4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * Deprecated the `ENGINE` API. Engines should be replaced with providers
31 * Reworked the recorded ERR codes to make better space for system errors.
32 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
33 given code is a system error (true) or an OpenSSL error (false).
37 * Reworked the test perl framework to better allow parallel testing.
39 *Nicola Tuveri and David von Oheimb*
41 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
42 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
46 * 'Configure' has been changed to figure out the configuration target if
47 none is given on the command line. Consequently, the 'config' script is
48 now only a mere wrapper. All documentation is changed to only mention
51 *Rich Salz and Richard Levitte*
53 * Added a library context that applications as well as other
54 libraries can use to form a separate context within which libcrypto
55 operations are performed.
57 There are two ways this can be used:
59 - Directly, by passing a library context to functions that take
60 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
62 - Indirectly, by creating a new library context and then assigning
63 it as the new default, with `OPENSSL_CTX_set0_default`.
65 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
66 apart from the functions directly related to `OPENSSL_CTX`, accept
67 NULL to indicate that the default library context should be used.
69 Library code that changes the default library context using
70 `OPENSSL_CTX_set0_default` should take care to restore it with a
71 second call before returning to the caller.
75 * Handshake now fails if Extended Master Secret extension is dropped
80 * Dropped interactive mode from the 'openssl' program. From now on,
81 the `openssl` command without arguments is equivalent to `openssl
86 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
87 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
88 While the old function names have been retained for backward compatibility
89 they should not be used in new developments
90 because their return values are confusing: Unlike other `_cmp()` functions
91 they do not return 0 in case their arguments are equal.
95 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
96 EC_GROUP_get_field_type().
100 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
101 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
102 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
103 Applications should rely on the library automatically assigning a suitable
104 EC_METHOD internally upon EC_GROUP construction.
108 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
109 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
110 assigned internally without application intervention.
111 Users of EC_GROUP_new() should switch to a different suitable constructor.
115 * Add CAdES-BES signature verification support, mostly derived
116 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
118 *Filipe Raimundo da Silva*
120 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
124 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
125 functions are not widely used and now OpenSSL automatically perform this
126 conversion when needed.
130 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
131 EC_KEY_precompute_mult(). These functions are not widely used and
132 applications should instead switch to named curves which OpenSSL has
133 hardcoded lookup tables for.
137 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
138 should instead use the L<EC_POINT_mul(3)> function.
142 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
143 that are not applicable to the new provider model. Applications should
144 instead use EVP_default_properties_is_fips_enabled() and
145 EVP_default_properties_enable_fips().
149 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
150 is set, an unexpected EOF is ignored, it pretends a close notify was received
151 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
155 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
156 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
157 used and applications should instead use the
158 L<EC_POINT_set_affine_coordinates(3)> and
159 L<EC_POINT_get_affine_coordinates(3)> functions.
163 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
164 arrays to be more easily constructed via a series of utility functions.
165 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
166 the various push functions and finally convert to a passable OSSL_PARAM
167 array using OSSL_PARAM_BLD_to_param().
171 * The security strength of SHA1 and MD5 based signatures in TLS has been
172 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
173 working at the default security level of 1 and instead requires security
174 level 0. The security level can be changed either using the cipher string
175 with @SECLEVEL, or calling SSL_CTX_set_security_level().
179 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
180 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
181 internal keys, if they correspond to one of those built in types.
185 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
186 contain a provider side internal key.
190 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
191 They are old functions that we don't use, and that you could disable with
192 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
196 * Project text documents not yet having a proper file name extension
197 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
198 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
202 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
203 have been converted to Markdown with the goal to produce documents
204 which not only look pretty when viewed online in the browser, but
205 remain well readable inside a plain text editor.
207 To achieve this goal, a 'minimalistic' Markdown style has been applied
208 which avoids formatting elements that interfere too much with the
209 reading flow in the text file. For example, it
211 * avoids [ATX headings][] and uses [setext headings][] instead
212 (which works for `<h1>` and `<h2>` headings only).
213 * avoids [inline links][] and uses [reference links][] instead.
214 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
216 [ATX headings]: https://github.github.com/gfm/#atx-headings
217 [setext headings]: https://github.github.com/gfm/#setext-headings
218 [inline links]: https://github.github.com/gfm/#inline-link
219 [reference links]: https://github.github.com/gfm/#reference-link
220 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
221 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
223 *Matthias St. Pierre*
225 * The test suite is changed to preserve results of each test recipe.
226 A new directory test-runs/ with subdirectories named like the
227 test recipes are created in the build tree for this purpose.
231 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
232 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
233 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
235 *David von Oheimb, Martin Peylo*
237 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
238 The legacy OCSP-focused and only partly documented API is retained.
239 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
243 * Added `util/check-format.pl`, a tool for checking adherence to the
244 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
245 The checks performed are incomplete and yield some false positives.
246 Still the tool should be useful for detecting most typical glitches.
250 * BIO_do_connect and BIO_do_handshake have been extended:
251 If domain name resolution yields multiple IP addresses all of them are tried
252 after connect() failures.
256 * All of the low level RSA functions have been deprecated including:
258 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
259 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
260 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
261 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
262 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
263 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
264 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
265 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
266 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
267 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
268 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
269 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
270 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
271 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
272 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
273 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
274 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
275 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
276 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
277 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
278 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
279 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
280 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
281 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
282 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
283 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
284 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
285 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
287 Use of these low level functions has been informally discouraged for a long
288 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
289 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
290 L<EVP_PKEY_decrypt(3)>.
294 * X509 certificates signed using SHA1 are no longer allowed at security
296 In TLS/SSL the default security level is 1. It can be set either
297 using the cipher string with @SECLEVEL, or calling
298 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
299 a call to SSL_CTX_use_certificate() will fail if the security level is not
301 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
302 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
307 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
308 modified to use PKEY APIs. These commands are now in maintenance mode
309 and no new features will be added to them.
313 * The command line utility rsautl has been deprecated.
314 Instead use the pkeyutl program.
318 * The command line utilities genrsa and rsa have been modified to use PKEY
319 APIs These commands are now in maintenance mode and no new features will
324 * All of the low level DH functions have been deprecated including:
326 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
327 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
328 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
329 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
330 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
331 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
332 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
333 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
334 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
335 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
336 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
337 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
338 DH_meth_get_generate_params and DH_meth_set_generate_params.
340 Use of these low level functions has been informally discouraged for a long
341 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
342 and L<EVP_PKEY_derive(3)>.
346 * All of the low level DSA functions have been deprecated including:
348 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
349 DSA_get_default_method, DSA_set_method, DSA_get_method,
350 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
351 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
352 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
353 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
354 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
355 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
356 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
357 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
358 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
359 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
360 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
363 Use of these low level functions has been informally discouraged for a long
364 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
365 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
369 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
370 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
371 This means that applications don't have to look at the curve NID and
372 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
373 However, they still can, that EVP_PKEY_set_alias_type() call acts as
374 a no-op when the EVP_PKEY is already of the given type.
376 Parameter and key generation is also reworked to make it possible
377 to generate EVP_PKEY_SM2 parameters and keys without having to go
378 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
379 However, code that does the latter will still work as before.
383 * Deprecated low level ECDH and ECDSA functions. These include:
385 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
386 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
389 Use of these low level functions has been informally discouraged for a long
390 time. Instead applications should use the EVP_PKEY_derive(3),
391 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
395 * Deprecated the EC_KEY_METHOD functions. These include:
397 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
398 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
399 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
400 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
401 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
402 EC_KEY_METHOD_get_verify.
404 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
408 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
409 and EVP_PKEY_decrypt() instead.
410 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
411 and EVP_PKEY_encrypt() instead.
415 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
416 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
417 a new formulation to include all the things it can be used for,
418 as well as words of caution.
422 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
423 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
427 * All of the low level HMAC functions have been deprecated including:
429 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
430 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
433 Use of these low level functions has been informally discouraged for a long
434 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
435 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
436 and L<EVP_MAC_final(3)>.
440 * Over two thousand fixes were made to the documentation, including:
441 - Common options (such as -rand/-writerand, TLS version control, etc)
442 were refactored and point to newly-enhanced descriptions in openssl.pod.
443 - Added style conformance for all options (with help from Richard Levitte),
444 documented all reported missing options, added a CI build to check
445 that all options are documented and that no unimplemented options
447 - Documented some internals, such as all use of environment variables.
448 - Addressed all internal broken L<> references.
452 * All of the low level CMAC functions have been deprecated including:
454 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
455 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
457 Use of these low level functions has been informally discouraged for a long
458 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
459 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
460 and L<EVP_MAC_final(3)>.
464 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
465 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
468 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
469 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
470 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
471 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
472 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
473 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
474 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
475 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
476 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
477 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
479 Use of these low level functions has been informally discouraged
480 for a long time. Applications should use the EVP_DigestInit_ex(3),
481 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
485 * Corrected the documentation of the return values from the `EVP_DigestSign*`
486 set of functions. The documentation mentioned negative values for some
487 errors, but this was never the case, so the mention of negative values
490 Code that followed the documentation and thereby check with something
491 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
495 * All of the low level cipher functions have been deprecated including:
497 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
498 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
499 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
500 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
501 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
502 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
503 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
504 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
505 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
506 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
507 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
508 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
509 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
510 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
511 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
512 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
513 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
514 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
515 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
516 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
517 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
518 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
519 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
520 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
521 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
522 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
523 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
524 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
525 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
527 Use of these low level functions has been informally discouraged for
528 a long time. Applications should use the high level EVP APIs, e.g.
529 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
530 equivalently named decrypt functions instead.
532 *Matt Caswell and Paul Dale*
534 * Removed include/openssl/opensslconf.h.in and replaced it with
535 include/openssl/configuration.h.in, which differs in not including
536 <openssl/macros.h>. A short header include/openssl/opensslconf.h
537 was added to include both.
539 This allows internal hacks where one might need to modify the set
540 of configured macros, for example this if deprecated symbols are
541 still supposed to be available internally:
543 #include <openssl/configuration.h>
545 #undef OPENSSL_NO_DEPRECATED
546 #define OPENSSL_SUPPRESS_DEPRECATED
548 #include <openssl/macros.h>
550 This should not be used by applications that use the exported
551 symbols, as that will lead to linking errors.
555 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
556 used in exponentiation with 512-bit moduli. No EC algorithms are
557 affected. Analysis suggests that attacks against 2-prime RSA1024,
558 3-prime RSA1536, and DSA1024 as a result of this defect would be very
559 difficult to perform and are not believed likely. Attacks against DH512
560 are considered just feasible. However, for an attack the target would
561 have to re-use the DH512 private key, which is not recommended anyway.
562 Also applications directly using the low level API BN_mod_exp may be
563 affected if they use BN_FLG_CONSTTIME.
568 * Most memory-debug features have been deprecated, and the functionality
569 replaced with no-ops.
573 * Added documentation for the STACK API. OpenSSL only defines the STACK
574 functions where they are used.
578 * Introduced a new method type and API, OSSL_SERIALIZER, to
579 represent generic serializers. An implementation is expected to
580 be able to serialize an object associated with a given name (such
581 as an algorithm name for an asymmetric key) into forms given by
582 implementation properties.
584 Serializers are primarily used from inside libcrypto, through
585 calls to functions like EVP_PKEY_print_private(),
586 PEM_write_bio_PrivateKey() and similar.
588 Serializers are specified in such a way that they can be made to
589 directly handle the provider side portion of an object, if this
590 provider side part comes from the same provider as the serializer
591 itself, but can also be made to handle objects in parametrized
592 form (as an OSSL_PARAM array of data). This allows a provider to
593 offer generic serializers as a service for any other provider.
597 * Added a .pragma directive to the syntax of configuration files, to
598 allow varying behavior in a supported and predictable manner.
599 Currently added pragma:
603 This allows dollar signs to be a keyword character unless it's
604 followed by a opening brace or parenthesis. This is useful for
605 platforms where dollar signs are commonly used in names, such as
606 volume names and system directory names on VMS.
610 * Added functionality to create an EVP_PKEY from user data. This
611 is effectively the same as creating a RSA, DH or DSA object and
612 then assigning them to an EVP_PKEY, but directly using algorithm
613 agnostic EVP functions. A benefit is that this should be future
614 proof for public key algorithms to come.
618 * Change the interpretation of the '--api' configuration option to
619 mean that this is a desired API compatibility level with no
620 further meaning. The previous interpretation, that this would
621 also mean to remove all deprecated symbols up to and including
622 the given version, no requires that 'no-deprecated' is also used
623 in the configuration.
625 When building applications, the desired API compatibility level
626 can be set with the OPENSSL_API_COMPAT macro like before. For
627 API compatibility version below 3.0, the old style numerical
628 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
629 For version 3.0 and on, the value is expected to be the decimal
630 value calculated from the major and minor version like this:
632 MAJOR * 10000 + MINOR * 100
636 -DOPENSSL_API_COMPAT=30000 For 3.0
637 -DOPENSSL_API_COMPAT=30200 For 3.2
639 To hide declarations that are deprecated up to and including the
640 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
641 given when building the application as well.
645 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
646 access to certificate and CRL stores via URIs and OSSL_STORE
649 This adds the following functions:
651 - X509_LOOKUP_store()
652 - X509_STORE_load_file()
653 - X509_STORE_load_path()
654 - X509_STORE_load_store()
655 - SSL_add_store_cert_subjects_to_stack()
656 - SSL_CTX_set_default_verify_store()
657 - SSL_CTX_load_verify_file()
658 - SSL_CTX_load_verify_dir()
659 - SSL_CTX_load_verify_store()
663 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
664 The presence of this system service is determined at run-time.
668 * Added functionality to create an EVP_PKEY context based on data
669 for methods from providers. This takes an algorithm name and a
670 property query string and simply stores them, with the intent
671 that any operation that uses this context will use those strings
672 to fetch the needed methods implicitly, thereby making the port
673 of application written for pre-3.0 OpenSSL easier.
677 * The undocumented function NCONF_WIN32() has been deprecated; for
678 conversion details see the HISTORY section of doc/man5/config.pod
682 * Introduced the new functions EVP_DigestSignInit_ex() and
683 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
684 EVP_DigestVerifyUpdate() have been converted to functions. See the man
685 pages for further details.
689 * Over two thousand fixes were made to the documentation, including:
690 adding missing command flags, better style conformance, documentation
693 *Rich Salz, Richard Levitte*
695 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
696 X25519, X448, Ed25519 and Ed448.
700 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
705 * Deprecated the public definition of ERR_STATE as well as the function
706 ERR_get_state(). This is done in preparation of making ERR_STATE an
711 * Added ERR functionality to give callers access to the stored function
712 names that have replaced the older function code based functions.
714 New functions are ERR_get_error_func(), ERR_peek_error_func(),
715 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
716 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
717 and ERR_peek_last_error_all().
719 These functions have become deprecated: ERR_get_error_line_data(),
720 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
721 ERR_func_error_string().
725 * Extended testing to be verbose for failing tests only. The make variables
726 VERBOSE_FAILURE or VF can be used to enable this:
728 $ make VF=1 test # Unix
729 $ mms /macro=(VF=1) test ! OpenVMS
730 $ nmake VF=1 test # Windows
734 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
735 used even when parsing explicit parameters, when loading a serialized key
736 or calling `EC_GROUP_new_from_ecpkparameters()`/
737 `EC_GROUP_new_from_ecparameters()`.
738 This prevents bypass of security hardening and performance gains,
739 especially for curves with specialized EC_METHODs.
740 By default, if a key encoded with explicit parameters is loaded and later
741 serialized, the output is still encoded with explicit parameters, even if
742 internally a "named" EC_GROUP is used for computation.
746 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
747 this change, EC_GROUP_set_generator would accept order and/or cofactor as
748 NULL. After this change, only the cofactor parameter can be NULL. It also
749 does some minimal sanity checks on the passed order.
754 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
755 An attack is simple, if the first CMS_recipientInfo is valid but the
756 second CMS_recipientInfo is chosen ciphertext. If the second
757 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
758 encryption key will be replaced by garbage, and the message cannot be
759 decoded, but if the RSA decryption fails, the correct encryption key is
760 used and the recipient will not notice the attack.
761 As a work around for this potential attack the length of the decrypted
762 key must be equal to the cipher default key length, in case the
763 certifiate is not given and all recipientInfo are tried out.
764 The old behaviour can be re-enabled in the CMS code by setting the
765 CMS_DEBUG_DECRYPT flag.
769 * Early start up entropy quality from the DEVRANDOM seed source has been
770 improved for older Linux systems. The RAND subsystem will wait for
771 /dev/random to be producing output before seeding from /dev/urandom.
772 The seeded state is stored for future library initialisations using
773 a system global shared memory segment. The shared memory identifier
774 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
775 the desired value. The default identifier is 114.
779 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
780 when primes for RSA keys are computed.
781 Since we previously always generated primes == 2 (mod 3) for RSA keys,
782 the 2-prime and 3-prime RSA modules were easy to distinguish, since
783 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
784 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
785 This avoids possible fingerprinting of newly generated RSA modules.
789 * Correct the extended master secret constant on EBCDIC systems. Without this
790 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
791 negotiate EMS will fail. Unfortunately this also means that TLS connections
792 between EBCDIC systems with this fix, and EBCDIC systems without this
793 fix will fail if they negotiate EMS.
797 * Changed the library initialisation so that the config file is now loaded
798 by default. This was already the case for libssl. It now occurs for both
799 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
800 OPENSSL_init_crypto() to suppress automatic loading of a config file.
804 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
805 where the former acts as a replacement for ERR_put_error(), and the
806 latter replaces the combination ERR_put_error()+ERR_add_error_data().
807 ERR_raise_data() adds more flexibility by taking a format string and
808 an arbitrary number of arguments following it, to be processed with
813 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
814 to check if a named provider is loaded and available. When called, it
815 will also activate all fallback providers if such are still present.
819 * Enforce a minimum DH modulus size of 512 bits.
823 * Changed DH parameters to generate the order q subgroup instead of 2q.
824 Previously generated DH parameters are still accepted by DH_check
825 but DH_generate_key works around that by clearing bit 0 of the
826 private key for those. This avoids leaking bit 0 of the private key.
830 * Significantly reduce secure memory usage by the randomness pools.
834 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
839 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
840 algorithms. An implementation of a key exchange algorithm can be obtained
841 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
842 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
843 the older EVP_PKEY_derive_init() function. See the man pages for the new
844 functions for further details.
848 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
852 * Removed the function names from error messages and deprecated the
855 * Removed NextStep support and the macro OPENSSL_UNISTD
859 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
860 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
861 Also removed "export var as function" capability; we do not export
862 variables, only functions.
866 * RC5_32_set_key has been changed to return an int type, with 0 indicating
867 an error and 1 indicating success. In previous versions of OpenSSL this
868 was a void type. If a key was set longer than the maximum possible this
873 * Support SM2 signing and verification schemes with X509 certificate.
877 * Use SHA256 as the default digest for TS query in the ts app.
881 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
882 This checks that the salt length is at least 128 bits, the derived key
883 length is at least 112 bits, and that the iteration count is at least 1000.
884 For backwards compatibility these checks are disabled by default in the
885 default provider, but are enabled by default in the fips provider.
886 To enable or disable these checks use the control
887 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
891 * Default cipher lists/suites are now available via a function, the
892 #defines are deprecated.
896 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
897 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
898 for Windows Store apps easier. Also, the "no-uplink" option has been added.
902 * Join the directories crypto/x509 and crypto/x509v3
906 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
907 This changes the size when using the genpkey app when no size is given. It
908 fixes an omission in earlier changes that changed all RSA, DSA and DH
909 generation apps to use 2048 bits by default.
913 * Added command 'openssl kdf' that uses the EVP_KDF API.
917 * Added command 'openssl mac' that uses the EVP_MAC API.
921 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
922 as default directories. Also added the command 'openssl info'
923 for scripting purposes.
927 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
928 deprecated. These undocumented functions were never integrated into the EVP
929 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
930 Bi-directional IGE mode. These modes were never formally standardised and
931 usage of these functions is believed to be very small. In particular
932 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
933 is ever used. The security implications are believed to be minimal, but
934 this issue was never fixed for backwards compatibility reasons. New code
935 should not use these modes.
939 * Add prediction resistance to the DRBG reseeding process.
943 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
944 mandated by IEEE Std 1619-2018.
948 * Added newline escaping functionality to a filename when using openssl dgst.
949 This output format is to replicate the output format found in the `*sum`
950 checksum programs. This aims to preserve backward compatibility.
952 *Matt Eaton, Richard Levitte, and Paul Dale*
954 * Removed the heartbeat message in DTLS feature, as it has very
955 little usage and doesn't seem to fulfill a valuable purpose.
956 The configuration option is now deprecated.
960 * Changed the output of 'openssl {digestname} < file' to display the
961 digest name in its output.
