. Engelschall*
17864 * More extension code. Incomplete support for subject and issuer alt
17865 name, issuer and authority key id. Change the i2v function parameters
17866 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17867 what that's for :-) Fix to ASN1 macro which messed up
17868 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17872 * Preliminary support for ENUMERATED type. This is largely copied from the
17877 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17879 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17881 * Make sure `make rehash` target really finds the `openssl` program.
17883 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17885 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17886 like to hear about it if this slows down other processors.
17890 * Add CygWin32 platform information to Configure script.
17892 *Alan Batie <batie@aahz.jf.intel.com>*
17894 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17896 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17898 * New program nseq to manipulate netscape certificate sequences
17902 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17907 * Fixes to BN code. Previously the default was to define BN_RECURSION
17908 but the BN code had some problems that would cause failures when
17909 doing certificate verification and some other functions.
17911 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17913 * Add ASN1 and PEM code to support netscape certificate sequences.
17917 * Add ASN1 and PEM code to support netscape certificate sequences.
17921 * Add several PKIX and private extended key usage OIDs.
17925 * Modify the 'ca' program to handle the new extension code. Modify
17926 openssl.cnf for new extension format, add comments.
17930 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17931 and add a sample to openssl.cnf so req -x509 now adds appropriate
17936 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17937 error code, add initial support to X509_print() and x509 application.
17941 * Takes a deep breath and start adding X509 V3 extension support code. Add
17942 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17943 stuff is currently isolated and isn't even compiled yet.
17947 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17948 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17949 Removed the versions check from X509 routines when loading extensions:
17950 this allows certain broken certificates that don't set the version
17951 properly to be processed.
17955 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17956 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17957 can still be regenerated with "make depend".
17961 * Spelling mistake in C version of CAST-128.
17963 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17965 * Changes to the error generation code. The perl script err-code.pl
17966 now reads in the old error codes and retains the old numbers, only
17967 adding new ones if necessary. It also only changes the .err files if new
17968 codes are added. The makefiles have been modified to only insert errors
17969 when needed (to avoid needlessly modifying header files). This is done
17970 by only inserting errors if the .err file is newer than the auto generated
17971 C file. To rebuild all the error codes from scratch (the old behaviour)
17972 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17973 or delete all the .err files.
17977 * CAST-128 was incorrectly implemented for short keys. The C version has
17978 been fixed, but is untested. The assembler versions are also fixed, but
17979 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17980 to regenerate it if needed.
17981 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17982 Hagino <itojun@kame.net>*
17984 * File was opened incorrectly in randfile.c.
17986 *Ulf Möller <ulf@fitug.de>*
17988 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17989 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17990 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17991 al: it's just almost always a UTCTime. Note this patch adds new error
17992 codes so do a "make errors" if there are problems.
17996 * Correct Linux 1 recognition in config.
17998 *Ulf Möller <ulf@fitug.de>*
18000 * Remove pointless MD5 hash when using DSA keys in ca.
18002 *Anonymous <nobody@replay.com>*
18004 * Generate an error if given an empty string as a cert directory. Also
18005 generate an error if handed NULL (previously returned 0 to indicate an
18006 error, but didn't set one).
18008 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18010 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18014 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18015 parameters. This was causing a warning which killed off the Win32 compile.
18019 * Remove C++ style comments from crypto/bn/bn_local.h.
18021 *Neil Costigan <neil.costigan@celocom.com>*
18023 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18024 based on a text string, looking up short and long names and finally
18025 "dot" format. The "dot" format stuff didn't work. Added new function
18026 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18027 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18028 OID is not part of the table.
18032 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18033 X509_LOOKUP_by_alias().
18037 * Sort openssl functions by name.
18041 * Get the gendsa program working (hopefully) and add it to app list. Remove
18042 encryption from sample DSA keys (in case anyone is interested the password
18047 * Make *all* `*_free` functions accept a NULL pointer.
18049 *Frans Heymans <fheymans@isaserver.be>*
18051 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18054 *Anonymous <nobody@replay.com>*
18056 * s_server should send the CAfile as acceptable CAs, not its own cert.