965 * Added a new generic trace API which provides support for enabling
966 instrumentation through trace output. This feature is mainly intended
967 as an aid for developers and is disabled by default. To utilize it,
968 OpenSSL needs to be configured with the `enable-trace` option.
970 If the tracing API is enabled, the application can activate trace output
971 by registering BIOs as trace channels for a number of tracing and debugging
974 The 'openssl' application has been expanded to enable any of the types
975 available via environment variables defined by the user, and serves as
976 one possible example on how to use this functionality.
978 *Richard Levitte & Matthias St. Pierre*
980 * Added build tests for C++. These are generated files that only do one
981 thing, to include one public OpenSSL head file each. This tests that
982 the public header files can be usefully included in a C++ application.
984 This test isn't enabled by default. It can be enabled with the option
985 'enable-buildtest-c++'.
989 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
993 * Add KMAC to EVP_MAC.
997 * Added property based algorithm implementation selection framework to
1002 * Added SCA hardening for modular field inversion in EC_GROUP through
1003 a new dedicated field_inv() pointer in EC_METHOD.
1004 This also addresses a leakage affecting conversions from projective
1005 to affine coordinates.
1007 *Billy Bob Brumley, Nicola Tuveri*
1009 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1010 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1011 those algorithms that were already supported through the EVP_PKEY API
1012 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1013 and scrypt are now wrappers that call EVP_KDF.
1017 * Build devcrypto engine as a dynamic engine.
1019 *Eneas U de Queiroz*
1021 * Add keyed BLAKE2 to EVP_MAC.
1025 * Fix a bug in the computation of the endpoint-pair shared secret used
1026 by DTLS over SCTP. This breaks interoperability with older versions
1027 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1028 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1029 interoperability with such broken implementations. However, enabling
1030 this switch breaks interoperability with correct implementations.
1032 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1033 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1037 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1041 * Change the license to the Apache License v2.0.
1045 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1047 - Major releases (indicated by incrementing the MAJOR release number)
1048 may introduce incompatible API/ABI changes.
1049 - Minor releases (indicated by incrementing the MINOR release number)
1050 may introduce new features but retain API/ABI compatibility.
1051 - Patch releases (indicated by incrementing the PATCH number)
1052 are intended for bug fixes and other improvements of existing
1053 features only (like improving performance or adding documentation)
1054 and retain API/ABI compatibility.
1058 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1062 * Remove the 'dist' target and add a tarball building script. The
1063 'dist' target has fallen out of use, and it shouldn't be
1064 necessary to configure just to create a source distribution.
1068 * Recreate the OS390-Unix config target. It no longer relies on a
1069 special script like it did for OpenSSL pre-1.1.0.
1073 * Instead of having the source directories listed in Configure, add
1074 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1079 * Add GMAC to EVP_MAC.
1083 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1087 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1088 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1089 to facilitate the continued use of MACs through raw private keys in
1090 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1094 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1095 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1099 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1100 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1101 are retained for backwards compatibility.
1105 * AES-XTS mode now enforces that its two keys are different to mitigate
1106 the attacked described in "Efficient Instantiations of Tweakable
1107 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1108 Details of this attack can be obtained from:
1109 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1113 * Rename the object files, i.e. give them other names than in previous
1114 versions. Their names now include the name of the final product, as
1115 well as its type mnemonic (bin, lib, shlib).
1119 * Added new option for 'openssl list', '-objects', which will display the
1120 list of built in objects, i.e. OIDs with names.
1124 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1125 improves application performance by removing data copies and providing
1126 applications with zero-copy system calls such as sendfile and splice.
1130 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1131 option is set, openssl cleanses (zeroize) plaintext bytes from
1132 internal buffers after delivering them to the application. Note,
1133 the application is still responsible for cleansing other copies
1134 (e.g.: data received by SSL_read(3)).
1141 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1143 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1145 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1146 while reading in libssl then we would report an error back to the
1147 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1148 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1149 therefore give a hint as to what went wrong.
1153 * Check that ed25519 and ed448 are allowed by the security level. Previously
1154 signature algorithms not using an MD were not being checked that they were
1155 allowed by the security level.
1159 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1160 was not quite right. The behaviour was not consistent between resumption
1161 and normal handshakes, and also not quite consistent with historical
1162 behaviour. The behaviour in various scenarios has been clarified and
1163 it has been updated to make it match historical behaviour as closely as
1168 * *[VMS only]* The header files that the VMS compilers include automatically,
1169 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1170 that the C++ compiler doesn't understand. This is a shortcoming in the
1171 compiler, but can be worked around with `__cplusplus` guards.
1173 C++ applications that use OpenSSL libraries must be compiled using the
1174 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1175 functions. Otherwise, only functions with symbols of less than 31
1176 characters can be used, as the linker will not be able to successfully
1177 resolve symbols with longer names.
1181 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1182 The presence of this system service is determined at run-time.
1186 * Added newline escaping functionality to a filename when using openssl dgst.
1187 This output format is to replicate the output format found in the `*sum`
1188 checksum programs. This aims to preserve backward compatibility.
1190 *Matt Eaton, Richard Levitte, and Paul Dale*
1192 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1197 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1199 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1200 number generator (RNG). This was intended to include protection in the
1201 event of a fork() system call in order to ensure that the parent and child
1202 processes did not share the same RNG state. However this protection was not
1203 being used in the default case.
1205 A partial mitigation for this issue is that the output from a high
1206 precision timer is mixed into the RNG state so the likelihood of a parent
1207 and child process sharing state is significantly reduced.
1209 If an application already calls OPENSSL_init_crypto() explicitly using
1210 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1213 *Matthias St. Pierre*
1215 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1216 used even when parsing explicit parameters, when loading a serialized key
1217 or calling `EC_GROUP_new_from_ecpkparameters()`/
1218 `EC_GROUP_new_from_ecparameters()`.
1219 This prevents bypass of security hardening and performance gains,
1220 especially for curves with specialized EC_METHODs.
1221 By default, if a key encoded with explicit parameters is loaded and later
1222 serialized, the output is still encoded with explicit parameters, even if
1223 internally a "named" EC_GROUP is used for computation.
1227 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1228 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1229 NULL. After this change, only the cofactor parameter can be NULL. It also
1230 does some minimal sanity checks on the passed order.
1235 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1236 An attack is simple, if the first CMS_recipientInfo is valid but the
1237 second CMS_recipientInfo is chosen ciphertext. If the second
1238 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1239 encryption key will be replaced by garbage, and the message cannot be
1240 decoded, but if the RSA decryption fails, the correct encryption key is
1241 used and the recipient will not notice the attack.
1242 As a work around for this potential attack the length of the decrypted
1243 key must be equal to the cipher default key length, in case the
1244 certifiate is not given and all recipientInfo are tried out.
1245 The old behaviour can be re-enabled in the CMS code by setting the
1246 CMS_DEBUG_DECRYPT flag.
1251 * Early start up entropy quality from the DEVRANDOM seed source has been
1252 improved for older Linux systems. The RAND subsystem will wait for
1253 /dev/random to be producing output before seeding from /dev/urandom.
1254 The seeded state is stored for future library initialisations using
1255 a system global shared memory segment. The shared memory identifier
1256 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1257 the desired value. The default identifier is 114.
1261 * Correct the extended master secret constant on EBCDIC systems. Without this
1262 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1263 negotiate EMS will fail. Unfortunately this also means that TLS connections
1264 between EBCDIC systems with this fix, and EBCDIC systems without this
1265 fix will fail if they negotiate EMS.
1269 * Use Windows installation paths in the mingw builds
1271 Mingw isn't a POSIX environment per se, which means that Windows
1272 paths should be used for installation.
1277 * Changed DH_check to accept parameters with order q and 2q subgroups.
1278 With order 2q subgroups the bit 0 of the private key is not secret
1279 but DH_generate_key works around that by clearing bit 0 of the
1280 private key for those. This avoids leaking bit 0 of the private key.
1284 * Significantly reduce secure memory usage by the randomness pools.
1288 * Revert the DEVRANDOM_WAIT feature for Linux systems
1290 The DEVRANDOM_WAIT feature added a select() call to wait for the
1291 /dev/random device to become readable before reading from the
1292 /dev/urandom device.
1294 It turned out that this change had negative side effects on
1295 performance which were not acceptable. After some discussion it
1296 was decided to revert this feature and leave it up to the OS
1297 resp. the platform maintainer to ensure a proper initialization
1298 during early boot time.
1300 *Matthias St. Pierre*
1302 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1304 * Add build tests for C++. These are generated files that only do one
1305 thing, to include one public OpenSSL head file each. This tests that
1306 the public header files can be usefully included in a C++ application.
1308 This test isn't enabled by default. It can be enabled with the option
1309 'enable-buildtest-c++'.
1313 * Enable SHA3 pre-hashing for ECDSA and DSA.
1317 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1318 This changes the size when using the genpkey app when no size is given. It
1319 fixes an omission in earlier changes that changed all RSA, DSA and DH
1320 generation apps to use 2048 bits by default.
1324 * Reorganize the manual pages to consistently have RETURN VALUES,
1325 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1326 util/fix-doc-nits accordingly.
1328 *Paul Yang, Joshua Lock*
1330 * Add the missing accessor EVP_PKEY_get0_engine()
1334 * Have apps like 's_client' and 's_server' output the signature scheme
1335 along with other cipher suite parameters when debugging.
1339 * Make OPENSSL_config() error agnostic again.
1343 * Do the error handling in RSA decryption constant time.
1347 * Prevent over long nonces in ChaCha20-Poly1305.
1349 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1350 for every encryption operation. RFC 7539 specifies that the nonce value
1351 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1352 and front pads the nonce with 0 bytes if it is less than 12
1353 bytes. However it also incorrectly allows a nonce to be set of up to 16
1354 bytes. In this case only the last 12 bytes are significant and any
1355 additional leading bytes are ignored.
1357 It is a requirement of using this cipher that nonce values are
1358 unique. Messages encrypted using a reused nonce value are susceptible to
1359 serious confidentiality and integrity attacks. If an application changes
1360 the default nonce length to be longer than 12 bytes and then makes a
1361 change to the leading bytes of the nonce expecting the new value to be a
1362 new unique nonce then such an application could inadvertently encrypt
1363 messages with a reused nonce.
1365 Additionally the ignored bytes in a long nonce are not covered by the
1366 integrity guarantee of this cipher. Any application that relies on the
1367 integrity of these ignored leading bytes of a long nonce may be further
1368 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1369 is safe because no such use sets such a long nonce value. However user
1370 applications that use this cipher directly and set a non-default nonce
1371 length to be longer than 12 bytes may be vulnerable.
1373 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1379 * Add DEVRANDOM_WAIT feature for Linux systems
1381 On older Linux systems where the getrandom() system call is not available,
1382 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1383 Contrary to getrandom(), the /dev/urandom device will not block during
1384 early boot when the kernel CSPRNG has not been seeded yet.
1386 To mitigate this known weakness, use select() to wait for /dev/random to
1387 become readable before reading from /dev/urandom.
1389 * Ensure that SM2 only uses SM3 as digest algorithm
1393 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1395 * Change the info callback signals for the start and end of a post-handshake
1396 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1397 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1398 confused by this and assume that a TLSv1.2 renegotiation has started. This
1399 can break KeyUpdate handling. Instead we no longer signal the start and end
1400 of a post handshake message exchange (although the messages themselves are
1401 still signalled). This could break some applications that were expecting
1402 the old signals. However without this KeyUpdate is not usable for many
1407 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1409 * Timing vulnerability in DSA signature generation
1411 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1412 timing side channel attack. An attacker could use variations in the signing
1413 algorithm to recover the private key.
1415 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1420 * Timing vulnerability in ECDSA signature generation
1422 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1423 timing side channel attack. An attacker could use variations in the signing
1424 algorithm to recover the private key.
1426 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1431 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1432 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1433 of two gigabytes and the error handling improved.
1435 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1436 categorized as a normal bug, not a security issue, because the DRBG reseeds
1437 automatically and is fully functional even without additional randomness
1438 provided by the application.
1440 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1442 * Add a new ClientHello callback. Provides a callback interface that gives
1443 the application the ability to adjust the nascent SSL object at the
1444 earliest stage of ClientHello processing, immediately after extensions have
1445 been collected but before they have been processed. In particular, this
1446 callback can adjust the supported TLS versions in response to the contents
1451 * Add SM2 base algorithm support.
1455 * s390x assembly pack: add (improved) hardware-support for the following
1456 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1457 aes-cfb/cfb8, aes-ecb.
1461 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1462 parameter is no longer accepted, as it leads to a corrupt table. NULL
1463 pem_str is reserved for alias entries only.
1467 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1468 step for prime curves. The new implementation is based on formulae from
1469 differential addition-and-doubling in homogeneous projective coordinates
1470 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1471 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1472 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1473 to work in projective coordinates.
1475 *Billy Bob Brumley, Nicola Tuveri*
1477 * Change generating and checking of primes so that the error rate of not
1478 being prime depends on the intended use based on the size of the input.
1479 For larger primes this will result in more rounds of Miller-Rabin.
1480 The maximal error rate for primes with more than 1080 bits is lowered
1483 *Kurt Roeckx, Annie Yousar*
1485 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1489 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1490 moving between systems, and to avoid confusion when a Windows build is
1491 done with mingw vs with MSVC. For POSIX installs, there's still a
1492 symlink or copy named 'tsget' to avoid that confusion as well.
1496 * Revert blinding in ECDSA sign and instead make problematic addition
1497 length-invariant. Switch even to fixed-length Montgomery multiplication.
1501 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1502 step for binary curves. The new implementation is based on formulae from
1503 differential addition-and-doubling in mixed Lopez-Dahab projective
1504 coordinates, modified to independently blind the operands.
1506 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1508 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1509 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1510 EC_METHODs to implement their own specialized "ladder step", to take
1511 advantage of more favorable coordinate systems or more efficient
1512 differential addition-and-doubling algorithms.
1514 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1516 * Modified the random device based seed sources to keep the relevant
1517 file descriptors open rather than reopening them on each access.
1518 This allows such sources to operate in a chroot() jail without
1519 the associated device nodes being available. This behaviour can be
1520 controlled using RAND_keep_random_devices_open().
1524 * Numerous side-channel attack mitigations have been applied. This may have
1525 performance impacts for some algorithms for the benefit of improved
1526 security. Specific changes are noted in this change log by their respective
1531 * AIX shared library support overhaul. Switch to AIX "natural" way of
1532 handling shared libraries, which means collecting shared objects of
1533 different versions and bitnesses in one common archive. This allows to
1534 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1535 doesn't affect the way 3rd party applications are linked, only how
1536 multi-version installation is managed.
1540 * Make ec_group_do_inverse_ord() more robust and available to other
1541 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1542 mitigations are applied to the fallback BN_mod_inverse().
1543 When using this function rather than BN_mod_inverse() directly, new
1544 EC cryptosystem implementations are then safer-by-default.
1548 * Add coordinate blinding for EC_POINT and implement projective
1549 coordinate blinding for generic prime curves as a countermeasure to
1550 chosen point SCA attacks.
1552 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1554 * Add blinding to ECDSA and DSA signatures to protect against side channel
1555 attacks discovered by Keegan Ryan (NCC Group).
1559 * Enforce checking in the pkeyutl command line app to ensure that the input
1560 length does not exceed the maximum supported digest length when performing
1561 a sign, verify or verifyrecover operation.
1565 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1566 I/O in combination with something like select() or poll() will hang. This
1567 can be turned off again using SSL_CTX_clear_mode().
1568 Many applications do not properly handle non-application data records, and
1569 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1570 around the problems in those applications, but can also break some.
1571 It's recommended to read the manpages about SSL_read(), SSL_write(),
1572 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1573 SSL_CTX_set_read_ahead() again.
1577 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1578 now allow empty (zero character) pass phrases.
1582 * Apply blinding to binary field modular inversion and remove patent
1583 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1587 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1588 binary and prime elliptic curves.
1592 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1593 constant time fixed point multiplication.
1597 * Revise elliptic curve scalar multiplication with timing attack
1598 defenses: ec_wNAF_mul redirects to a constant time implementation
1599 when computing fixed point and variable point multiplication (which
1600 in OpenSSL are mostly used with secret scalars in keygen, sign,
1601 ECDH derive operations).
1602 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1605 * Updated CONTRIBUTING
1609 * Updated DRBG / RAND to request nonce and additional low entropy
1610 randomness from the system.
1612 *Matthias St. Pierre*
1614 * Updated 'openssl rehash' to use OpenSSL consistent default.
1618 * Moved the load of the ssl_conf module to libcrypto, which helps
1619 loading engines that libssl uses before libssl is initialised.
1623 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1627 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1629 *Ingo Schwarze, Rich Salz*
1631 * Added output of accepting IP address and port for 'openssl s_server'
1635 * Added a new API for TLSv1.3 ciphersuites:
1636 SSL_CTX_set_ciphersuites()
1637 SSL_set_ciphersuites()
1641 * Memory allocation failures consistently add an error to the error
1646 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1647 in libcrypto when run as setuid/setgid.
1651 * Load any config file by default when libssl is used.
1655 * Added new public header file <openssl/rand_drbg.h> and documentation
1656 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1658 *Matthias St. Pierre*
1660 * QNX support removed (cannot find contributors to get their approval
1661 for the license change).
1665 * TLSv1.3 replay protection for early data has been implemented. See the
1666 SSL_read_early_data() man page for further details.
1670 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1671 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1672 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1673 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1674 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1675 configuration has been separated out. See the ciphers man page or the
1676 SSL_CTX_set_ciphersuites() man page for more information.
1680 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1681 in responder mode now supports the new "-multi" option, which
1682 spawns the specified number of child processes to handle OCSP
1683 requests. The "-timeout" option now also limits the OCSP
1684 responder's patience to wait to receive the full client request
1685 on a newly accepted connection. Child processes are respawned
1686 as needed, and the CA index file is automatically reloaded
1687 when changed. This makes it possible to run the "ocsp" responder
1688 as a long-running service, making the OpenSSL CA somewhat more
1689 feature-complete. In this mode, most diagnostic messages logged
1690 after entering the event loop are logged via syslog(3) rather than
1695 * Added support for X448 and Ed448. Heavily based on original work by
1700 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1701 objects loaded. This adds the functions OSSL_STORE_expect() and
1702 OSSL_STORE_find() as well as needed tools to construct searches and
1703 get the search data out of them.
1707 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1708 version of OpenSSL should review their configuration settings to ensure
1709 that they are still appropriate for TLSv1.3. For further information see:
1710 <https://wiki.openssl.org/index.php/TLS1.3>
1714 * Grand redesign of the OpenSSL random generator
1716 The default RAND method now utilizes an AES-CTR DRBG according to
1717 NIST standard SP 800-90Ar1. The new random generator is essentially
1718 a port of the default random generator from the OpenSSL FIPS 2.0
1719 object module. It is a hybrid deterministic random bit generator
1720 using an AES-CTR bit stream and which seeds and reseeds itself
1721 automatically using trusted system entropy sources.
1723 Some of its new features are:
1724 - Support for multiple DRBG instances with seed chaining.
1725 - The default RAND method makes use of a DRBG.
1726 - There is a public and private DRBG instance.
1727 - The DRBG instances are fork-safe.
1728 - Keep all global DRBG instances on the secure heap if it is enabled.
1729 - The public and private DRBG instance are per thread for lock free
1732 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1734 * Changed Configure so it only says what it does and doesn't dump
1735 so much data. Instead, ./configdata.pm should be used as a script
1736 to display all sorts of configuration data.
1740 * Added processing of "make variables" to Configure.
1744 * Added SHA512/224 and SHA512/256 algorithm support.
1748 * The last traces of Netware support, first removed in 1.1.0, have
1753 * Get rid of Makefile.shared, and in the process, make the processing
1754 of certain files (rc.obj, or the .def/.map/.opt files produced from
1755 the ordinal files) more visible and hopefully easier to trace and
1756 debug (or make silent).
1760 * Make it possible to have environment variable assignments as
1761 arguments to config / Configure.
1765 * Add multi-prime RSA (RFC 8017) support.
1769 * Add SM3 implemented according to GB/T 32905-2016
1770 *Jack Lloyd <jack.lloyd@ribose.com>,*
1771 *Ronald Tse <ronald.tse@ribose.com>,*
1772 *Erick Borsboom <erick.borsboom@ribose.com>*
1774 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1775 as documented in RFC6066.
1776 Based on a patch from Tomasz Moń
1778 *Filipe Raimundo da Silva*
1780 * Add SM4 implemented according to GB/T 32907-2016.
1781 *Jack Lloyd <jack.lloyd@ribose.com>,*
1782 *Ronald Tse <ronald.tse@ribose.com>,*
1783 *Erick Borsboom <erick.borsboom@ribose.com>*
1785 * Reimplement -newreq-nodes and ERR_error_string_n; the
1786 original author does not agree with the license change.
1790 * Add ARIA AEAD TLS support.
1794 * Some macro definitions to support VS6 have been removed. Visual
1795 Studio 6 has not worked since 1.1.0
1799 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1800 without clearing the errors.
1804 * Add "atfork" functions. If building on a system that without
1805 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1806 requirements. The RAND facility now uses/requires this.
1814 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1815 not possible to disable entirely. However, it's still possible to
1816 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1819 To disable, configure with 'no-ui-console'. 'no-ui' is still
1820 possible to use as an alias. Check at compile time with the
1821 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1822 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1826 * Add a STORE module, which implements a uniform and URI based reader of
1827 stores that can contain keys, certificates, CRLs and numerous other
1828 objects. The main API is loosely based on a few stdio functions,
1829 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1830 OSSL_STORE_error and OSSL_STORE_close.
1831 The implementation uses backends called "loaders" to implement arbitrary
1832 URI schemes. There is one built in "loader" for the 'file' scheme.
1836 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1837 then adjusted to work on FreeBSD 8.4 as well.
1838 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1839 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1843 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1844 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1845 error code calls like this:
1847 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1849 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1850 that can be encoded in C. For the foreseeable future, this will only
1853 *Richard Levitte and Tim Hudson*
1855 * Removed BSD cryptodev engine.
1859 * Add a build target 'build_all_generated', to build all generated files
1860 and only that. This can be used to prepare everything that requires
1861 things like perl for a system that lacks perl and then move everything
1862 to that system and do the rest of the build there.
1866 * In the UI interface, make it possible to duplicate the user data. This
1867 can be used by engines that need to retain the data for a longer time
1868 than just the call where this user data is passed.
1872 * Ignore the '-named_curve auto' value for compatibility of applications
1875 *Tomas Mraz <tmraz@fedoraproject.org>*
1877 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1878 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1879 alerts across multiple records (some of which could be empty). In practice
1880 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1881 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1882 support this at all. Supporting it adds significant complexity to the
1883 record layer, and its removal is unlikely to cause interoperability
1888 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1889 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1890 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1895 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1896 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1898 *Richard Levitte, Andy Polyakov*
1900 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1905 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1906 platform rather than 'mingw'.
1910 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1911 success if they are asked to add an object which already exists
1912 in the store. This change cascades to other functions which load
1913 certificates and CRLs.
1917 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1918 facilitate stack unwinding even from assembly subroutines.
1922 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1923 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1927 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1928 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1929 which is the minimum version we support.
1933 * Certificate time validation (X509_cmp_time) enforces stricter
1934 compliance with RFC 5280. Fractional seconds and timezone offsets
1935 are no longer allowed.
1939 * Add support for ARIA
1943 * s_client will now send the Server Name Indication (SNI) extension by
1944 default unless the new "-noservername" option is used. The server name is
1945 based on the host provided to the "-connect" option unless overridden by
1946 using "-servername".
1950 * Add support for SipHash
1954 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1955 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1956 prevent issues where no progress is being made and the peer continually
1957 sends unrecognised record types, using up resources processing them.
1961 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1962 using the algorithm defined in
1963 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1967 * Heartbeat support has been removed; the ABI is changed for now.
1969 *Richard Levitte, Rich Salz*
1971 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1975 * The RSA "null" method, which was partially supported to avoid patent
1976 issues, has been replaced to always returns NULL.
1983 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1985 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1986 used even when parsing explicit parameters, when loading a serialized key
1987 or calling `EC_GROUP_new_from_ecpkparameters()`/
1988 `EC_GROUP_new_from_ecparameters()`.
1989 This prevents bypass of security hardening and performance gains,
1990 especially for curves with specialized EC_METHODs.
1991 By default, if a key encoded with explicit parameters is loaded and later
1992 serialized, the output is still encoded with explicit parameters, even if
1993 internally a "named" EC_GROUP is used for computation.
1997 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1998 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1999 NULL. After this change, only the cofactor parameter can be NULL. It also
2000 does some minimal sanity checks on the passed order.
2005 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2006 An attack is simple, if the first CMS_recipientInfo is valid but the
2007 second CMS_recipientInfo is chosen ciphertext. If the second
2008 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2009 encryption key will be replaced by garbage, and the message cannot be
2010 decoded, but if the RSA decryption fails, the correct encryption key is
2011 used and the recipient will not notice the attack.
2012 As a work around for this potential attack the length of the decrypted
2013 key must be equal to the cipher default key length, in case the
2014 certifiate is not given and all recipientInfo are tried out.
2015 The old behaviour can be re-enabled in the CMS code by setting the
2016 CMS_DEBUG_DECRYPT flag.
2021 * Use Windows installation paths in the mingw builds
2023 Mingw isn't a POSIX environment per se, which means that Windows
2024 paths should be used for installation.
2029 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2031 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2032 This changes the size when using the genpkey app when no size is given. It
2033 fixes an omission in earlier changes that changed all RSA, DSA and DH
2034 generation apps to use 2048 bits by default.
2038 * Prevent over long nonces in ChaCha20-Poly1305.
2040 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2041 for every encryption operation. RFC 7539 specifies that the nonce value
2042 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2043 and front pads the nonce with 0 bytes if it is less than 12
2044 bytes. However it also incorrectly allows a nonce to be set of up to 16
2045 bytes. In this case only the last 12 bytes are significant and any
2046 additional leading bytes are ignored.
2048 It is a requirement of using this cipher that nonce values are
2049 unique. Messages encrypted using a reused nonce value are susceptible to
2050 serious confidentiality and integrity attacks. If an application changes
2051 the default nonce length to be longer than 12 bytes and then makes a
2052 change to the leading bytes of the nonce expecting the new value to be a
2053 new unique nonce then such an application could inadvertently encrypt
2054 messages with a reused nonce.
2056 Additionally the ignored bytes in a long nonce are not covered by the
2057 integrity guarantee of this cipher. Any application that relies on the
2058 integrity of these ignored leading bytes of a long nonce may be further
2059 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2060 is safe because no such use sets such a long nonce value. However user
2061 applications that use this cipher directly and set a non-default nonce
2062 length to be longer than 12 bytes may be vulnerable.
2064 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2070 * Added SCA hardening for modular field inversion in EC_GROUP through
2071 a new dedicated field_inv() pointer in EC_METHOD.
2072 This also addresses a leakage affecting conversions from projective
2073 to affine coordinates.
2075 *Billy Bob Brumley, Nicola Tuveri*
2077 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2078 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2082 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2086 * Remove the 'dist' target and add a tarball building script. The
2087 'dist' target has fallen out of use, and it shouldn't be
2088 necessary to configure just to create a source distribution.
2092 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2094 * Timing vulnerability in DSA signature generation
2096 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2097 timing side channel attack. An attacker could use variations in the signing
2098 algorithm to recover the private key.
2100 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2105 * Timing vulnerability in ECDSA signature generation
2107 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2108 timing side channel attack. An attacker could use variations in the signing
2109 algorithm to recover the private key.
2111 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2116 * Add coordinate blinding for EC_POINT and implement projective
2117 coordinate blinding for generic prime curves as a countermeasure to
2118 chosen point SCA attacks.
2120 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2122 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2124 * Client DoS due to large DH parameter
2126 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2127 malicious server can send a very large prime value to the client. This will
2128 cause the client to spend an unreasonably long period of time generating a
2129 key for this prime resulting in a hang until the client has finished. This
2130 could be exploited in a Denial Of Service attack.
2132 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2137 * Cache timing vulnerability in RSA Key Generation
2139 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2140 a cache timing side channel attack. An attacker with sufficient access to
2141 mount cache timing attacks during the RSA key generation process could
2142 recover the private key.
2144 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2145 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2150 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2151 parameter is no longer accepted, as it leads to a corrupt table. NULL
2152 pem_str is reserved for alias entries only.
2156 * Revert blinding in ECDSA sign and instead make problematic addition
2157 length-invariant. Switch even to fixed-length Montgomery multiplication.
2161 * Change generating and checking of primes so that the error rate of not
2162 being prime depends on the intended use based on the size of the input.
2163 For larger primes this will result in more rounds of Miller-Rabin.
2164 The maximal error rate for primes with more than 1080 bits is lowered
2167 *Kurt Roeckx, Annie Yousar*
2169 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2173 * Add blinding to ECDSA and DSA signatures to protect against side channel
2174 attacks discovered by Keegan Ryan (NCC Group).
2178 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2179 now allow empty (zero character) pass phrases.
2183 * Certificate time validation (X509_cmp_time) enforces stricter
2184 compliance with RFC 5280. Fractional seconds and timezone offsets
2185 are no longer allowed.
2189 * Fixed a text canonicalisation bug in CMS
2191 Where a CMS detached signature is used with text content the text goes
2192 through a canonicalisation process first prior to signing or verifying a
2193 signature. This process strips trailing space at the end of lines, converts
2194 line terminators to CRLF and removes additional trailing line terminators
2195 at the end of a file. A bug in the canonicalisation process meant that
2196 some characters, such as form-feed, were incorrectly treated as whitespace
2197 and removed. This is contrary to the specification (RFC5485). This fix
2198 could mean that detached text data signed with an earlier version of
2199 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2200 signed with a fixed OpenSSL may fail to verify with an earlier version of
2201 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2202 and use the "-binary" flag (for the "cms" command line application) or set
2203 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2207 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2209 * Constructed ASN.1 types with a recursive definition could exceed the stack
2211 Constructed ASN.1 types with a recursive definition (such as can be found
2212 in PKCS7) could eventually exceed the stack given malicious input with
2213 excessive recursion. This could result in a Denial Of Service attack. There
2214 are no such structures used within SSL/TLS that come from untrusted sources
2215 so this is considered safe.
2217 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2223 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2225 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2226 effectively reduced to only comparing the least significant bit of each
2227 byte. This allows an attacker to forge messages that would be considered as
2228 authenticated in an amount of tries lower than that guaranteed by the
2229 security claims of the scheme. The module can only be compiled by the
2230 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2232 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2238 * Add a build target 'build_all_generated', to build all generated files
2239 and only that. This can be used to prepare everything that requires
2240 things like perl for a system that lacks perl and then move everything
2241 to that system and do the rest of the build there.
2245 * Backport SSL_OP_NO_RENGOTIATION
2247 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2248 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2249 changes this is no longer possible in 1.1.0. Therefore the new
2250 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2251 1.1.0 to provide equivalent functionality.
2253 Note that if an application built against 1.1.0h headers (or above) is run
2254 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2255 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2259 * Removed the OS390-Unix config target. It relied on a script that doesn't
2264 * rsaz_1024_mul_avx2 overflow bug on x86_64
2266 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2267 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2268 Analysis suggests that attacks against RSA and DSA as a result of this
2269 defect would be very difficult to perform and are not believed likely.
2270 Attacks against DH1024 are considered just feasible, because most of the
2271 work necessary to deduce information about a private key may be performed
2272 offline. The amount of resources required for such an attack would be
2273 significant. However, for an attack on TLS to be meaningful, the server
2274 would have to share the DH1024 private key among multiple clients, which is
2275 no longer an option since CVE-2016-0701.
2277 This only affects processors that support the AVX2 but not ADX extensions
2278 like Intel Haswell (4th generation).
2280 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2281 was originally found via the OSS-Fuzz project.
2286 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2288 * bn_sqrx8x_internal carry bug on x86_64
2290 There is a carry propagating bug in the x86_64 Montgomery squaring
2291 procedure. No EC algorithms are affected. Analysis suggests that attacks
2292 against RSA and DSA as a result of this defect would be very difficult to
2293 perform and are not believed likely. Attacks against DH are considered just
2294 feasible (although very difficult) because most of the work necessary to
2295 deduce information about a private key may be performed offline. The amount
2296 of resources required for such an attack would be very significant and
2297 likely only accessible to a limited number of attackers. An attacker would
2298 additionally need online access to an unpatched system using the target
2299 private key in a scenario with persistent DH parameters and a private
2300 key that is shared between multiple clients.
2302 This only affects processors that support the BMI1, BMI2 and ADX extensions
2303 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2305 This issue was reported to OpenSSL by the OSS-Fuzz project.
2310 * Malformed X.509 IPAddressFamily could cause OOB read
2312 If an X.509 certificate has a malformed IPAddressFamily extension,
2313 OpenSSL could do a one-byte buffer overread. The most likely result
2314 would be an erroneous display of the certificate in text format.
2316 This issue was reported to OpenSSL by the OSS-Fuzz project.
2321 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2323 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2324 platform rather than 'mingw'.
2328 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2329 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2330 which is the minimum version we support.
2334 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2336 * Encrypt-Then-Mac renegotiation crash
2338 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2339 negotiated where it was not in the original handshake (or vice-versa) then
2340 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2341 and servers are affected.
2343 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2348 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2350 * Truncated packet could crash via OOB read
2352 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2353 cipher is being used, then a truncated packet can cause that host to
2354 perform an out-of-bounds read, usually resulting in a crash.
2356 This issue was reported to OpenSSL by Robert Święcki of Google.
2361 * Bad (EC)DHE parameters cause a client crash
2363 If a malicious server supplies bad parameters for a DHE or ECDHE key
2364 exchange then this can result in the client attempting to dereference a
2365 NULL pointer leading to a client crash. This could be exploited in a Denial
2368 This issue was reported to OpenSSL by Guido Vranken.
2373 * BN_mod_exp may produce incorrect results on x86_64
2375 There is a carry propagating bug in the x86_64 Montgomery squaring
2376 procedure. No EC algorithms are affected. Analysis suggests that attacks
2377 against RSA and DSA as a result of this defect would be very difficult to
2378 perform and are not believed likely. Attacks against DH are considered just
2379 feasible (although very difficult) because most of the work necessary to
2380 deduce information about a private key may be performed offline. The amount
2381 of resources required for such an attack would be very significant and
2382 likely only accessible to a limited number of attackers. An attacker would
2383 additionally need online access to an unpatched system using the target
2384 private key in a scenario with persistent DH parameters and a private
2385 key that is shared between multiple clients. For example this can occur by
2386 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2387 similar to CVE-2015-3193 but must be treated as a separate problem.
2389 This issue was reported to OpenSSL by the OSS-Fuzz project.
2394 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2396 * ChaCha20/Poly1305 heap-buffer-overflow
2398 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2399 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2400 crash. This issue is not considered to be exploitable beyond a DoS.
2402 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2407 * CMS Null dereference
2409 Applications parsing invalid CMS structures can crash with a NULL pointer
2410 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2411 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2412 structure callback if an attempt is made to free certain invalid encodings.
2413 Only CHOICE structures using a callback which do not handle NULL value are
2416 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2421 * Montgomery multiplication may produce incorrect results
2423 There is a carry propagating bug in the Broadwell-specific Montgomery
2424 multiplication procedure that handles input lengths divisible by, but
2425 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2426 and DH private keys are impossible. This is because the subroutine in
2427 question is not used in operations with the private key itself and an input
2428 of the attacker's direct choice. Otherwise the bug can manifest itself as
2429 transient authentication and key negotiation failures or reproducible
2430 erroneous outcome of public-key operations with specially crafted input.
2431 Among EC algorithms only Brainpool P-512 curves are affected and one
2432 presumably can attack ECDH key negotiation. Impact was not analyzed in
2433 detail, because pre-requisites for attack are considered unlikely. Namely
2434 multiple clients have to choose the curve in question and the server has to
2435 share the private key among them, neither of which is default behaviour.
2436 Even then only clients that chose the curve will be affected.
2438 This issue was publicly reported as transient failures and was not
2439 initially recognized as a security issue. Thanks to Richard Morgan for
2440 providing reproducible case.
2445 * Removed automatic addition of RPATH in shared libraries and executables,
2446 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2450 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2452 * Fix Use After Free for large message sizes
2454 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2455 message larger than approx 16k is received then the underlying buffer to
2456 store the incoming message is reallocated and moved. Unfortunately a
2457 dangling pointer to the old location is left which results in an attempt to
2458 write to the previously freed location. This is likely to result in a
2459 crash, however it could potentially lead to execution of arbitrary code.
2461 This issue only affects OpenSSL 1.1.0a.
2463 This issue was reported to OpenSSL by Robert Święcki.
2468 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2470 * OCSP Status Request extension unbounded memory growth
2472 A malicious client can send an excessively large OCSP Status Request
2473 extension. If that client continually requests renegotiation, sending a
2474 large OCSP Status Request extension each time, then there will be unbounded
2475 memory growth on the server. This will eventually lead to a Denial Of
2476 Service attack through memory exhaustion. Servers with a default
2477 configuration are vulnerable even if they do not support OCSP. Builds using
2478 the "no-ocsp" build time option are not affected.
2480 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2485 * SSL_peek() hang on empty record
2487 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2488 sends an empty record. This could be exploited by a malicious peer in a
2489 Denial Of Service attack.
2491 This issue was reported to OpenSSL by Alex Gaynor.
2496 * Excessive allocation of memory in tls_get_message_header() and
2497 dtls1_preprocess_fragment()
2499 A (D)TLS message includes 3 bytes for its length in the header for the
2500 message. This would allow for messages up to 16Mb in length. Messages of
2501 this length are excessive and OpenSSL includes a check to ensure that a
2502 peer is sending reasonably sized messages in order to avoid too much memory
2503 being consumed to service a connection. A flaw in the logic of version
2504 1.1.0 means that memory for the message is allocated too early, prior to
2505 the excessive message length check. Due to way memory is allocated in
2506 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2507 to service a connection. This could lead to a Denial of Service through
2508 memory exhaustion. However, the excessive message length check still takes
2509 place, and this would cause the connection to immediately fail. Assuming
2510 that the application calls SSL_free() on the failed connection in a timely
2511 manner then the 21Mb of allocated memory will then be immediately freed
2512 again. Therefore the excessive memory allocation will be transitory in
2513 nature. This then means that there is only a security impact if:
2515 1) The application does not call SSL_free() in a timely manner in the event
2516 that the connection fails
2518 2) The application is working in a constrained environment where there is
2519 very little free memory
2521 3) The attacker initiates multiple connection attempts such that there are
2522 multiple connections in a state where memory has been allocated for the
2523 connection; SSL_free() has not yet been called; and there is insufficient
2524 memory to service the multiple requests.
2526 Except in the instance of (1) above any Denial Of Service is likely to be
2527 transitory because as soon as the connection fails the memory is
2528 subsequently freed again in the SSL_free() call. However there is an
2529 increased risk during this period of application crashes due to the lack of
2530 memory - which would then mean a more serious Denial of Service.
2532 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2533 (CVE-2016-6307 and CVE-2016-6308)
2537 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2538 had to be removed. Primary reason is that vendor assembler can't
2539 assemble our modules with -KPIC flag. As result it, assembly
2540 support, was not even available as option. But its lack means
2541 lack of side-channel resistant code, which is incompatible with
2542 security by todays standards. Fortunately gcc is readily available
2543 prepackaged option, which we firmly point at...
2547 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2549 * Windows command-line tool supports UTF-8 opt-in option for arguments
2550 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2551 (to any value) allows Windows user to access PKCS#12 file generated
2552 with Windows CryptoAPI and protected with non-ASCII password, as well
2553 as files generated under UTF-8 locale on Linux also protected with
2558 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2559 have been disabled by default and removed from DEFAULT, just like RC4.
2560 See the RC4 item below to re-enable both.
2564 * The method for finding the storage location for the Windows RAND seed file
2565 has changed. First we check %RANDFILE%. If that is not set then we check
2566 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2567 all else fails we fall back to C:\.
2571 * The EVP_EncryptUpdate() function has had its return type changed from void
2572 to int. A return of 0 indicates and error while a return of 1 indicates
2577 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2578 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2579 off the constant time implementation for RSA, DSA and DH have been made
2580 no-ops and deprecated.
2584 * Windows RAND implementation was simplified to only get entropy by
2585 calling CryptGenRandom(). Various other RAND-related tickets
2588 *Joseph Wylie Yandle, Rich Salz*
2590 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2591 and `OPENSSL_LH_`, respectively. The old names are available
2592 with API compatibility. They new names are now completely documented.
2596 * Unify TYPE_up_ref(obj) methods signature.
2597 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2598 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2599 int (instead of void) like all others TYPE_up_ref() methods.
2600 So now these methods also check the return value of CRYPTO_atomic_add(),
2601 and the validity of object reference counter.
2603 *fdasilvayy@gmail.com*
2605 * With Windows Visual Studio builds, the .pdb files are installed
2606 alongside the installed libraries and executables. For a static
2607 library installation, ossl_static.pdb is the associate compiler
2608 generated .pdb file to be used when linking programs.
2612 * Remove openssl.spec. Packaging files belong with the packagers.
2616 * Automatic Darwin/OSX configuration has had a refresh, it will now
2617 recognise x86_64 architectures automatically. You can still decide
2618 to build for a different bitness with the environment variable
2619 KERNEL_BITS (can be 32 or 64), for example:
2621 KERNEL_BITS=32 ./config
2625 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2626 256 bit AES and HMAC with SHA256.
2630 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2634 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2638 * To enable users to have their own config files and build file templates,
2639 Configure looks in the directory indicated by the environment variable
2640 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2641 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2642 name and is used as is.
2646 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2647 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2648 X509_CERT_FILE_CTX was removed.
2652 * "shared" builds are now the default. To create only static libraries use
2653 the "no-shared" Configure option.