18058 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18060 * Don't blow it for numeric -newkey arguments to apps/req.
18062 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18064 * Temp key "for export" tests were wrong in s3_srvr.c.
18066 *Anonymous <nobody@replay.com>*
18068 * Add prototype for temp key callback functions
18069 SSL_CTX_set_tmp_{rsa,dh}_callback().
18073 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18074 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18078 * X509_name_add_entry() freed the wrong thing after an error.
18080 *Arne Ansper <arne@ats.cyber.ee>*
18082 * rsa_eay.c would attempt to free a NULL context.
18084 *Arne Ansper <arne@ats.cyber.ee>*
18086 * BIO_s_socket() had a broken should_retry() on Windoze.
18088 *Arne Ansper <arne@ats.cyber.ee>*
18090 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18092 *Arne Ansper <arne@ats.cyber.ee>*
18094 * Make sure the already existing X509_STORE->depth variable is initialized
18095 in X509_STORE_new(), but document the fact that this variable is still
18096 unused in the certificate verification process.
18098 *Ralf S. Engelschall*
18100 * Fix the various library and apps files to free up pkeys obtained from
18101 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18105 * Fix reference counting in X509_PUBKEY_get(). This makes
18106 demos/maurice/example2.c work, amongst others, probably.
18108 *Steve Henson and Ben Laurie*
18110 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18111 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18112 are no longer created. This way we have a single and consistent command
18113 line interface `openssl <command>`, similar to `cvs <command>`.
18115 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18117 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18118 BIT STRING wrapper always have zero unused bits.
18122 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18126 * Make the top-level INSTALL documentation easier to understand.
18130 * Makefiles updated to exit if an error occurs in a sub-directory
18131 make (including if user presses ^C) [Paul Sutton]
18133 * Make Montgomery context stuff explicit in RSA data structure.
18137 * Fix build order of pem and err to allow for generated pem.h.
18141 * Fix renumbering bug in X509_NAME_delete_entry().
18145 * Enhanced the err-ins.pl script so it makes the error library number
18146 global and can add a library name. This is needed for external ASN1 and
18147 other error libraries.
18151 * Fixed sk_insert which never worked properly.
18155 * Fix ASN1 macros so they can handle indefinite length constructed
18156 EXPLICIT tags. Some non standard certificates use these: they can now
18161 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18162 into a single doc/ssleay.txt bundle. This way the information is still
18163 preserved but no longer messes up this directory. Now it's new room for
18164 the new set of documentation files.
18166 *Ralf S. Engelschall*
18168 * SETs were incorrectly DER encoded. This was a major pain, because they
18169 shared code with SEQUENCEs, which aren't coded the same. This means that
18170 almost everything to do with SETs or SEQUENCEs has either changed name or
18171 number of arguments.
18173 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18175 * Fix test data to work with the above.
18179 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18180 was already fixed by Eric for 0.9.1 it seems.
18182 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18184 * Autodetect FreeBSD3.
18188 * Fix various bugs in Configure. This affects the following platforms:
18192 unixware-2.0-pentium
18197 * Eliminate generated files from CVS. Reorder tests to regenerate files
18198 before they are needed.
18202 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18206 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18208 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18209 changed SSLeay to OpenSSL in version strings.
18211 *Ralf S. Engelschall*
18213 * Some fixups to the top-level documents.
18217 * Fixed the nasty bug where rsaref.h was not found under compile-time
18218 because the symlink to include/ was missing.
18220 *Ralf S. Engelschall*
18222 * Incorporated the popular no-RSA/DSA-only patches
18223 which allow to compile a RSA-free SSLeay.
18225 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18227 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18228 when "ssleay" is still not found.
18230 *Ralf S. Engelschall*
18232 * Added more platforms to Configure: Cray T3E, HPUX 11,
18234 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18236 * Updated the README file.
18238 *Ralf S. Engelschall*
18240 * Added various .cvsignore files in the CVS repository subdirs
18241 to make a "cvs update" really silent.
18243 *Ralf S. Engelschall*
18245 * Recompiled the error-definition header files and added
18246 missing symbols to the Win32 linker tables.