2657 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2658 All of these option have not worked for some while and are fundamental
2663 * Make various cleanup routines no-ops and mark them as deprecated. Most
2664 global cleanup functions are no longer required because they are handled
2665 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2666 Explicitly de-initing can cause problems (e.g. where a library that uses
2667 OpenSSL de-inits, but an application is still using it). The affected
2668 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2669 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2670 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2671 COMP_zlib_cleanup().
2675 * --strict-warnings no longer enables runtime debugging options
2676 such as REF_DEBUG. Instead, debug options are automatically
2677 enabled with '--debug' builds.
2679 *Andy Polyakov, Emilia Käsper*
2681 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2682 have been moved out of the public header files. New functions for managing
2683 these have been added.
2687 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2688 objects have been moved out of the public header files. New
2689 functions for managing these have been added.
2693 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2694 have been moved out of the public header files. New functions for managing
2695 these have been added.
2699 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2700 moved out of the public header files. New functions for managing these
2705 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2709 * Removed the mk1mf build scripts.
2713 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2714 it is always safe to #include a header now.
2718 * Removed the aged BC-32 config and all its supporting scripts
2722 * Removed support for Ultrix, Netware, and OS/2.
2726 * Add support for HKDF.
2728 *Alessandro Ghedini*
2730 * Add support for blake2b and blake2s
2734 * Added support for "pipelining". Ciphers that have the
2735 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2736 encryptions/decryptions simultaneously. There are currently no built-in
2737 ciphers with this property but the expectation is that engines will be able
2738 to offer it to significantly improve throughput. Support has been extended
2739 into libssl so that multiple records for a single connection can be
2740 processed in one go (for >=TLS 1.1).
2744 * Added the AFALG engine. This is an async capable engine which is able to
2745 offload work to the Linux kernel. In this initial version it only supports
2746 AES128-CBC. The kernel must be version 4.1.0 or greater.
2750 * OpenSSL now uses a new threading API. It is no longer necessary to
2751 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2752 are two supported threading models: pthreads and windows threads. It is
2753 also possible to configure OpenSSL at compile time for "no-threads". The
2754 old threading API should no longer be used. The functions have been
2755 replaced with "no-op" compatibility macros.
2757 *Alessandro Ghedini, Matt Caswell*
2759 * Modify behavior of ALPN to invoke callback after SNI/servername
2760 callback, such that updates to the SSL_CTX affect ALPN.
2764 * Add SSL_CIPHER queries for authentication and key-exchange.
2768 * Changes to the DEFAULT cipherlist:
2769 - Prefer (EC)DHE handshakes over plain RSA.
2770 - Prefer AEAD ciphers over legacy ciphers.
2771 - Prefer ECDSA over RSA when both certificates are available.
2772 - Prefer TLSv1.2 ciphers/PRF.
2773 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2778 * Change the ECC default curve list to be this, in order: x25519,
2779 secp256r1, secp521r1, secp384r1.
2783 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2784 disabled by default. They can be re-enabled using the
2785 enable-weak-ssl-ciphers option to Configure.
2789 * If the server has ALPN configured, but supports no protocols that the
2790 client advertises, send a fatal "no_application_protocol" alert.
2791 This behaviour is SHALL in RFC 7301, though it isn't universally
2792 implemented by other servers.
2796 * Add X25519 support.
2797 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2798 for public and private key encoding using the format documented in
2799 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2800 key generation and key derivation.
2802 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2807 * Deprecate SRP_VBASE_get_by_user.
2808 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2809 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2810 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2811 seed, even if the seed is configured.
2813 Users should use SRP_VBASE_get1_by_user instead. Note that in
2814 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2815 also that even though configuring the SRP seed attempts to hide
2816 invalid usernames by continuing the handshake with fake
2817 credentials, this behaviour is not constant time and no strong
2818 guarantees are made that the handshake is indistinguishable from
2819 that of a valid user.
2823 * Configuration change; it's now possible to build dynamic engines
2824 without having to build shared libraries and vice versa. This
2825 only applies to the engines in engines/, those in crypto/engine/
2826 will always be built into libcrypto (i.e. "static").
2828 Building dynamic engines is enabled by default; to disable, use
2829 the configuration option "disable-dynamic-engine".
2831 The only requirements for building dynamic engines are the
2832 presence of the DSO module and building with position independent
2833 code, so they will also automatically be disabled if configuring
2834 with "disable-dso" or "disable-pic".
2836 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2837 are also taken away from openssl/opensslconf.h, as they are
2842 * Configuration change; if there is a known flag to compile
2843 position independent code, it will always be applied on the
2844 libcrypto and libssl object files, and never on the application
2845 object files. This means other libraries that use routines from
2846 libcrypto / libssl can be made into shared libraries regardless
2847 of how OpenSSL was configured.
2849 If this isn't desirable, the configuration options "disable-pic"
2850 or "no-pic" can be used to disable the use of PIC. This will
2851 also disable building shared libraries and dynamic engines.
2855 * Removed JPAKE code. It was experimental and has no wide use.
2859 * The INSTALL_PREFIX Makefile variable has been renamed to
2860 DESTDIR. That makes for less confusion on what this variable
2861 is for. Also, the configuration option --install_prefix is
2866 * Heartbeat for TLS has been removed and is disabled by default
2867 for DTLS; configure with enable-heartbeats. Code that uses the
2868 old #define's might need to be updated.
2870 *Emilia Käsper, Rich Salz*
2872 * Rename REF_CHECK to REF_DEBUG.
2876 * New "unified" build system
2878 The "unified" build system is aimed to be a common system for all
2879 platforms we support. With it comes new support for VMS.
2881 This system builds supports building in a different directory tree
2882 than the source tree. It produces one Makefile (for unix family
2883 or lookalikes), or one descrip.mms (for VMS).
2885 The source of information to make the Makefile / descrip.mms is
2886 small files called 'build.info', holding the necessary
2887 information for each directory with source to compile, and a
2888 template in Configurations, like unix-Makefile.tmpl or
2891 With this change, the library names were also renamed on Windows
2892 and on VMS. They now have names that are closer to the standard
2893 on Unix, and include the major version number, and in certain
2894 cases, the architecture they are built for. See "Notes on shared
2895 libraries" in INSTALL.
2897 We rely heavily on the perl module Text::Template.
2901 * Added support for auto-initialisation and de-initialisation of the library.
2902 OpenSSL no longer requires explicit init or deinit routines to be called,
2903 except in certain circumstances. See the OPENSSL_init_crypto() and
2904 OPENSSL_init_ssl() man pages for further information.
2908 * The arguments to the DTLSv1_listen function have changed. Specifically the
2909 "peer" argument is now expected to be a BIO_ADDR object.
2911 * Rewrite of BIO networking library. The BIO library lacked consistent
2912 support of IPv6, and adding it required some more extensive
2913 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2914 which hold all types of addresses and chains of address information.
2915 It also introduces a new API, with functions like BIO_socket,
2916 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2917 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2918 have been adapted accordingly.
2922 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2927 * CRIME protection: disable compression by default, even if OpenSSL is
2928 compiled with zlib enabled. Applications can still enable compression
2929 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2930 using the SSL_CONF library to configure compression.
2934 * The signature of the session callback configured with
2935 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2936 was explicitly marked as `const unsigned char*` instead of
2941 * Always DPURIFY. Remove the use of uninitialized memory in the
2942 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2946 * Removed many obsolete configuration items, including
2947 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2948 MD2_CHAR, MD2_INT, MD2_LONG
2950 IDEA_SHORT, IDEA_LONG
2951 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2953 *Rich Salz, with advice from Andy Polyakov*
2955 * Many BN internals have been moved to an internal header file.
2957 *Rich Salz with help from Andy Polyakov*
2959 * Configuration and writing out the results from it has changed.
2960 Files such as Makefile include/openssl/opensslconf.h and are now
2961 produced through general templates, such as Makefile.in and
2962 crypto/opensslconf.h.in and some help from the perl module
2965 Also, the center of configuration information is no longer
2966 Makefile. Instead, Configure produces a perl module in
2967 configdata.pm which holds most of the config data (in the hash
2968 table %config), the target data that comes from the target
2969 configuration in one of the `Configurations/*.conf` files (in
2974 * To clarify their intended purposes, the Configure options
2975 --prefix and --openssldir change their semantics, and become more
2976 straightforward and less interdependent.
2978 --prefix shall be used exclusively to give the location INSTALLTOP
2979 where programs, scripts, libraries, include files and manuals are
2980 going to be installed. The default is now /usr/local.
2982 --openssldir shall be used exclusively to give the default
2983 location OPENSSLDIR where certificates, private keys, CRLs are
2984 managed. This is also where the default openssl.cnf gets
2986 If the directory given with this option is a relative path, the
2987 values of both the --prefix value and the --openssldir value will
2988 be combined to become OPENSSLDIR.
2989 The default for --openssldir is INSTALLTOP/ssl.
2991 Anyone who uses --openssldir to specify where OpenSSL is to be
2992 installed MUST change to use --prefix instead.
2996 * The GOST engine was out of date and therefore it has been removed. An up
2997 to date GOST engine is now being maintained in an external repository.
2998 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2999 support for GOST ciphersuites (these are only activated if a GOST engine
3004 * EGD is no longer supported by default; use enable-egd when
3007 *Ben Kaduk and Rich Salz*
3009 * The distribution now has Makefile.in files, which are used to
3010 create Makefile's when Configure is run. *Configure must be run
3011 before trying to build now.*
3015 * The return value for SSL_CIPHER_description() for error conditions
3020 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3022 Obtaining and performing DNSSEC validation of TLSA records is
3023 the application's responsibility. The application provides
3024 the TLSA records of its choice to OpenSSL, and these are then
3025 used to authenticate the peer.
3027 The TLSA records need not even come from DNS. They can, for
3028 example, be used to implement local end-entity certificate or
3029 trust-anchor "pinning", where the "pin" data takes the form
3030 of TLSA records, which can augment or replace verification
3031 based on the usual WebPKI public certification authorities.
3035 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3036 continues to support deprecated interfaces in default builds.
3037 However, applications are strongly advised to compile their
3038 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3039 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3040 or the 1.1.0 releases.
3042 In environments in which all applications have been ported to
3043 not use any deprecated interfaces OpenSSL's Configure script
3044 should be used with the --api=1.1.0 option to entirely remove
3045 support for the deprecated features from the library and
3046 unconditionally disable them in the installed headers.
3047 Essentially the same effect can be achieved with the "no-deprecated"
3048 argument to Configure, except that this will always restrict
3049 the build to just the latest API, rather than a fixed API
3052 As applications are ported to future revisions of the API,
3053 they should update their compile-time OPENSSL_API_COMPAT define
3054 accordingly, but in most cases should be able to continue to
3055 compile with later releases.
3057 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3058 0x10000000L and 0x00908000L, respectively. However those
3059 versions did not support the OPENSSL_API_COMPAT feature, and
3060 so applications are not typically tested for explicit support
3061 of just the undeprecated features of either release.
3065 * Add support for setting the minimum and maximum supported protocol.
3066 It can bet set via the SSL_set_min_proto_version() and
3067 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3068 MaxProtocol. It's recommended to use the new APIs to disable
3069 protocols instead of disabling individual protocols using
3070 SSL_set_options() or SSL_CONF's Protocol. This change also
3071 removes support for disabling TLS 1.2 in the OpenSSL TLS
3072 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3076 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3080 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3081 and integrates ECDSA and ECDH functionality into EC. Implementations can
3082 now redirect key generation and no longer need to convert to or from
3085 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3086 include the ec.h header file instead.
3090 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3091 ciphers who are no longer supported and drops support the ephemeral RSA key
3092 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3096 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3097 opaque. For HMAC_CTX, the following constructors and destructors
3100 HMAC_CTX *HMAC_CTX_new(void);
3101 void HMAC_CTX_free(HMAC_CTX *ctx);
3103 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3104 destroy such methods has been added. See EVP_MD_meth_new(3) and
3105 EVP_CIPHER_meth_new(3) for documentation.
3108 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3109 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3110 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3111 an already created structure.
3112 2) For consistency with the majority of our object creators and
3113 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3114 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3115 for deprecated builds.
3119 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3120 cryptographic operations to be performed asynchronously as long as an
3121 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3122 further details. Libssl has also had this capability integrated with the
3123 introduction of the new mode SSL_MODE_ASYNC and associated error
3124 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3125 pages. This work was developed in partnership with Intel Corp.
3129 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3130 always enabled now. If you want to disable the support you should
3131 exclude it using the list of supported ciphers. This also means that the
3132 "-no_ecdhe" option has been removed from s_server.
3136 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3137 SSL_{CTX_}set1_curves() which can set a list.
3141 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3142 curve you want to support using SSL_{CTX_}set1_curves().
3146 * State machine rewrite. The state machine code has been significantly
3147 refactored in order to remove much duplication of code and solve issues
3148 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3149 further details). This change does have some associated API changes.
3150 Notably the SSL_state() function has been removed and replaced by
3151 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3152 SSL_set_state() has been removed altogether. The previous handshake states
3153 defined in ssl.h and ssl3.h have also been removed.
3157 * All instances of the string "ssleay" in the public API were replaced
3158 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3159 Some error codes related to internal RSA_eay API's were renamed.
3163 * The demo files in crypto/threads were moved to demo/threads.
3167 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3170 *Matt Caswell, Rich Salz*
3172 * New ASN.1 embed macro.
3174 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3175 structure is not allocated: it is part of the parent. That is instead of
3183 This reduces memory fragmentation and make it impossible to accidentally
3184 set a mandatory field to NULL.
3186 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3187 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3188 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3193 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3197 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3198 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3199 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3200 DES and RC4 ciphersuites.
3204 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3205 This changes the decoding behaviour for some invalid messages,
3206 though the change is mostly in the more lenient direction, and
3207 legacy behaviour is preserved as much as possible.
3211 * Fix no-stdio build.
3212 *David Woodhouse <David.Woodhouse@intel.com> and also*
3213 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3215 * New testing framework
3216 The testing framework has been largely rewritten and is now using
3217 perl and the perl modules Test::Harness and an extended variant of
3218 Test::More called OpenSSL::Test to do its work. All test scripts in
3219 test/ have been rewritten into test recipes, and all direct calls to
3220 executables in test/Makefile have become individual recipes using the
3221 simplified testing OpenSSL::Test::Simple.
3223 For documentation on our testing modules, do:
3225 perldoc test/testlib/OpenSSL/Test/Simple.pm
3226 perldoc test/testlib/OpenSSL/Test.pm
3230 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3231 are used; the latter aborts on memory leaks (usually checked on exit).
3232 Some undocumented "set malloc, etc., hooks" functions were removed
3233 and others were changed. All are now documented.
3237 * In DSA_generate_parameters_ex, if the provided seed is too short,
3240 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3242 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3243 from RFC4279, RFC4785, RFC5487, RFC5489.
3245 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3246 original RSA_PSK patch.
3250 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3251 era flag was never set throughout the codebase (only read). Also removed
3252 SSL3_FLAGS_POP_BUFFER which was only used if
3253 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3257 * Changed the default name options in the "ca", "crl", "req" and "x509"
3258 to be "oneline" instead of "compat".
3262 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3263 not aware of clients that still exhibit this bug, and the workaround
3264 hasn't been working properly for a while.
3268 * The return type of BIO_number_read() and BIO_number_written() as well as
3269 the corresponding num_read and num_write members in the BIO structure has
3270 changed from unsigned long to uint64_t. On platforms where an unsigned
3271 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3276 * Given the pervasive nature of TLS extensions it is inadvisable to run
3277 OpenSSL without support for them. It also means that maintaining
3278 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3279 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3283 * Removed support for the two export grade static DH ciphersuites
3284 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3285 were newly added (along with a number of other static DH ciphersuites) to
3286 1.0.2. However the two export ones have *never* worked since they were
3287 introduced. It seems strange in any case to be adding new export
3288 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3292 * Version negotiation has been rewritten. In particular SSLv23_method(),
3293 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3294 and turned into macros which simply call the new preferred function names
3295 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3296 should use the new names instead. Also as part of this change the ssl23.h
3297 header file has been removed.
3301 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3302 code and the associated standard is no longer considered fit-for-purpose.
3306 * RT2547 was closed. When generating a private key, try to make the
3307 output file readable only by the owner. This behavior change might
3308 be noticeable when interacting with other software.
3310 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3315 * Added HTTP GET support to the ocsp command.
3319 * Changed default digest for the dgst and enc commands from MD5 to
3324 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3328 * Added support for TLS extended master secret from
3329 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3330 initial patch which was a great help during development.
3334 * All libssl internal structures have been removed from the public header
3335 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3336 now redundant). Users should not attempt to access internal structures
3337 directly. Instead they should use the provided API functions.
3341 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3342 Access to deprecated functions can be re-enabled by running config with
3343 "enable-deprecated". In addition applications wishing to use deprecated
3344 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3345 will, by default, disable some transitive includes that previously existed
3346 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3350 * Added support for OCB mode. OpenSSL has been granted a patent license
3351 compatible with the OpenSSL license for use of OCB. Details are available
3352 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3353 for OCB can be removed by calling config with no-ocb.
3357 * SSLv2 support has been removed. It still supports receiving a SSLv2
3358 compatible client hello.
3362 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3363 done while fixing the error code for the key-too-small case.
3365 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3367 * CA.sh has been removed; use CA.pl instead.
3371 * Removed old DES API.
3375 * Remove various unsupported platforms:
3381 Sinix/ReliantUNIX RM400
3386 16-bit platforms such as WIN16
3390 * Clean up OPENSSL_NO_xxx #define's
3391 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3392 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3393 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3394 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3395 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3396 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3397 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3398 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3399 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3400 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3404 * Cleaned up dead code
3405 Remove all but one '#ifdef undef' which is to be looked at.
3409 * Clean up calling of xxx_free routines.
3410 Just like free(), fix most of the xxx_free routines to accept
3411 NULL. Remove the non-null checks from callers. Save much code.
3415 * Add secure heap for storage of private keys (when possible).
3416 Add BIO_s_secmem(), CBIGNUM, etc.
3417 Contributed by Akamai Technologies under our Corporate CLA.
3421 * Experimental support for a new, fast, unbiased prime candidate generator,
3422 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3424 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3426 * New output format NSS in the sess_id command line tool. This allows
3427 exporting the session id and the master key in NSS keylog format.
3429 *Martin Kaiser <martin@kaiser.cx>*
3431 * Harmonize version and its documentation. -f flag is used to display
3434 *mancha <mancha1@zoho.com>*
3436 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3437 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3439 *mancha <mancha1@zoho.com>*
3441 * Fix some double frees. These are not thought to be exploitable.
3443 *mancha <mancha1@zoho.com>*
3445 * A missing bounds check in the handling of the TLS heartbeat extension
3446 can be used to reveal up to 64k of memory to a connected client or
3449 Thanks for Neel Mehta of Google Security for discovering this bug and to
3450 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3451 preparing the fix [CVE-2014-0160][]
3453 *Adam Langley, Bodo Moeller*
3455 * Fix for the attack described in the paper "Recovering OpenSSL
3456 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3457 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3458 <http://eprint.iacr.org/2014/140>
3460 Thanks to Yuval Yarom and Naomi Benger for discovering this
3461 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3463 *Yuval Yarom and Naomi Benger*
3465 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3466 this fixes a limitation in previous versions of OpenSSL.
3470 * Experimental encrypt-then-mac support.
3472 Experimental support for encrypt then mac from
3473 draft-gutmann-tls-encrypt-then-mac-02.txt
3475 To enable it set the appropriate extension number (0x42 for the test
3476 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3478 For non-compliant peers (i.e. just about everything) this should have no
3481 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3485 * Add EVP support for key wrapping algorithms, to avoid problems with
3486 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3487 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3488 algorithms and include tests cases.
3492 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3497 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3498 MGF1 digest and OAEP label.
3502 * Make openssl verify return errors.
3504 *Chris Palmer <palmer@google.com> and Ben Laurie*
3506 * New function ASN1_TIME_diff to calculate the difference between two
3507 ASN1_TIME structures or one structure and the current time.
3511 * Update fips_test_suite to support multiple command line options. New
3512 test to induce all self test errors in sequence and check expected
3517 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3518 sign or verify all in one operation.
3522 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3523 test programs and fips_test_suite. Includes functionality to parse
3524 the minimal script output of fipsalgest.pl directly.
3528 * Add authorisation parameter to FIPS_module_mode_set().
3532 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3536 * Use separate DRBG fields for internal and external flags. New function
3537 FIPS_drbg_health_check() to perform on demand health checking. Add
3538 generation tests to fips_test_suite with reduced health check interval to
3539 demonstrate periodic health checking. Add "nodh" option to
3540 fips_test_suite to skip very slow DH test.
3544 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3549 * More extensive health check for DRBG checking many more failure modes.
3550 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3551 combination: call this in fips_test_suite.
3555 * Add support for canonical generation of DSA parameter 'g'. See
3558 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3559 POST to handle HMAC cases.
3563 * Add functions FIPS_module_version() and FIPS_module_version_text()
3564 to return numerical and string versions of the FIPS module number.
3568 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3569 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3570 outside the validated module in the FIPS capable OpenSSL.
3574 * Minor change to DRBG entropy callback semantics. In some cases
3575 there is no multiple of the block length between min_len and
3576 max_len. Allow the callback to return more than max_len bytes
3577 of entropy but discard any extra: it is the callback's responsibility
3578 to ensure that the extra data discarded does not impact the
3579 requested amount of entropy.
3583 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3584 information in FIPS186-3, SP800-57 and SP800-131A.
3588 * CCM support via EVP. Interface is very similar to GCM case except we
3589 must supply all data in one chunk (i.e. no update, final) and the
3590 message length must be supplied if AAD is used. Add algorithm test
3595 * Initial version of POST overhaul. Add POST callback to allow the status
3596 of POST to be monitored and/or failures induced. Modify fips_test_suite
3597 to use callback. Always run all selftests even if one fails.
3601 * XTS support including algorithm test driver in the fips_gcmtest program.
3602 Note: this does increase the maximum key length from 32 to 64 bytes but
3603 there should be no binary compatibility issues as existing applications
3604 will never use XTS mode.
3608 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3609 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3610 performs algorithm blocking for unapproved PRNG types. Also do not
3611 set PRNG type in FIPS_mode_set(): leave this to the application.
3612 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3613 the standard OpenSSL PRNG: set additional data to a date time vector.
3617 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3618 This shouldn't present any incompatibility problems because applications
3619 shouldn't be using these directly and any that are will need to rethink
3620 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3624 * Extensive self tests and health checking required by SP800-90 DRBG.
3625 Remove strength parameter from FIPS_drbg_instantiate and always
3626 instantiate at maximum supported strength.
3630 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3634 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3638 * New function DH_compute_key_padded() to compute a DH key and pad with
3639 leading zeroes if needed: this complies with SP800-56A et al.
3643 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3644 anything, incomplete, subject to change and largely untested at present.
3648 * Modify fipscanisteronly build option to only build the necessary object
3649 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3653 * Add experimental option FIPSSYMS to give all symbols in
3654 fipscanister.o and FIPS or fips prefix. This will avoid
3655 conflicts with future versions of OpenSSL. Add perl script
3656 util/fipsas.pl to preprocess assembly language source files
3657 and rename any affected symbols.
3661 * Add selftest checks and algorithm block of non-fips algorithms in
3662 FIPS mode. Remove DES2 from selftests.
3666 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3667 return internal method without any ENGINE dependencies. Add new
3668 tiny fips sign and verify functions.
3672 * New build option no-ec2m to disable characteristic 2 code.
3676 * New build option "fipscanisteronly". This only builds fipscanister.o
3677 and (currently) associated fips utilities. Uses the file Makefile.fips
3678 instead of Makefile.org as the prototype.
3682 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3683 Update fips_gcmtest to use IV generator.
3687 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3688 setting output buffer to NULL. The `*Final` function must be
3689 called although it will not retrieve any additional data. The tag
3690 can be set or retrieved with a ctrl. The IV length is by default 12
3691 bytes (96 bits) but can be set to an alternative value. If the IV
3692 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3697 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3698 underlying do_cipher function handles all cipher semantics itself
3699 including padding and finalisation. This is useful if (for example)
3700 an ENGINE cipher handles block padding itself. The behaviour of
3701 do_cipher is subtly changed if this flag is set: the return value
3702 is the number of characters written to the output buffer (zero is
3703 no longer an error code) or a negative error code. Also if the
3704 input buffer is NULL and length 0 finalisation should be performed.
3708 * If a candidate issuer certificate is already part of the constructed
3709 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3713 * Improve forward-security support: add functions
3715 void SSL_CTX_set_not_resumable_session_callback(
3716 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3717 void SSL_set_not_resumable_session_callback(
3718 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3720 for use by SSL/TLS servers; the callback function will be called whenever a
3721 new session is created, and gets to decide whether the session may be
3722 cached to make it resumable (return 0) or not (return 1). (As by the
3723 SSL/TLS protocol specifications, the session_id sent by the server will be
3724 empty to indicate that the session is not resumable; also, the server will
3725 not generate RFC 4507 (RFC 5077) session tickets.)
3727 A simple reasonable callback implementation is to return is_forward_secure.
3728 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3729 by the SSL/TLS server library, indicating whether it can provide forward
3732 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3734 * New -verify_name option in command line utilities to set verification
3739 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3740 Add CMAC pkey methods.
3744 * Experimental renegotiation in s_server -www mode. If the client
3745 browses /reneg connection is renegotiated. If /renegcert it is
3746 renegotiated requesting a certificate.
3750 * Add an "external" session cache for debugging purposes to s_server. This
3751 should help trace issues which normally are only apparent in deployed
3752 multi-process servers.
3756 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3757 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3758 BIO_set_cipher() and some obscure PEM functions were changed so they
3759 can now return an error. The RAND changes required a change to the
3760 RAND_METHOD structure.
3764 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3765 a gcc attribute to warn if the result of a function is ignored. This
3766 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3767 whose return value is often ignored.
3771 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3772 These allow SCTs (signed certificate timestamps) to be requested and
3773 validated when establishing a connection.
3775 *Rob Percival <robpercival@google.com>*
3780 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3782 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3783 used even when parsing explicit parameters, when loading a serialized key
3784 or calling `EC_GROUP_new_from_ecpkparameters()`/
3785 `EC_GROUP_new_from_ecparameters()`.
3786 This prevents bypass of security hardening and performance gains,
3787 especially for curves with specialized EC_METHODs.
3788 By default, if a key encoded with explicit parameters is loaded and later
3789 serialized, the output is still encoded with explicit parameters, even if
3790 internally a "named" EC_GROUP is used for computation.
3794 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3795 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3796 NULL. After this change, only the cofactor parameter can be NULL. It also
3797 does some minimal sanity checks on the passed order.
3802 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3803 An attack is simple, if the first CMS_recipientInfo is valid but the
3804 second CMS_recipientInfo is chosen ciphertext. If the second
3805 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3806 encryption key will be replaced by garbage, and the message cannot be
3807 decoded, but if the RSA decryption fails, the correct encryption key is
3808 used and the recipient will not notice the attack.
3809 As a work around for this potential attack the length of the decrypted
3810 key must be equal to the cipher default key length, in case the
3811 certifiate is not given and all recipientInfo are tried out.
3812 The old behaviour can be re-enabled in the CMS code by setting the
3813 CMS_DEBUG_DECRYPT flag.
3818 * Document issue with installation paths in diverse Windows builds
3820 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3821 binaries and run-time config file.
3826 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3828 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3829 This changes the size when using the genpkey app when no size is given. It
3830 fixes an omission in earlier changes that changed all RSA, DSA and DH
3831 generation apps to use 2048 bits by default.
3835 * Add FIPS support for Android Arm 64-bit
3837 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3838 Module in Version 2.0.10. For some reason, the corresponding target
3839 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3840 built with FIPS support on Android Arm 64-bit. This omission has been
3843 *Matthias St. Pierre*
3845 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3847 * 0-byte record padding oracle
3849 If an application encounters a fatal protocol error and then calls
3850 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3851 then OpenSSL can respond differently to the calling application if a 0 byte
3852 record is received with invalid padding compared to if a 0 byte record is
3853 received with an invalid MAC. If the application then behaves differently
3854 based on that in a way that is detectable to the remote peer, then this
3855 amounts to a padding oracle that could be used to decrypt data.
3857 In order for this to be exploitable "non-stitched" ciphersuites must be in
3858 use. Stitched ciphersuites are optimised implementations of certain
3859 commonly used ciphersuites. Also the application must call SSL_shutdown()
3860 twice even if a protocol error has occurred (applications should not do
3861 this but some do anyway).
3863 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3864 Aviram, with additional investigation by Steven Collison and Andrew
3865 Hourselt. It was reported to OpenSSL on 10th December 2018.
3870 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3874 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3876 * Microarchitecture timing vulnerability in ECC scalar multiplication
3878 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3879 shown to be vulnerable to a microarchitecture timing side channel attack.
3880 An attacker with sufficient access to mount local timing attacks during
3881 ECDSA signature generation could recover the private key.
3883 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3884 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3890 * Timing vulnerability in DSA signature generation
3892 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3893 timing side channel attack. An attacker could use variations in the signing
3894 algorithm to recover the private key.
3896 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3901 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3902 Module, accidentally introduced while backporting security fixes from the
3903 development branch and hindering the use of ECC in FIPS mode.
3907 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3909 * Client DoS due to large DH parameter
3911 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3912 malicious server can send a very large prime value to the client. This will
3913 cause the client to spend an unreasonably long period of time generating a
3914 key for this prime resulting in a hang until the client has finished. This
3915 could be exploited in a Denial Of Service attack.
3917 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3922 * Cache timing vulnerability in RSA Key Generation
3924 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3925 a cache timing side channel attack. An attacker with sufficient access to
3926 mount cache timing attacks during the RSA key generation process could
3927 recover the private key.
3929 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3930 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3935 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3936 parameter is no longer accepted, as it leads to a corrupt table. NULL
3937 pem_str is reserved for alias entries only.
3941 * Revert blinding in ECDSA sign and instead make problematic addition
3942 length-invariant. Switch even to fixed-length Montgomery multiplication.
3946 * Change generating and checking of primes so that the error rate of not
3947 being prime depends on the intended use based on the size of the input.
3948 For larger primes this will result in more rounds of Miller-Rabin.
3949 The maximal error rate for primes with more than 1080 bits is lowered
3952 *Kurt Roeckx, Annie Yousar*
3954 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3958 * Add blinding to ECDSA and DSA signatures to protect against side channel
3959 attacks discovered by Keegan Ryan (NCC Group).
3963 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3964 now allow empty (zero character) pass phrases.
3968 * Certificate time validation (X509_cmp_time) enforces stricter
3969 compliance with RFC 5280. Fractional seconds and timezone offsets
3970 are no longer allowed.
3974 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3976 * Constructed ASN.1 types with a recursive definition could exceed the stack
3978 Constructed ASN.1 types with a recursive definition (such as can be found
3979 in PKCS7) could eventually exceed the stack given malicious input with
3980 excessive recursion. This could result in a Denial Of Service attack. There
3981 are no such structures used within SSL/TLS that come from untrusted sources
3982 so this is considered safe.
3984 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3990 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3992 * Read/write after SSL object in error state
3994 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3995 mechanism. The intent was that if a fatal error occurred during a handshake
3996 then OpenSSL would move into the error state and would immediately fail if
3997 you attempted to continue the handshake. This works as designed for the
3998 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3999 SSL_connect()), however due to a bug it does not work correctly if
4000 SSL_read() or SSL_write() is called directly. In that scenario, if the
4001 handshake fails then a fatal error will be returned in the initial function
4002 call. If SSL_read()/SSL_write() is subsequently called by the application
4003 for the same SSL object then it will succeed and the data is passed without
4004 being decrypted/encrypted directly from the SSL/TLS record layer.
4006 In order to exploit this issue an application bug would have to be present
4007 that resulted in a call to SSL_read()/SSL_write() being issued after having
4008 already received a fatal error.
4010 This issue was reported to OpenSSL by David Benjamin (Google).
4015 * rsaz_1024_mul_avx2 overflow bug on x86_64
4017 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4018 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4019 Analysis suggests that attacks against RSA and DSA as a result of this
4020 defect would be very difficult to perform and are not believed likely.
4021 Attacks against DH1024 are considered just feasible, because most of the
4022 work necessary to deduce information about a private key may be performed
4023 offline. The amount of resources required for such an attack would be
4024 significant. However, for an attack on TLS to be meaningful, the server
4025 would have to share the DH1024 private key among multiple clients, which is
4026 no longer an option since CVE-2016-0701.
4028 This only affects processors that support the AVX2 but not ADX extensions
4029 like Intel Haswell (4th generation).
4031 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4032 was originally found via the OSS-Fuzz project.
4037 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4039 * bn_sqrx8x_internal carry bug on x86_64
4041 There is a carry propagating bug in the x86_64 Montgomery squaring
4042 procedure. No EC algorithms are affected. Analysis suggests that attacks
4043 against RSA and DSA as a result of this defect would be very difficult to
4044 perform and are not believed likely. Attacks against DH are considered just
4045 feasible (although very difficult) because most of the work necessary to
4046 deduce information about a private key may be performed offline. The amount
4047 of resources required for such an attack would be very significant and
4048 likely only accessible to a limited number of attackers. An attacker would
4049 additionally need online access to an unpatched system using the target
4050 private key in a scenario with persistent DH parameters and a private
4051 key that is shared between multiple clients.
4053 This only affects processors that support the BMI1, BMI2 and ADX extensions
4054 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4056 This issue was reported to OpenSSL by the OSS-Fuzz project.
4061 * Malformed X.509 IPAddressFamily could cause OOB read
4063 If an X.509 certificate has a malformed IPAddressFamily extension,
4064 OpenSSL could do a one-byte buffer overread. The most likely result
4065 would be an erroneous display of the certificate in text format.
4067 This issue was reported to OpenSSL by the OSS-Fuzz project.
4072 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4074 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4075 platform rather than 'mingw'.
4079 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4081 * Truncated packet could crash via OOB read
4083 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4084 cipher is being used, then a truncated packet can cause that host to
4085 perform an out-of-bounds read, usually resulting in a crash.
4087 This issue was reported to OpenSSL by Robert Święcki of Google.
4092 * BN_mod_exp may produce incorrect results on x86_64
4094 There is a carry propagating bug in the x86_64 Montgomery squaring
4095 procedure. No EC algorithms are affected. Analysis suggests that attacks
4096 against RSA and DSA as a result of this defect would be very difficult to
4097 perform and are not believed likely. Attacks against DH are considered just
4098 feasible (although very difficult) because most of the work necessary to
4099 deduce information about a private key may be performed offline. The amount
4100 of resources required for such an attack would be very significant and
4101 likely only accessible to a limited number of attackers. An attacker would
4102 additionally need online access to an unpatched system using the target
4103 private key in a scenario with persistent DH parameters and a private
4104 key that is shared between multiple clients. For example this can occur by
4105 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4106 similar to CVE-2015-3193 but must be treated as a separate problem.
4108 This issue was reported to OpenSSL by the OSS-Fuzz project.
4113 * Montgomery multiplication may produce incorrect results
4115 There is a carry propagating bug in the Broadwell-specific Montgomery
4116 multiplication procedure that handles input lengths divisible by, but
4117 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4118 and DH private keys are impossible. This is because the subroutine in
4119 question is not used in operations with the private key itself and an input
4120 of the attacker's direct choice. Otherwise the bug can manifest itself as
4121 transient authentication and key negotiation failures or reproducible
4122 erroneous outcome of public-key operations with specially crafted input.
4123 Among EC algorithms only Brainpool P-512 curves are affected and one
4124 presumably can attack ECDH key negotiation. Impact was not analyzed in
4125 detail, because pre-requisites for attack are considered unlikely. Namely
4126 multiple clients have to choose the curve in question and the server has to
4127 share the private key among them, neither of which is default behaviour.
4128 Even then only clients that chose the curve will be affected.
4130 This issue was publicly reported as transient failures and was not
4131 initially recognized as a security issue. Thanks to Richard Morgan for
4132 providing reproducible case.
4137 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4138 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4139 prevent issues where no progress is being made and the peer continually
4140 sends unrecognised record types, using up resources processing them.
4144 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4146 * Missing CRL sanity check
4148 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4149 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4150 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4152 This issue only affects the OpenSSL 1.0.2i
4157 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4159 * OCSP Status Request extension unbounded memory growth
4161 A malicious client can send an excessively large OCSP Status Request
4162 extension. If that client continually requests renegotiation, sending a
4163 large OCSP Status Request extension each time, then there will be unbounded
4164 memory growth on the server. This will eventually lead to a Denial Of
4165 Service attack through memory exhaustion. Servers with a default
4166 configuration are vulnerable even if they do not support OCSP. Builds using
4167 the "no-ocsp" build time option are not affected.
4169 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4174 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4177 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4183 * OOB write in MDC2_Update()
4185 An overflow can occur in MDC2_Update() either if called directly or
4186 through the EVP_DigestUpdate() function using MDC2. If an attacker
4187 is able to supply very large amounts of input data after a previous
4188 call to EVP_EncryptUpdate() with a partial block then a length check
4189 can overflow resulting in a heap corruption.
4191 The amount of data needed is comparable to SIZE_MAX which is impractical
4194 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4199 * Malformed SHA512 ticket DoS
4201 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4202 DoS attack where a malformed ticket will result in an OOB read which will
4205 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4206 a custom server callback and ticket lookup mechanism.
4208 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4213 * OOB write in BN_bn2dec()
4215 The function BN_bn2dec() does not check the return value of BN_div_word().
4216 This can cause an OOB write if an application uses this function with an
4217 overly large BIGNUM. This could be a problem if an overly large certificate
4218 or CRL is printed out from an untrusted source. TLS is not affected because
4219 record limits will reject an oversized certificate before it is parsed.
4221 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4226 * OOB read in TS_OBJ_print_bio()
4228 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4229 the total length the OID text representation would use and not the amount
4230 of data written. This will result in OOB reads when large OIDs are
4233 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4238 * Pointer arithmetic undefined behaviour
4240 Avoid some undefined pointer arithmetic
4242 A common idiom in the codebase is to check limits in the following manner:
4245 Where "p" points to some malloc'd data of SIZE bytes and
4248 "len" here could be from some externally supplied data (e.g. from a TLS
4251 The rules of C pointer arithmetic are such that "p + len" is only well
4252 defined where len <= SIZE. Therefore the above idiom is actually
4253 undefined behaviour.
4255 For example this could cause problems if some malloc implementation
4256 provides an address for "p" such that "p + len" actually overflows for
4257 values of len that are too big and therefore p + len < limit.
4259 This issue was reported to OpenSSL by Guido Vranken
4264 * Constant time flag not preserved in DSA signing
4266 Operations in the DSA signing algorithm should run in constant time in
4267 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4268 implementation means that a non-constant time codepath is followed for
4269 certain operations. This has been demonstrated through a cache-timing
4270 attack to be sufficient for an attacker to recover the private DSA key.
4272 This issue was reported by César Pereida (Aalto University), Billy Brumley
4273 (Tampere University of Technology), and Yuval Yarom (The University of
4274 Adelaide and NICTA).
4279 * DTLS buffered message DoS
4281 In a DTLS connection where handshake messages are delivered out-of-order
4282 those messages that OpenSSL is not yet ready to process will be buffered
4283 for later use. Under certain circumstances, a flaw in the logic means that
4284 those messages do not get removed from the buffer even though the handshake
4285 has been completed. An attacker could force up to approx. 15 messages to
4286 remain in the buffer when they are no longer required. These messages will
4287 be cleared when the DTLS connection is closed. The default maximum size for
4288 a message is 100k. Therefore the attacker could force an additional 1500k
4289 to be consumed per connection. By opening many simulataneous connections an
4290 attacker could cause a DoS attack through memory exhaustion.
4292 This issue was reported to OpenSSL by Quan Luo.
4297 * DTLS replay protection DoS
4299 A flaw in the DTLS replay attack protection mechanism means that records
4300 that arrive for future epochs update the replay protection "window" before
4301 the MAC for the record has been validated. This could be exploited by an
4302 attacker by sending a record for the next epoch (which does not have to
4303 decrypt or have a valid MAC), with a very large sequence number. This means
4304 that all subsequent legitimate packets are dropped causing a denial of
4305 service for a specific DTLS connection.
4307 This issue was reported to OpenSSL by the OCAP audit team.
4312 * Certificate message OOB reads
4314 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4315 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4316 theoretical DoS risk but this has not been observed in practice on common
4319 The messages affected are client certificate, client certificate request
4320 and server certificate. As a result the attack can only be performed
4321 against a client or a server which enables client authentication.
4323 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4328 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4330 * Prevent padding oracle in AES-NI CBC MAC check
4332 A MITM attacker can use a padding oracle attack to decrypt traffic
4333 when the connection uses an AES CBC cipher and the server support
4336 This issue was introduced as part of the fix for Lucky 13 padding
4337 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4338 constant time by making sure that always the same bytes are read and
4339 compared against either the MAC or padding bytes. But it no longer
4340 checked that there was enough data to have both the MAC and padding
4343 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4348 * Fix EVP_EncodeUpdate overflow
4350 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4351 Base64 encoding of binary data. If an attacker is able to supply very large
4352 amounts of input data then a length check can overflow resulting in a heap
4355 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4356 the `PEM_write_bio*` family of functions. These are mainly used within the
4357 OpenSSL command line applications, so any application which processes data
4358 from an untrusted source and outputs it as a PEM file should be considered
4359 vulnerable to this issue. User applications that call these APIs directly
4360 with large amounts of untrusted data may also be vulnerable.