18248 *Ralf S. Engelschall*
18250 * Cleaned up the top-level documents;
18251 o new files: CHANGES and LICENSE
18252 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18253 o merged COPYRIGHT into LICENSE
18254 o removed obsolete TODO file
18255 o renamed MICROSOFT to INSTALL.W32
18257 *Ralf S. Engelschall*
18259 * Removed dummy files from the 0.9.1b source tree:
18260 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18261 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18262 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18263 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18264 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18266 *Ralf S. Engelschall*
18268 * Added various platform portability fixes.
18272 * The Genesis of the OpenSSL rpject:
18273 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18274 Young and Tim J. Hudson created while they were working for C2Net until
18277 *The OpenSSL Project*
18279 ### Changes between 0.9.0b and 0.9.1b [not released]
18281 * Updated a few CA certificates under certs/
18285 * Changed some BIGNUM api stuff.
18289 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18290 DGUX x86, Linux Alpha, etc.
18294 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18295 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18300 * Add -strparse option to asn1pars program which parses nested
18303 *Dr Stephen Henson <shenson@bigfoot.com>*
18305 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18309 * DSA fix for "ca" program.
18313 * Added "-genkey" option to "dsaparam" program.
18317 * Added RIPE MD160 (rmd160) message digest.
18321 * Added -a (all) option to "ssleay version" command.
18325 * Added PLATFORM define which is the id given to Configure.
18329 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18333 * Extended the ASN.1 parser routines.
18337 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18341 * Added a BN_CTX to the BN library.
18345 * Fixed the weak key values in DES library
18349 * Changed API in EVP library for cipher aliases.
18353 * Added support for RC2/64bit cipher.
18357 * Converted the lhash library to the crypto/mem.c functions.
18361 * Added more recognized ASN.1 object ids.
18365 * Added more RSA padding checks for SSL/TLS.
18369 * Added BIO proxy/filter functionality.
18373 * Added extra_certs to SSL_CTX which can be used
18374 send extra CA certificates to the client in the CA cert chain sending
18375 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18379 * Now Fortezza is denied in the authentication phase because
18380 this is key exchange mechanism is not supported by SSLeay at all.
18384 * Additional PKCS1 checks.
18388 * Support the string "TLSv1" for all TLS v1 ciphers.
18392 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18393 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18397 * Fixed a few memory leaks.
18401 * Fixed various code and comment typos.
18405 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18406 bytes sent in the client random.
18408 *Edward Bishop <ebishop@spyglass.com>*
18412 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18413 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18414 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18415 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18416 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18417 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18418 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18419 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18420 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18421 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18422 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18423 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18424 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18425 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18426 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18427 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18428 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18429 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18430 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18431 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18432 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18433 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18434 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18435 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18436 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18437 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18438 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18439 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18440 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18441 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18442 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18443 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18444 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18445 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18446 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18447 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18448 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18449 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18450 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18451 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18452 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18453 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18454 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18455 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18456 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18457 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18458 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18459 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18460 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18461 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18462 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18463 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18464 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18465 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18466 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18467 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18468 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18469 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18470 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18471 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18472 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18473 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18474 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18475 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18476 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18477 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18478 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18479 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18480 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18481 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18482 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18483 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18484 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18485 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18486 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18487 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18488 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18489 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18490 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18491 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18492 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18493 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18494 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18495 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18496 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18497 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18498 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18499 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18500 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18501 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18502 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18503 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18504 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18505 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18506 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18507 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18508 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18509 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18510 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18511 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18512 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18513 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18514 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18515 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18516 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18517 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18518 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18519 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18520 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18521 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18522 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18523 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18524 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18525 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18526 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18527 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18528 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18529 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18530 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18531 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18532 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18533 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18534 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18535 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18536 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18537 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18538 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18539 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18540 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18541 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18542 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18543 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18544 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18545 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18546 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18547 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18548 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18549 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18550 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18551 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18552 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18553 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18554 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18555 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18556 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18557 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18558 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18559 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18560 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18561 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18562 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18563 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18564 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18565 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18566 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18567 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18568 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18569 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18570 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18571 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655