4362 This issue was reported by Guido Vranken.
4367 * Fix EVP_EncryptUpdate overflow
4369 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4370 is able to supply very large amounts of input data after a previous call to
4371 EVP_EncryptUpdate() with a partial block then a length check can overflow
4372 resulting in a heap corruption. Following an analysis of all OpenSSL
4373 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4374 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4375 the first called function after an EVP_EncryptInit(), and therefore that
4376 specific call must be safe. The second form is where the length passed to
4377 EVP_EncryptUpdate() can be seen from the code to be some small value and
4378 therefore there is no possibility of an overflow. Since all instances are
4379 one of these two forms, it is believed that there can be no overflows in
4380 internal code due to this problem. It should be noted that
4381 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4382 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4383 of these calls have also been analysed too and it is believed there are no
4384 instances in internal usage where an overflow could occur.
4386 This issue was reported by Guido Vranken.
4391 * Prevent ASN.1 BIO excessive memory allocation
4393 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4394 a short invalid encoding can cause allocation of large amounts of memory
4395 potentially consuming excessive resources or exhausting memory.
4397 Any application parsing untrusted data through d2i BIO functions is
4398 affected. The memory based functions such as d2i_X509() are *not* affected.
4399 Since the memory based functions are used by the TLS library, TLS
4400 applications are not affected.
4402 This issue was reported by Brian Carpenter.
4409 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4410 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4411 in arbitrary stack data being returned in the buffer.
4413 This issue was reported by Guido Vranken.
4418 * Modify behavior of ALPN to invoke callback after SNI/servername
4419 callback, such that updates to the SSL_CTX affect ALPN.
4423 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4428 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4429 methods are enabled and ssl2 is disabled the methods return NULL.
4433 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4435 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4436 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4437 provide any "EXPORT" or "LOW" strength ciphers.
4441 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4442 is by default disabled at build-time. Builds that are not configured with
4443 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4444 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4445 will need to explicitly call either of:
4447 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4449 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4451 as appropriate. Even if either of those is used, or the application
4452 explicitly uses the version-specific SSLv2_method() or its client and
4453 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4454 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4455 ciphers, and SSLv2 56-bit DES are no longer available.
4460 * Fix a double-free in DSA code
4462 A double free bug was discovered when OpenSSL parses malformed DSA private
4463 keys and could lead to a DoS attack or memory corruption for applications
4464 that receive DSA private keys from untrusted sources. This scenario is
4467 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4473 * Disable SRP fake user seed to address a server memory leak.
4475 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4477 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4478 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4479 was changed to ignore the "fake user" SRP seed, even if the seed
4482 Users should use SRP_VBASE_get1_by_user instead. Note that in
4483 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4484 also that even though configuring the SRP seed attempts to hide
4485 invalid usernames by continuing the handshake with fake
4486 credentials, this behaviour is not constant time and no strong
4487 guarantees are made that the handshake is indistinguishable from
4488 that of a valid user.
4493 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4495 In the BN_hex2bn function the number of hex digits is calculated using an
4496 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4497 large values of `i` this can result in `bn_expand` not allocating any
4498 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4499 field as NULL leading to a subsequent NULL ptr deref. For very large values
4500 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4501 In this case memory is allocated to the internal BIGNUM data field, but it
4502 is insufficiently sized leading to heap corruption. A similar issue exists
4503 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4504 is ever called by user applications with very large untrusted hex/dec data.
4505 This is anticipated to be a rare occurrence.
4507 All OpenSSL internal usage of these functions use data that is not expected
4508 to be untrusted, e.g. config file data or application command line
4509 arguments. If user developed applications generate config file data based
4510 on untrusted data then it is possible that this could also lead to security
4511 consequences. This is also anticipated to be rare.
4513 This issue was reported to OpenSSL by Guido Vranken.
4518 * Fix memory issues in `BIO_*printf` functions
4520 The internal `fmtstr` function used in processing a "%s" format string in
4521 the `BIO_*printf` functions could overflow while calculating the length of a
4522 string and cause an OOB read when printing very long strings.
4524 Additionally the internal `doapr_outch` function can attempt to write to an
4525 OOB memory location (at an offset from the NULL pointer) in the event of a
4526 memory allocation failure. In 1.0.2 and below this could be caused where
4527 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4528 could be in processing a very long "%s" format string. Memory leaks can
4531 The first issue may mask the second issue dependent on compiler behaviour.
4532 These problems could enable attacks where large amounts of untrusted data
4533 is passed to the `BIO_*printf` functions. If applications use these functions
4534 in this way then they could be vulnerable. OpenSSL itself uses these
4535 functions when printing out human-readable dumps of ASN.1 data. Therefore
4536 applications that print this data could be vulnerable if the data is from
4537 untrusted sources. OpenSSL command line applications could also be
4538 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4539 as command line arguments.
4541 Libssl is not considered directly vulnerable. Additionally certificates etc
4542 received via remote connections via libssl are also unlikely to be able to
4543 trigger these issues because of message size limits enforced within libssl.
4545 This issue was reported to OpenSSL Guido Vranken.
4550 * Side channel attack on modular exponentiation
4552 A side-channel attack was found which makes use of cache-bank conflicts on
4553 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4554 of RSA keys. The ability to exploit this issue is limited as it relies on
4555 an attacker who has control of code in a thread running on the same
4556 hyper-threaded core as the victim thread which is performing decryptions.
4558 This issue was reported to OpenSSL by Yuval Yarom, The University of
4559 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4560 Nadia Heninger, University of Pennsylvania with more information at
4561 <http://cachebleed.info>.
4566 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4567 if no keysize is specified with default_bits. This fixes an
4568 omission in an earlier change that changed all RSA/DSA key generation
4569 apps to use 2048 bits by default.
4573 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4575 * DH small subgroups
4577 Historically OpenSSL only ever generated DH parameters based on "safe"
4578 primes. More recently (in version 1.0.2) support was provided for
4579 generating X9.42 style parameter files such as those required for RFC 5114
4580 support. The primes used in such files may not be "safe". Where an
4581 application is using DH configured with parameters based on primes that are
4582 not "safe" then an attacker could use this fact to find a peer's private
4583 DH exponent. This attack requires that the attacker complete multiple
4584 handshakes in which the peer uses the same private DH exponent. For example
4585 this could be used to discover a TLS server's private DH exponent if it's
4586 reusing the private DH exponent or it's using a static DH ciphersuite.
4588 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4589 TLS. It is not on by default. If the option is not set then the server
4590 reuses the same private DH exponent for the life of the server process and
4591 would be vulnerable to this attack. It is believed that many popular
4592 applications do set this option and would therefore not be at risk.
4594 The fix for this issue adds an additional check where a "q" parameter is
4595 available (as is the case in X9.42 based parameters). This detects the
4596 only known attack, and is the only possible defense for static DH
4597 ciphersuites. This could have some performance impact.
4599 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4600 default and cannot be disabled. This could have some performance impact.
4602 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4607 * SSLv2 doesn't block disabled ciphers
4609 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4610 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4611 been disabled, provided that the SSLv2 protocol was not also disabled via
4614 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4615 and Sebastian Schinzel.
4620 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4622 * BN_mod_exp may produce incorrect results on x86_64
4624 There is a carry propagating bug in the x86_64 Montgomery squaring
4625 procedure. No EC algorithms are affected. Analysis suggests that attacks
4626 against RSA and DSA as a result of this defect would be very difficult to
4627 perform and are not believed likely. Attacks against DH are considered just
4628 feasible (although very difficult) because most of the work necessary to
4629 deduce information about a private key may be performed offline. The amount
4630 of resources required for such an attack would be very significant and
4631 likely only accessible to a limited number of attackers. An attacker would
4632 additionally need online access to an unpatched system using the target
4633 private key in a scenario with persistent DH parameters and a private
4634 key that is shared between multiple clients. For example this can occur by
4635 default in OpenSSL DHE based SSL/TLS ciphersuites.
4637 This issue was reported to OpenSSL by Hanno Böck.
4642 * Certificate verify crash with missing PSS parameter
4644 The signature verification routines will crash with a NULL pointer
4645 dereference if presented with an ASN.1 signature using the RSA PSS
4646 algorithm and absent mask generation function parameter. Since these
4647 routines are used to verify certificate signature algorithms this can be
4648 used to crash any certificate verification operation and exploited in a
4649 DoS attack. Any application which performs certificate verification is
4650 vulnerable including OpenSSL clients and servers which enable client
4653 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4658 * X509_ATTRIBUTE memory leak
4660 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4661 memory. This structure is used by the PKCS#7 and CMS routines so any
4662 application which reads PKCS#7 or CMS data from untrusted sources is
4663 affected. SSL/TLS is not affected.
4665 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4671 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4672 This changes the decoding behaviour for some invalid messages,
4673 though the change is mostly in the more lenient direction, and
4674 legacy behaviour is preserved as much as possible.
4678 * In DSA_generate_parameters_ex, if the provided seed is too short,
4681 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4683 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4685 * Alternate chains certificate forgery
4687 During certificate verification, OpenSSL will attempt to find an
4688 alternative certificate chain if the first attempt to build such a chain
4689 fails. An error in the implementation of this logic can mean that an
4690 attacker could cause certain checks on untrusted certificates to be
4691 bypassed, such as the CA flag, enabling them to use a valid leaf
4692 certificate to act as a CA and "issue" an invalid certificate.
4694 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4699 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4701 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4702 incompatibility in the handling of HMAC. The previous ABI has now been
4707 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4709 * Malformed ECParameters causes infinite loop
4711 When processing an ECParameters structure OpenSSL enters an infinite loop
4712 if the curve specified is over a specially malformed binary polynomial
4715 This can be used to perform denial of service against any
4716 system which processes public keys, certificate requests or
4717 certificates. This includes TLS clients and TLS servers with
4718 client authentication enabled.
4720 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4725 * Exploitable out-of-bounds read in X509_cmp_time
4727 X509_cmp_time does not properly check the length of the ASN1_TIME
4728 string and can read a few bytes out of bounds. In addition,
4729 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4732 An attacker can use this to craft malformed certificates and CRLs of
4733 various sizes and potentially cause a segmentation fault, resulting in
4734 a DoS on applications that verify certificates or CRLs. TLS clients
4735 that verify CRLs are affected. TLS clients and servers with client
4736 authentication enabled may be affected if they use custom verification
4739 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4740 independently by Hanno Böck.
4745 * PKCS7 crash with missing EnvelopedContent
4747 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4748 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4749 with missing content and trigger a NULL pointer dereference on parsing.
4751 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4752 structures from untrusted sources are affected. OpenSSL clients and
4753 servers are not affected.
4755 This issue was reported to OpenSSL by Michal Zalewski (Google).
4760 * CMS verify infinite loop with unknown hash function
4762 When verifying a signedData message the CMS code can enter an infinite loop
4763 if presented with an unknown hash function OID. This can be used to perform
4764 denial of service against any system which verifies signedData messages using
4766 This issue was reported to OpenSSL by Johannes Bauer.
4771 * Race condition handling NewSessionTicket
4773 If a NewSessionTicket is received by a multi-threaded client when attempting to
4774 reuse a previous ticket then a race condition can occur potentially leading to
4775 a double free of the ticket data.
4780 * Only support 256-bit or stronger elliptic curves with the
4781 'ecdh_auto' setting (server) or by default (client). Of supported
4782 curves, prefer P-256 (both).
4786 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4788 * ClientHello sigalgs DoS fix
4790 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4791 invalid signature algorithms extension a NULL pointer dereference will
4792 occur. This can be exploited in a DoS attack against the server.
4794 This issue was was reported to OpenSSL by David Ramos of Stanford
4798 *Stephen Henson and Matt Caswell*
4800 * Multiblock corrupted pointer fix
4802 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4803 feature only applies on 64 bit x86 architecture platforms that support AES
4804 NI instructions. A defect in the implementation of "multiblock" can cause
4805 OpenSSL's internal write buffer to become incorrectly set to NULL when
4806 using non-blocking IO. Typically, when the user application is using a
4807 socket BIO for writing, this will only result in a failed connection.
4808 However if some other BIO is used then it is likely that a segmentation
4809 fault will be triggered, thus enabling a potential DoS attack.
4811 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4816 * Segmentation fault in DTLSv1_listen fix
4818 The DTLSv1_listen function is intended to be stateless and processes the
4819 initial ClientHello from many peers. It is common for user code to loop
4820 over the call to DTLSv1_listen until a valid ClientHello is received with
4821 an associated cookie. A defect in the implementation of DTLSv1_listen means
4822 that state is preserved in the SSL object from one invocation to the next
4823 that can lead to a segmentation fault. Errors processing the initial
4824 ClientHello can trigger this scenario. An example of such an error could be
4825 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4828 This issue was reported to OpenSSL by Per Allansson.
4833 * Segmentation fault in ASN1_TYPE_cmp fix
4835 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4836 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4837 certificate signature algorithm consistency this can be used to crash any
4838 certificate verification operation and exploited in a DoS attack. Any
4839 application which performs certificate verification is vulnerable including
4840 OpenSSL clients and servers which enable client authentication.
4845 * Segmentation fault for invalid PSS parameters fix
4847 The signature verification routines will crash with a NULL pointer
4848 dereference if presented with an ASN.1 signature using the RSA PSS
4849 algorithm and invalid parameters. Since these routines are used to verify
4850 certificate signature algorithms this can be used to crash any
4851 certificate verification operation and exploited in a DoS attack. Any
4852 application which performs certificate verification is vulnerable including
4853 OpenSSL clients and servers which enable client authentication.
4855 This issue was was reported to OpenSSL by Brian Carpenter.
4860 * ASN.1 structure reuse memory corruption fix
4862 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4863 memory corruption via an invalid write. Such reuse is and has been
4864 strongly discouraged and is believed to be rare.
4866 Applications that parse structures containing CHOICE or ANY DEFINED BY
4867 components may be affected. Certificate parsing (d2i_X509 and related
4868 functions) are however not affected. OpenSSL clients and servers are
4874 * PKCS7 NULL pointer dereferences fix
4876 The PKCS#7 parsing code does not handle missing outer ContentInfo
4877 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4878 missing content and trigger a NULL pointer dereference on parsing.
4880 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4881 otherwise parse PKCS#7 structures from untrusted sources are
4882 affected. OpenSSL clients and servers are not affected.
4884 This issue was reported to OpenSSL by Michal Zalewski (Google).
4889 * DoS via reachable assert in SSLv2 servers fix
4891 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4892 servers that both support SSLv2 and enable export cipher suites by sending
4893 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4895 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4896 (OpenSSL development team).
4901 * Empty CKE with client auth and DHE fix
4903 If client auth is used then a server can seg fault in the event of a DHE
4904 ciphersuite being selected and a zero length ClientKeyExchange message
4905 being sent by the client. This could be exploited in a DoS attack.
4910 * Handshake with unseeded PRNG fix
4912 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4913 with an unseeded PRNG. The conditions are:
4914 - The client is on a platform where the PRNG has not been seeded
4915 automatically, and the user has not seeded manually
4916 - A protocol specific client method version has been used (i.e. not
4917 SSL_client_methodv23)
4918 - A ciphersuite is used that does not require additional random data from
4919 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4921 If the handshake succeeds then the client random that has been used will
4922 have been generated from a PRNG with insufficient entropy and therefore the
4923 output may be predictable.
4925 For example using the following command with an unseeded openssl will
4926 succeed on an unpatched platform:
4928 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4933 * Use After Free following d2i_ECPrivatekey error fix
4935 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4936 could cause a use after free condition. This, in turn, could cause a double
4937 free in several private key parsing functions (such as d2i_PrivateKey
4938 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4939 for applications that receive EC private keys from untrusted
4940 sources. This scenario is considered rare.
4942 This issue was discovered by the BoringSSL project and fixed in their
4948 * X509_to_X509_REQ NULL pointer deref fix
4950 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4951 the certificate key is invalid. This function is rarely used in practice.
4953 This issue was discovered by Brian Carpenter.
4958 * Removed the export ciphers from the DEFAULT ciphers
4962 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4964 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4965 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4966 So far those who have to target multiple platforms would compromise
4967 and argue that binary targeting say ARMv5 would still execute on
4968 ARMv8. "Universal" build resolves this compromise by providing
4969 near-optimal performance even on newer platforms.
4973 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4974 (other platforms pending).
4976 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4978 * Add support for the SignedCertificateTimestampList certificate and
4979 OCSP response extensions from RFC6962.
4983 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4984 for corner cases. (Certain input points at infinity could lead to
4985 bogus results, with non-infinity inputs mapped to infinity too.)
4989 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4990 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4991 common cases are optimized and there still is room for further
4992 improvements. Vector Permutation AES for Altivec is also added.
4996 * Add support for little-endian ppc64 Linux target.
4998 *Marcelo Cerri (IBM)*
5000 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5001 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5002 are optimized and there still is room for further improvements.
5003 Both 32- and 64-bit modes are supported.
5005 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5007 * Improved ARMv7 NEON support.
5011 * Support for SPARC Architecture 2011 crypto extensions, first
5012 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5013 SHA256/512, MD5, GHASH and modular exponentiation.
5015 *Andy Polyakov, David Miller*
5017 * Accelerated modular exponentiation for Intel processors, a.k.a.
5020 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5022 * Support for new and upcoming Intel processors, including AVX2,
5023 BMI and SHA ISA extensions. This includes additional "stitched"
5024 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5027 This work was sponsored by Intel Corp.
5031 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5032 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5033 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5037 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5038 this fixes a limitation in previous versions of OpenSSL.
5042 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5043 MGF1 digest and OAEP label.
5047 * Add EVP support for key wrapping algorithms, to avoid problems with
5048 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5049 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5050 algorithms and include tests cases.
5054 * Add functions to allocate and set the fields of an ECDSA_METHOD
5057 *Douglas E. Engert, Steve Henson*
5059 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5060 difference in days and seconds between two tm or ASN1_TIME structures.
5064 * Add -rev test option to s_server to just reverse order of characters
5065 received by client and send back to server. Also prints an abbreviated
5066 summary of the connection parameters.
5070 * New option -brief for s_client and s_server to print out a brief summary
5071 of connection parameters.
5075 * Add callbacks for arbitrary TLS extensions.
5077 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5079 * New option -crl_download in several openssl utilities to download CRLs
5080 from CRLDP extension in certificates.
5084 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5088 * New function X509_CRL_diff to generate a delta CRL from the difference
5089 of two full CRLs. Add support to "crl" utility.
5093 * New functions to set lookup_crls function and to retrieve
5094 X509_STORE from X509_STORE_CTX.
5098 * Print out deprecated issuer and subject unique ID fields in
5103 * Extend OCSP I/O functions so they can be used for simple general purpose
5104 HTTP as well as OCSP. New wrapper function which can be used to download
5105 CRLs using the OCSP API.
5109 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5113 * `SSL_CONF*` functions. These provide a common framework for application
5114 configuration using configuration files or command lines.
5118 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5119 message callback and prints the results. Needs compile time option
5120 "enable-ssl-trace". New options to s_client and s_server to enable
5125 * New ctrl and macro to retrieve supported points extensions.
5126 Print out extension in s_server and s_client.
5130 * New functions to retrieve certificate signature and signature
5135 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5140 * New Suite B modes for TLS code. These use and enforce the requirements
5141 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5142 only use Suite B curves. The Suite B modes can be set by using the
5143 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5147 * New chain verification flags for Suite B levels of security. Check
5148 algorithms are acceptable when flags are set in X509_verify_cert.
5152 * Make tls1_check_chain return a set of flags indicating checks passed
5153 by a certificate chain. Add additional tests to handle client
5154 certificates: checks for matching certificate type and issuer name
5159 * If an attempt is made to use a signature algorithm not in the peer
5160 preference list abort the handshake. If client has no suitable
5161 signature algorithms in response to a certificate request do not
5162 use the certificate.
5166 * If server EC tmp key is not in client preference list abort handshake.
5170 * Add support for certificate stores in CERT structure. This makes it
5171 possible to have different stores per SSL structure or one store in
5172 the parent SSL_CTX. Include distinct stores for certificate chain
5173 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5174 to build and store a certificate chain in CERT structure: returning
5175 an error if the chain cannot be built: this will allow applications
5176 to test if a chain is correctly configured.
5178 Note: if the CERT based stores are not set then the parent SSL_CTX
5179 store is used to retain compatibility with existing behaviour.
5183 * New function ssl_set_client_disabled to set a ciphersuite disabled
5184 mask based on the current session, check mask when sending client
5185 hello and checking the requested ciphersuite.
5189 * New ctrls to retrieve and set certificate types in a certificate
5190 request message. Print out received values in s_client. If certificate
5191 types is not set with custom values set sensible values based on
5192 supported signature algorithms.
5196 * Support for distinct client and server supported signature algorithms.
5200 * Add certificate callback. If set this is called whenever a certificate
5201 is required by client or server. An application can decide which
5202 certificate chain to present based on arbitrary criteria: for example
5203 supported signature algorithms. Add very simple example to s_server.
5204 This fixes many of the problems and restrictions of the existing client
5205 certificate callback: for example you can now clear an existing
5206 certificate and specify the whole chain.
5210 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5211 the certificate can be used for (if anything). Set valid_flags field
5212 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5213 to have similar checks in it.
5215 Add new "cert_flags" field to CERT structure and include a "strict mode".
5216 This enforces some TLS certificate requirements (such as only permitting
5217 certificate signature algorithms contained in the supported algorithms
5218 extension) which some implementations ignore: this option should be used
5219 with caution as it could cause interoperability issues.
5223 * Update and tidy signature algorithm extension processing. Work out
5224 shared signature algorithms based on preferences and peer algorithms
5225 and print them out in s_client and s_server. Abort handshake if no
5226 shared signature algorithms.
5230 * Add new functions to allow customised supported signature algorithms
5231 for SSL and SSL_CTX structures. Add options to s_client and s_server
5236 * New function SSL_certs_clear() to delete all references to certificates
5237 from an SSL structure. Before this once a certificate had been added
5238 it couldn't be removed.
5242 * Integrate hostname, email address and IP address checking with certificate
5243 verification. New verify options supporting checking in openssl utility.
5247 * Fixes and wildcard matching support to hostname and email checking
5248 functions. Add manual page.
5250 *Florian Weimer (Red Hat Product Security Team)*
5252 * New functions to check a hostname email or IP address against a
5253 certificate. Add options x509 utility to print results of checks against
5258 * Fix OCSP checking.
5260 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5262 * Initial experimental support for explicitly trusted non-root CAs.
5263 OpenSSL still tries to build a complete chain to a root but if an
5264 intermediate CA has a trust setting included that is used. The first
5265 setting is used: whether to trust (e.g., -addtrust option to the x509
5270 * Add -trusted_first option which attempts to find certificates in the
5271 trusted store even if an untrusted chain is also supplied.
5275 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5276 platform support for Linux and Android.
5280 * Support for linux-x32, ILP32 environment in x86_64 framework.
5284 * Experimental multi-implementation support for FIPS capable OpenSSL.
5285 When in FIPS mode the approved implementations are used as normal,
5286 when not in FIPS mode the internal unapproved versions are used instead.
5287 This means that the FIPS capable OpenSSL isn't forced to use the
5288 (often lower performance) FIPS implementations outside FIPS mode.
5292 * Transparently support X9.42 DH parameters when calling
5293 PEM_read_bio_DHparameters. This means existing applications can handle
5294 the new parameter format automatically.
5298 * Initial experimental support for X9.42 DH parameter format: mainly
5299 to support use of 'q' parameter for RFC5114 parameters.
5303 * Add DH parameters from RFC5114 including test data to dhtest.
5307 * Support for automatic EC temporary key parameter selection. If enabled
5308 the most preferred EC parameters are automatically used instead of
5309 hardcoded fixed parameters. Now a server just has to call:
5310 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5311 support ECDH and use the most appropriate parameters.
5315 * Enhance and tidy EC curve and point format TLS extension code. Use
5316 static structures instead of allocation if default values are used.
5317 New ctrls to set curves we wish to support and to retrieve shared curves.
5318 Print out shared curves in s_server. New options to s_server and s_client
5319 to set list of supported curves.
5323 * New ctrls to retrieve supported signature algorithms and
5324 supported curve values as an array of NIDs. Extend openssl utility
5325 to print out received values.
5329 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5330 between NIDs and the more common NIST names such as "P-256". Enhance
5331 ecparam utility and ECC method to recognise the NIST names for curves.
5335 * Enhance SSL/TLS certificate chain handling to support different
5336 chains for each certificate instead of one chain in the parent SSL_CTX.
5340 * Support for fixed DH ciphersuite client authentication: where both
5341 server and client use DH certificates with common parameters.
5345 * Support for fixed DH ciphersuites: those requiring DH server
5350 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5352 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5353 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5354 X509_CINF_get_signature were reverted post internal team review.
5359 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5361 * OCSP Status Request extension unbounded memory growth
5363 A malicious client can send an excessively large OCSP Status Request
5364 extension. If that client continually requests renegotiation, sending a
5365 large OCSP Status Request extension each time, then there will be unbounded
5366 memory growth on the server. This will eventually lead to a Denial Of
5367 Service attack through memory exhaustion. Servers with a default
5368 configuration are vulnerable even if they do not support OCSP. Builds using
5369 the "no-ocsp" build time option are not affected.
5371 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5376 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5379 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5385 * OOB write in MDC2_Update()
5387 An overflow can occur in MDC2_Update() either if called directly or
5388 through the EVP_DigestUpdate() function using MDC2. If an attacker
5389 is able to supply very large amounts of input data after a previous
5390 call to EVP_EncryptUpdate() with a partial block then a length check
5391 can overflow resulting in a heap corruption.
5393 The amount of data needed is comparable to SIZE_MAX which is impractical
5396 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5401 * Malformed SHA512 ticket DoS
5403 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5404 DoS attack where a malformed ticket will result in an OOB read which will
5407 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5408 a custom server callback and ticket lookup mechanism.
5410 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5415 * OOB write in BN_bn2dec()
5417 The function BN_bn2dec() does not check the return value of BN_div_word().
5418 This can cause an OOB write if an application uses this function with an
5419 overly large BIGNUM. This could be a problem if an overly large certificate
5420 or CRL is printed out from an untrusted source. TLS is not affected because
5421 record limits will reject an oversized certificate before it is parsed.
5423 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5428 * OOB read in TS_OBJ_print_bio()
5430 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5431 the total length the OID text representation would use and not the amount
5432 of data written. This will result in OOB reads when large OIDs are
5435 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5440 * Pointer arithmetic undefined behaviour
5442 Avoid some undefined pointer arithmetic
5444 A common idiom in the codebase is to check limits in the following manner:
5447 Where "p" points to some malloc'd data of SIZE bytes and
5450 "len" here could be from some externally supplied data (e.g. from a TLS
5453 The rules of C pointer arithmetic are such that "p + len" is only well
5454 defined where len <= SIZE. Therefore the above idiom is actually
5455 undefined behaviour.
5457 For example this could cause problems if some malloc implementation
5458 provides an address for "p" such that "p + len" actually overflows for
5459 values of len that are too big and therefore p + len < limit.
5461 This issue was reported to OpenSSL by Guido Vranken
5466 * Constant time flag not preserved in DSA signing
5468 Operations in the DSA signing algorithm should run in constant time in
5469 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5470 implementation means that a non-constant time codepath is followed for
5471 certain operations. This has been demonstrated through a cache-timing
5472 attack to be sufficient for an attacker to recover the private DSA key.
5474 This issue was reported by César Pereida (Aalto University), Billy Brumley
5475 (Tampere University of Technology), and Yuval Yarom (The University of
5476 Adelaide and NICTA).
5481 * DTLS buffered message DoS
5483 In a DTLS connection where handshake messages are delivered out-of-order
5484 those messages that OpenSSL is not yet ready to process will be buffered
5485 for later use. Under certain circumstances, a flaw in the logic means that
5486 those messages do not get removed from the buffer even though the handshake
5487 has been completed. An attacker could force up to approx. 15 messages to
5488 remain in the buffer when they are no longer required. These messages will
5489 be cleared when the DTLS connection is closed. The default maximum size for
5490 a message is 100k. Therefore the attacker could force an additional 1500k
5491 to be consumed per connection. By opening many simulataneous connections an
5492 attacker could cause a DoS attack through memory exhaustion.
5494 This issue was reported to OpenSSL by Quan Luo.
5499 * DTLS replay protection DoS
5501 A flaw in the DTLS replay attack protection mechanism means that records
5502 that arrive for future epochs update the replay protection "window" before
5503 the MAC for the record has been validated. This could be exploited by an
5504 attacker by sending a record for the next epoch (which does not have to
5505 decrypt or have a valid MAC), with a very large sequence number. This means
5506 that all subsequent legitimate packets are dropped causing a denial of
5507 service for a specific DTLS connection.
5509 This issue was reported to OpenSSL by the OCAP audit team.
5514 * Certificate message OOB reads
5516 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5517 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5518 theoretical DoS risk but this has not been observed in practice on common
5521 The messages affected are client certificate, client certificate request
5522 and server certificate. As a result the attack can only be performed
5523 against a client or a server which enables client authentication.
5525 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5530 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5532 * Prevent padding oracle in AES-NI CBC MAC check
5534 A MITM attacker can use a padding oracle attack to decrypt traffic
5535 when the connection uses an AES CBC cipher and the server support
5538 This issue was introduced as part of the fix for Lucky 13 padding
5539 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5540 constant time by making sure that always the same bytes are read and
5541 compared against either the MAC or padding bytes. But it no longer
5542 checked that there was enough data to have both the MAC and padding
5545 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5550 * Fix EVP_EncodeUpdate overflow
5552 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5553 Base64 encoding of binary data. If an attacker is able to supply very large
5554 amounts of input data then a length check can overflow resulting in a heap
5557 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5558 the `PEM_write_bio*` family of functions. These are mainly used within the
5559 OpenSSL command line applications, so any application which processes data
5560 from an untrusted source and outputs it as a PEM file should be considered
5561 vulnerable to this issue. User applications that call these APIs directly
5562 with large amounts of untrusted data may also be vulnerable.
5564 This issue was reported by Guido Vranken.
5569 * Fix EVP_EncryptUpdate overflow
5571 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5572 is able to supply very large amounts of input data after a previous call to
5573 EVP_EncryptUpdate() with a partial block then a length check can overflow
5574 resulting in a heap corruption. Following an analysis of all OpenSSL
5575 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5576 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5577 the first called function after an EVP_EncryptInit(), and therefore that
5578 specific call must be safe. The second form is where the length passed to
5579 EVP_EncryptUpdate() can be seen from the code to be some small value and
5580 therefore there is no possibility of an overflow. Since all instances are
5581 one of these two forms, it is believed that there can be no overflows in
5582 internal code due to this problem. It should be noted that
5583 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5584 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5585 of these calls have also been analysed too and it is believed there are no
5586 instances in internal usage where an overflow could occur.
5588 This issue was reported by Guido Vranken.
5593 * Prevent ASN.1 BIO excessive memory allocation
5595 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5596 a short invalid encoding can casuse allocation of large amounts of memory
5597 potentially consuming excessive resources or exhausting memory.
5599 Any application parsing untrusted data through d2i BIO functions is
5600 affected. The memory based functions such as d2i_X509() are *not* affected.
5601 Since the memory based functions are used by the TLS library, TLS
5602 applications are not affected.
5604 This issue was reported by Brian Carpenter.
5611 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5612 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5613 in arbitrary stack data being returned in the buffer.
5615 This issue was reported by Guido Vranken.
5620 * Modify behavior of ALPN to invoke callback after SNI/servername
5621 callback, such that updates to the SSL_CTX affect ALPN.
5625 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5630 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5631 methods are enabled and ssl2 is disabled the methods return NULL.
5635 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5637 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5638 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5639 provide any "EXPORT" or "LOW" strength ciphers.
5643 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5644 is by default disabled at build-time. Builds that are not configured with
5645 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5646 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5647 will need to explicitly call either of:
5649 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5651 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5653 as appropriate. Even if either of those is used, or the application
5654 explicitly uses the version-specific SSLv2_method() or its client and
5655 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5656 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5657 ciphers, and SSLv2 56-bit DES are no longer available.
5662 * Fix a double-free in DSA code
5664 A double free bug was discovered when OpenSSL parses malformed DSA private
5665 keys and could lead to a DoS attack or memory corruption for applications
5666 that receive DSA private keys from untrusted sources. This scenario is
5669 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5675 * Disable SRP fake user seed to address a server memory leak.
5677 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5679 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5680 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5681 was changed to ignore the "fake user" SRP seed, even if the seed
5684 Users should use SRP_VBASE_get1_by_user instead. Note that in
5685 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5686 also that even though configuring the SRP seed attempts to hide
5687 invalid usernames by continuing the handshake with fake
5688 credentials, this behaviour is not constant time and no strong
5689 guarantees are made that the handshake is indistinguishable from
5690 that of a valid user.
5695 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5697 In the BN_hex2bn function the number of hex digits is calculated using an
5698 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5699 large values of `i` this can result in `bn_expand` not allocating any
5700 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5701 field as NULL leading to a subsequent NULL ptr deref. For very large values
5702 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5703 In this case memory is allocated to the internal BIGNUM data field, but it
5704 is insufficiently sized leading to heap corruption. A similar issue exists
5705 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5706 is ever called by user applications with very large untrusted hex/dec data.
5707 This is anticipated to be a rare occurrence.
5709 All OpenSSL internal usage of these functions use data that is not expected
5710 to be untrusted, e.g. config file data or application command line
5711 arguments. If user developed applications generate config file data based
5712 on untrusted data then it is possible that this could also lead to security
5713 consequences. This is also anticipated to be rare.
5715 This issue was reported to OpenSSL by Guido Vranken.
5720 * Fix memory issues in `BIO_*printf` functions
5722 The internal `fmtstr` function used in processing a "%s" format string in
5723 the `BIO_*printf` functions could overflow while calculating the length of a
5724 string and cause an OOB read when printing very long strings.
5726 Additionally the internal `doapr_outch` function can attempt to write to an
5727 OOB memory location (at an offset from the NULL pointer) in the event of a
5728 memory allocation failure. In 1.0.2 and below this could be caused where
5729 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5730 could be in processing a very long "%s" format string. Memory leaks can
5733 The first issue may mask the second issue dependent on compiler behaviour.
5734 These problems could enable attacks where large amounts of untrusted data
5735 is passed to the `BIO_*printf` functions. If applications use these functions
5736 in this way then they could be vulnerable. OpenSSL itself uses these
5737 functions when printing out human-readable dumps of ASN.1 data. Therefore
5738 applications that print this data could be vulnerable if the data is from
5739 untrusted sources. OpenSSL command line applications could also be
5740 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5741 as command line arguments.
5743 Libssl is not considered directly vulnerable. Additionally certificates etc
5744 received via remote connections via libssl are also unlikely to be able to
5745 trigger these issues because of message size limits enforced within libssl.
5747 This issue was reported to OpenSSL Guido Vranken.
5752 * Side channel attack on modular exponentiation
5754 A side-channel attack was found which makes use of cache-bank conflicts on
5755 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5756 of RSA keys. The ability to exploit this issue is limited as it relies on
5757 an attacker who has control of code in a thread running on the same
5758 hyper-threaded core as the victim thread which is performing decryptions.
5760 This issue was reported to OpenSSL by Yuval Yarom, The University of
5761 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5762 Nadia Heninger, University of Pennsylvania with more information at
5763 <http://cachebleed.info>.
5768 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5769 if no keysize is specified with default_bits. This fixes an
5770 omission in an earlier change that changed all RSA/DSA key generation
5771 apps to use 2048 bits by default.
5775 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5777 * Protection for DH small subgroup attacks
5779 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5780 switched on by default and cannot be disabled. This could have some
5785 * SSLv2 doesn't block disabled ciphers
5787 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5788 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5789 been disabled, provided that the SSLv2 protocol was not also disabled via
5792 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5793 and Sebastian Schinzel.
5798 * Reject DH handshakes with parameters shorter than 1024 bits.
5802 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5804 * Certificate verify crash with missing PSS parameter
5806 The signature verification routines will crash with a NULL pointer
5807 dereference if presented with an ASN.1 signature using the RSA PSS
5808 algorithm and absent mask generation function parameter. Since these
5809 routines are used to verify certificate signature algorithms this can be
5810 used to crash any certificate verification operation and exploited in a
5811 DoS attack. Any application which performs certificate verification is
5812 vulnerable including OpenSSL clients and servers which enable client
5815 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5820 * X509_ATTRIBUTE memory leak
5822 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5823 memory. This structure is used by the PKCS#7 and CMS routines so any
5824 application which reads PKCS#7 or CMS data from untrusted sources is
5825 affected. SSL/TLS is not affected.
5827 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5833 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5834 This changes the decoding behaviour for some invalid messages,
5835 though the change is mostly in the more lenient direction, and
5836 legacy behaviour is preserved as much as possible.
5840 * In DSA_generate_parameters_ex, if the provided seed is too short,
5841 use a random seed, as already documented.
5843 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5845 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5847 * Alternate chains certificate forgery
5849 During certificate verfification, OpenSSL will attempt to find an
5850 alternative certificate chain if the first attempt to build such a chain
5851 fails. An error in the implementation of this logic can mean that an
5852 attacker could cause certain checks on untrusted certificates to be
5853 bypassed, such as the CA flag, enabling them to use a valid leaf
5854 certificate to act as a CA and "issue" an invalid certificate.
5856 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5862 * Race condition handling PSK identify hint
5864 If PSK identity hints are received by a multi-threaded client then
5865 the values are wrongly updated in the parent SSL_CTX structure. This can
5866 result in a race condition potentially leading to a double free of the
5872 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5874 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5875 incompatibility in the handling of HMAC. The previous ABI has now been
5878 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5880 * Malformed ECParameters causes infinite loop
5882 When processing an ECParameters structure OpenSSL enters an infinite loop
5883 if the curve specified is over a specially malformed binary polynomial
5886 This can be used to perform denial of service against any
5887 system which processes public keys, certificate requests or
5888 certificates. This includes TLS clients and TLS servers with
5889 client authentication enabled.
5891 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5896 * Exploitable out-of-bounds read in X509_cmp_time
5898 X509_cmp_time does not properly check the length of the ASN1_TIME
5899 string and can read a few bytes out of bounds. In addition,
5900 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5903 An attacker can use this to craft malformed certificates and CRLs of
5904 various sizes and potentially cause a segmentation fault, resulting in
5905 a DoS on applications that verify certificates or CRLs. TLS clients
5906 that verify CRLs are affected. TLS clients and servers with client
5907 authentication enabled may be affected if they use custom verification
5910 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5911 independently by Hanno Böck.
5916 * PKCS7 crash with missing EnvelopedContent
5918 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5919 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5920 with missing content and trigger a NULL pointer dereference on parsing.
5922 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5923 structures from untrusted sources are affected. OpenSSL clients and
5924 servers are not affected.
5926 This issue was reported to OpenSSL by Michal Zalewski (Google).
5931 * CMS verify infinite loop with unknown hash function
5933 When verifying a signedData message the CMS code can enter an infinite loop
5934 if presented with an unknown hash function OID. This can be used to perform
5935 denial of service against any system which verifies signedData messages using
5937 This issue was reported to OpenSSL by Johannes Bauer.
5942 * Race condition handling NewSessionTicket
5944 If a NewSessionTicket is received by a multi-threaded client when attempting to
5945 reuse a previous ticket then a race condition can occur potentially leading to
5946 a double free of the ticket data.
5951 * Reject DH handshakes with parameters shorter than 768 bits.
5953 *Kurt Roeckx and Emilia Kasper*
5955 * dhparam: generate 2048-bit parameters by default.
5957 *Kurt Roeckx and Emilia Kasper*
5959 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5961 * Segmentation fault in ASN1_TYPE_cmp fix
5963 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5964 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5965 certificate signature algorithm consistency this can be used to crash any
5966 certificate verification operation and exploited in a DoS attack. Any
5967 application which performs certificate verification is vulnerable including
5968 OpenSSL clients and servers which enable client authentication.
5973 * ASN.1 structure reuse memory corruption fix
5975 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5976 memory corruption via an invalid write. Such reuse is and has been
5977 strongly discouraged and is believed to be rare.
5979 Applications that parse structures containing CHOICE or ANY DEFINED BY
5980 components may be affected. Certificate parsing (d2i_X509 and related
5981 functions) are however not affected. OpenSSL clients and servers are
5987 * PKCS7 NULL pointer dereferences fix
5989 The PKCS#7 parsing code does not handle missing outer ContentInfo
5990 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5991 missing content and trigger a NULL pointer dereference on parsing.
5993 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5994 otherwise parse PKCS#7 structures from untrusted sources are
5995 affected. OpenSSL clients and servers are not affected.
5997 This issue was reported to OpenSSL by Michal Zalewski (Google).
6002 * DoS via reachable assert in SSLv2 servers fix
6004 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6005 servers that both support SSLv2 and enable export cipher suites by sending
6006 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6008 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6009 (OpenSSL development team).
6014 * Use After Free following d2i_ECPrivatekey error fix
6016 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6017 could cause a use after free condition. This, in turn, could cause a double
6018 free in several private key parsing functions (such as d2i_PrivateKey
6019 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6020 for applications that receive EC private keys from untrusted
6021 sources. This scenario is considered rare.
6023 This issue was discovered by the BoringSSL project and fixed in their
6029 * X509_to_X509_REQ NULL pointer deref fix
6031 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6032 the certificate key is invalid. This function is rarely used in practice.
6034 This issue was discovered by Brian Carpenter.
6039 * Removed the export ciphers from the DEFAULT ciphers
6043 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6045 * Build fixes for the Windows and OpenVMS platforms
6047 *Matt Caswell and Richard Levitte*
6049 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6051 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6052 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6053 dereference. This could lead to a Denial Of Service attack. Thanks to
6054 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6059 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6060 dtls1_buffer_record function under certain conditions. In particular this
6061 could occur if an attacker sent repeated DTLS records with the same
6062 sequence number but for the next epoch. The memory leak could be exploited
6063 by an attacker in a Denial of Service attack through memory exhaustion.
6064 Thanks to Chris Mueller for reporting this issue.
6069 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6070 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6071 method would be set to NULL which could later result in a NULL pointer
6072 dereference. Thanks to Frank Schmirler for reporting this issue.
6077 * Abort handshake if server key exchange message is omitted for ephemeral
6080 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6081 reporting this issue.
6086 * Remove non-export ephemeral RSA code on client and server. This code
6087 violated the TLS standard by allowing the use of temporary RSA keys in
6088 non-export ciphersuites and could be used by a server to effectively
6089 downgrade the RSA key length used to a value smaller than the server
6090 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6091 INRIA or reporting this issue.
6096 * Fixed issue where DH client certificates are accepted without verification.
6097 An OpenSSL server will accept a DH certificate for client authentication
6098 without the certificate verify message. This effectively allows a client to
6099 authenticate without the use of a private key. This only affects servers
6100 which trust a client certificate authority which issues certificates
6101 containing DH keys: these are extremely rare and hardly ever encountered.
6102 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6108 * Ensure that the session ID context of an SSL is updated when its
6109 SSL_CTX is updated via SSL_set_SSL_CTX.
6111 The session ID context is typically set from the parent SSL_CTX,
6112 and can vary with the CTX.
6116 * Fix various certificate fingerprint issues.
6118 By using non-DER or invalid encodings outside the signed portion of a
6119 certificate the fingerprint can be changed without breaking the signature.
6120 Although no details of the signed portion of the certificate can be changed
6121 this can cause problems with some applications: e.g. those using the
6122 certificate fingerprint for blacklists.
6124 1. Reject signatures with non zero unused bits.
6126 If the BIT STRING containing the signature has non zero unused bits reject
6127 the signature. All current signature algorithms require zero unused bits.
6129 2. Check certificate algorithm consistency.
6131 Check the AlgorithmIdentifier inside TBS matches the one in the
6132 certificate signature. NB: this will result in signature failure
6133 errors for some broken certificates.
6135 Thanks to Konrad Kraszewski from Google for reporting this issue.
6137 3. Check DSA/ECDSA signatures use DER.
6139 Re-encode DSA/ECDSA signatures and compare with the original received
6140 signature. Return an error if there is a mismatch.
6142 This will reject various cases including garbage after signature
6143 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6144 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6145 (negative or with leading zeroes).
6147 Further analysis was conducted and fixes were developed by Stephen Henson
6148 of the OpenSSL core team.
6154 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6155 results on some platforms, including x86_64. This bug occurs at random
6156 with a very low probability, and is not known to be exploitable in any
6157 way, though its exact impact is difficult to determine. Thanks to Pieter
6158 Wuille (Blockstream) who reported this issue and also suggested an initial
6159 fix. Further analysis was conducted by the OpenSSL development team and
6160 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6161 the OpenSSL core team.
6166 * Do not resume sessions on the server if the negotiated protocol
6167 version does not match the session's version. Resuming with a different
6168 version, while not strictly forbidden by the RFC, is of questionable
6169 sanity and breaks all known clients.
6171 *David Benjamin, Emilia Käsper*
6173 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6174 early CCS messages during renegotiation. (Note that because
6175 renegotiation is encrypted, this early CCS was not exploitable.)
6179 * Tighten client-side session ticket handling during renegotiation:
6180 ensure that the client only accepts a session ticket if the server sends
6181 the extension anew in the ServerHello. Previously, a TLS client would
6182 reuse the old extension state and thus accept a session ticket if one was
6183 announced in the initial ServerHello.
6185 Similarly, ensure that the client requires a session ticket if one
6186 was advertised in the ServerHello. Previously, a TLS client would
6187 ignore a missing NewSessionTicket message.
6191 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6195 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6196 sends a carefully crafted handshake message, to cause OpenSSL to fail
6197 to free up to 64k of memory causing a memory leak. This could be
6198 exploited in a Denial Of Service attack. This issue affects OpenSSL
6199 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6200 whether SRTP is used or configured. Implementations of OpenSSL that
6201 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6203 The fix was developed by the OpenSSL team.
6208 * Session Ticket Memory Leak.
6210 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6211 integrity of that ticket is first verified. In the event of a session
6212 ticket integrity check failing, OpenSSL will fail to free memory
6213 causing a memory leak. By sending a large number of invalid session
6214 tickets an attacker could exploit this issue in a Denial Of Service
6220 * Build option no-ssl3 is incomplete.
6222 When OpenSSL is configured with "no-ssl3" as a build option, servers
6223 could accept and complete a SSL 3.0 handshake, and clients could be
6224 configured to send them.
6227 *Akamai and the OpenSSL team*
6229 * Add support for TLS_FALLBACK_SCSV.
6230 Client applications doing fallback retries should call
6231 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6234 *Adam Langley, Bodo Moeller*
6236 * Add additional DigestInfo checks.
6238 Re-encode DigestInto in DER and check against the original when
6239 verifying RSA signature: this will reject any improperly encoded
6240 DigestInfo structures.
6242 Note: this is a precautionary measure and no attacks are currently known.
6246 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6248 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6249 SRP code can be overrun an internal buffer. Add sanity check that
6250 g, A, B < N to SRP code.
6252 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6253 Group for discovering this issue.
6258 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6259 TLS 1.0 instead of higher protocol versions when the ClientHello message
6260 is badly fragmented. This allows a man-in-the-middle attacker to force a
6261 downgrade to TLS 1.0 even if both the server and the client support a
6262 higher protocol version, by modifying the client's TLS records.
6264 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6265 researching this issue.
6270 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6271 to a denial of service attack. A malicious server can crash the client
6272 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6273 ciphersuite and sending carefully crafted handshake messages.
6275 Thanks to Felix Gröbert (Google) for discovering and researching this
6281 * By sending carefully crafted DTLS packets an attacker could cause openssl
6282 to leak memory. This can be exploited through a Denial of Service attack.
6283 Thanks to Adam Langley for discovering and researching this issue.
6288 * An attacker can force openssl to consume large amounts of memory whilst
6289 processing DTLS handshake messages. This can be exploited through a
6290 Denial of Service attack.
6291 Thanks to Adam Langley for discovering and researching this issue.
6296 * An attacker can force an error condition which causes openssl to crash
6297 whilst processing DTLS packets due to memory being freed twice. This
6298 can be exploited through a Denial of Service attack.
6299 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6305 * If a multithreaded client connects to a malicious server using a resumed
6306 session and the server sends an ec point format extension it could write
6307 up to 255 bytes to freed memory.
6309 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6315 * A malicious server can crash an OpenSSL client with a null pointer
6316 dereference (read) by specifying an SRP ciphersuite even though it was not
6317 properly negotiated with the client. This can be exploited through a
6318 Denial of Service attack.
6320 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6321 discovering and researching this issue.
6326 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6327 X509_name_oneline, X509_name_print_ex et al. to leak some information
6328 from the stack. Applications may be affected if they echo pretty printing
6329 output to the attacker.
6331 Thanks to Ivan Fratric (Google) for discovering this issue.
6334 *Emilia Käsper, and Steve Henson*
6336 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6337 for corner cases. (Certain input points at infinity could lead to
6338 bogus results, with non-infinity inputs mapped to infinity too.)
6342 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6344 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6345 handshake can force the use of weak keying material in OpenSSL
6346 SSL/TLS clients and servers.
6348 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6349 researching this issue. [CVE-2014-0224][]
6351 *KIKUCHI Masashi, Steve Henson*
6353 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6354 OpenSSL DTLS client the code can be made to recurse eventually crashing
6357 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6360 *Imre Rad, Steve Henson*
6362 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6363 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6364 client or server. This is potentially exploitable to run arbitrary
6365 code on a vulnerable client or server.
6367 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6369 *Jüri Aedla, Steve Henson*
6371 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6372 are subject to a denial of service attack.
6374 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6375 this issue. [CVE-2014-3470][]
6377 *Felix Gröbert, Ivan Fratric, Steve Henson*
6379 * Harmonize version and its documentation. -f flag is used to display
6382 *mancha <mancha1@zoho.com>*
6384 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6385 in i2d_ECPrivateKey.
6387 *mancha <mancha1@zoho.com>*
6389 * Fix some double frees. These are not thought to be exploitable.
6391 *mancha <mancha1@zoho.com>*
6393 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6395 * A missing bounds check in the handling of the TLS heartbeat extension
6396 can be used to reveal up to 64k of memory to a connected client or
6399 Thanks for Neel Mehta of Google Security for discovering this bug and to
6400 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6401 preparing the fix [CVE-2014-0160][]
6403 *Adam Langley, Bodo Moeller*
6405 * Fix for the attack described in the paper "Recovering OpenSSL
6406 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6407 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6408 <http://eprint.iacr.org/2014/140>
6410 Thanks to Yuval Yarom and Naomi Benger for discovering this
6411 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6413 *Yuval Yarom and Naomi Benger*
6415 * TLS pad extension: draft-agl-tls-padding-03
6417 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6418 TLS client Hello record length value would otherwise be > 255 and
6419 less that 512 pad with a dummy extension containing zeroes so it
6420 is at least 512 bytes long.
6422 *Adam Langley, Steve Henson*
6424 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6426 * Fix for TLS record tampering bug. A carefully crafted invalid
6427 handshake could crash OpenSSL with a NULL pointer exception.
6428 Thanks to Anton Johansson for reporting this issues.
6431 * Keep original DTLS digest and encryption contexts in retransmission
6432 structures so we can use the previous session parameters if they need
6433 to be resent. [CVE-2013-6450][]
6437 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6438 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6439 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6440 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6441 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6442 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6444 *Rob Stradling, Adam Langley*
6446 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6448 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6449 supporting platforms or when small records were transferred.
6451 *Andy Polyakov, Steve Henson*
6453 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6455 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6457 This addresses the flaw in CBC record processing discovered by
6458 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6459 at: <http://www.isg.rhul.ac.uk/tls/>
6461 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6462 Security Group at Royal Holloway, University of London
6463 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6464 Emilia Käsper for the initial patch.
6467 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6469 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6470 ciphersuites which can be exploited in a denial of service attack.
6471 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6472 and detecting this bug and to Wolfgang Ettlinger
6473 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6478 * Return an error when checking OCSP signatures when key is NULL.
6479 This fixes a DoS attack. [CVE-2013-0166][]
6483 * Make openssl verify return errors.
6485 *Chris Palmer <palmer@google.com> and Ben Laurie*
6487 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6488 the right response is stapled. Also change SSL_get_certificate()
6489 so it returns the certificate actually sent.
6490 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6492 *Rob Stradling <rob.stradling@comodo.com>*
6494 * Fix possible deadlock when decoding public keys.
6498 * Don't use TLS 1.0 record version number in initial client hello
6503 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6505 * Sanity check record length before skipping explicit IV in TLS
6506 1.2, 1.1 and DTLS to fix DoS attack.
6508 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6509 fuzzing as a service testing platform.
6514 * Initialise tkeylen properly when encrypting CMS messages.
6515 Thanks to Solar Designer of Openwall for reporting this issue.
6519 * In FIPS mode don't try to use composite ciphers as they are not
6524 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6526 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6527 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6528 mean any application compiled against OpenSSL 1.0.0 headers setting
6529 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6530 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6531 0x10000000L Any application which was previously compiled against
6532 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6533 will need to be recompiled as a result. Letting be results in
6534 inability to disable specifically TLS 1.1 and in client context,
6535 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6539 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6540 disable just protocol X, but all protocols above X *if* there are
6541 protocols *below* X still enabled. In more practical terms it means
6542 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6543 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6544 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6549 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6551 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6552 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6553 in CRYPTO_realloc_clean.
6555 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6556 issue and to Adam Langley <agl@chromium.org> for fixing it.
6559 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6561 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6565 * Workarounds for some broken servers that "hang" if a client hello
6566 record length exceeds 255 bytes.
6568 1. Do not use record version number > TLS 1.0 in initial client
6569 hello: some (but not all) hanging servers will now work.
6570 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6571 the number of ciphers sent in the client hello. This should be
6572 set to an even number, such as 50, for example by passing:
6573 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6574 Most broken servers should now work.
6575 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6576 TLS 1.2 client support entirely.
6580 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6584 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6586 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6587 STRING form instead of a DigestInfo.
6591 * The format used for MDC2 RSA signatures is inconsistent between EVP
6592 and the RSA_sign/RSA_verify functions. This was made more apparent when
6593 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6594 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6595 the correct format in RSA_verify so both forms transparently work.
6599 * Some servers which support TLS 1.0 can choke if we initially indicate
6600 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6601 encrypted premaster secret. As a workaround use the maximum permitted
6602 client version in client hello, this should keep such servers happy
6603 and still work with previous versions of OpenSSL.
6607 * Add support for TLS/DTLS heartbeats.
6609 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6611 * Add support for SCTP.
6613 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6615 * Improved PRNG seeding for VOS.
6617 *Paul Green <Paul.Green@stratus.com>*
6619 * Extensive assembler packs updates, most notably:
6621 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6622 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6623 - x86_64: bit-sliced AES implementation;
6624 - ARM: NEON support, contemporary platforms optimizations;
6625 - s390x: z196 support;
6626 - `*`: GHASH and GF(2^m) multiplication implementations;
6630 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6631 (removal of unnecessary code)
6633 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6635 * Add TLS key material exporter from RFC 5705.
6639 * Add DTLS-SRTP negotiation from RFC 5764.
6643 * Add Next Protocol Negotiation,
6644 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6645 disabled with a no-npn flag to config or Configure. Code donated
6648 *Adam Langley <agl@google.com> and Ben Laurie*
6650 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6651 NIST-P256, NIST-P521, with constant-time single point multiplication on
6652 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6653 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6654 Code made available under Apache License version 2.0.
6656 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6657 line to include this in your build of OpenSSL, and run "make depend" (or
6658 "make update"). This enables the following EC_METHODs:
6660 EC_GFp_nistp224_method()
6661 EC_GFp_nistp256_method()
6662 EC_GFp_nistp521_method()
6664 EC_GROUP_new_by_curve_name() will automatically use these (while
6665 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6668 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6670 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6671 all platforms. Move ssize_t definition from e_os.h to the public
6672 header file e_os2.h as it now appears in public header file cms.h
6676 * New -sigopt option to the ca, req and x509 utilities. Additional
6677 signature parameters can be passed using this option and in
6682 * Add RSA PSS signing function. This will generate and set the
6683 appropriate AlgorithmIdentifiers for PSS based on those in the
6684 corresponding EVP_MD_CTX structure. No application support yet.
6688 * Support for companion algorithm specific ASN1 signing routines.
6689 New function ASN1_item_sign_ctx() signs a pre-initialised
6690 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6691 the appropriate parameters.
6695 * Add new algorithm specific ASN1 verification initialisation function
6696 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6697 handling will be the same no matter what EVP_PKEY_METHOD is used.
6698 Add a PSS handler to support verification of PSS signatures: checked
6699 against a number of sample certificates.
6703 * Add signature printing for PSS. Add PSS OIDs.
6705 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6707 * Add algorithm specific signature printing. An individual ASN1 method
6708 can now print out signatures instead of the standard hex dump.
6710 More complex signatures (e.g. PSS) can print out more meaningful
6711 information. Include DSA version that prints out the signature
6716 * Password based recipient info support for CMS library: implementing
6721 * Split password based encryption into PBES2 and PBKDF2 functions. This
6722 neatly separates the code into cipher and PBE sections and is required
6723 for some algorithms that split PBES2 into separate pieces (such as
6724 password based CMS).
6728 * Session-handling fixes:
6729 - Fix handling of connections that are resuming with a session ID,
6730 but also support Session Tickets.
6731 - Fix a bug that suppressed issuing of a new ticket if the client
6732 presented a ticket with an expired session.
6733 - Try to set the ticket lifetime hint to something reasonable.
6734 - Make tickets shorter by excluding irrelevant information.
6735 - On the client side, don't ignore renewed tickets.
6737 *Adam Langley, Bodo Moeller (Google)*
6739 * Fix PSK session representation.
6743 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6745 This work was sponsored by Intel.
6749 * Add GCM support to TLS library. Some custom code is needed to split
6750 the IV between the fixed (from PRF) and explicit (from TLS record)
6751 portions. This adds all GCM ciphersuites supported by RFC5288 and
6752 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6753 add a special AESGCM string for GCM only.
6757 * Expand range of ctrls for AES GCM. Permit setting invocation
6758 field on decrypt and retrieval of invocation field only on encrypt.
6762 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6763 As required by RFC5289 these ciphersuites cannot be used if for
6764 versions of TLS earlier than 1.2.
6768 * For FIPS capable OpenSSL interpret a NULL default public key method
6769 as unset and return the appropriate default but do *not* set the default.
6770 This means we can return the appropriate method in applications that
6771 switch between FIPS and non-FIPS modes.
6775 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6776 ENGINE is used then we cannot handle that in the FIPS module so we
6777 keep original code iff non-FIPS operations are allowed.
6781 * Add -attime option to openssl utilities.
6783 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6785 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6789 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6790 FIPS EC methods unconditionally for now.
6794 * New build option no-ec2m to disable characteristic 2 code.
6798 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6799 all cases can be covered as some introduce binary incompatibilities.
6803 * Redirect RSA operations to FIPS module including keygen,
6804 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6808 * Add similar low level API blocking to ciphers.
6812 * Low level digest APIs are not approved in FIPS mode: any attempt
6813 to use these will cause a fatal error. Applications that *really* want
6814 to use them can use the `private_*` version instead.
6818 * Redirect cipher operations to FIPS module for FIPS builds.
6822 * Redirect digest operations to FIPS module for FIPS builds.
6826 * Update build system to add "fips" flag which will link in fipscanister.o
6827 for static and shared library builds embedding a signature if needed.
6831 * Output TLS supported curves in preference order instead of numerical
6832 order. This is currently hardcoded for the highest order curves first.
6833 This should be configurable so applications can judge speed vs strength.
6837 * Add TLS v1.2 server support for client authentication.
6841 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6846 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6847 FIPS modules versions.
6851 * Add TLS v1.2 client side support for client authentication. Keep cache
6852 of handshake records longer as we don't know the hash algorithm to use
6853 until after the certificate request message is received.
6857 * Initial TLS v1.2 client support. Add a default signature algorithms
6858 extension including all the algorithms we support. Parse new signature
6859 format in client key exchange. Relax some ECC signing restrictions for
6860 TLS v1.2 as indicated in RFC5246.
6864 * Add server support for TLS v1.2 signature algorithms extension. Switch
6865 to new signature format when needed using client digest preference.
6866 All server ciphersuites should now work correctly in TLS v1.2. No client
6867 support yet and no support for client certificates.
6871 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6872 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6873 ciphersuites. At present only RSA key exchange ciphersuites work with
6874 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6875 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6876 and version checking.
6880 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6881 with this defined it will not be affected by any changes to ssl internal
6882 structures. Add several utility functions to allow openssl application
6883 to work with OPENSSL_NO_SSL_INTERN defined.
6887 * A long standing patch to add support for SRP from EdelWeb (Peter
6888 Sylvester and Christophe Renou) was integrated.
6889 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6890 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6893 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6897 * Permit abbreviated handshakes when renegotiating using the function
6898 SSL_renegotiate_abbreviated().
6900 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6902 * Add call to ENGINE_register_all_complete() to
6903 ENGINE_load_builtin_engines(), so some implementations get used
6904 automatically instead of needing explicit application support.
6908 * Add support for TLS key exporter as described in RFC5705.
6910 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6912 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6913 a few changes are required:
6915 Add SSL_OP_NO_TLSv1_1 flag.
6916 Add TLSv1_1 methods.
6917 Update version checking logic to handle version 1.1.
6918 Add explicit IV handling (ported from DTLS code).
6919 Add command line options to s_client/s_server.
6926 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6928 * X509_ATTRIBUTE memory leak
6930 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6931 memory. This structure is used by the PKCS#7 and CMS routines so any
6932 application which reads PKCS#7 or CMS data from untrusted sources is
6933 affected. SSL/TLS is not affected.
6935 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6941 * Race condition handling PSK identify hint
6943 If PSK identity hints are received by a multi-threaded client then
6944 the values are wrongly updated in the parent SSL_CTX structure. This can
6945 result in a race condition potentially leading to a double free of the
6951 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6953 * Malformed ECParameters causes infinite loop
6955 When processing an ECParameters structure OpenSSL enters an infinite loop
6956 if the curve specified is over a specially malformed binary polynomial
6959 This can be used to perform denial of service against any
6960 system which processes public keys, certificate requests or
6961 certificates. This includes TLS clients and TLS servers with
6962 client authentication enabled.
6964 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6969 * Exploitable out-of-bounds read in X509_cmp_time
6971 X509_cmp_time does not properly check the length of the ASN1_TIME
6972 string and can read a few bytes out of bounds. In addition,
6973 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6976 An attacker can use this to craft malformed certificates and CRLs of
6977 various sizes and potentially cause a segmentation fault, resulting in
6978 a DoS on applications that verify certificates or CRLs. TLS clients
6979 that verify CRLs are affected. TLS clients and servers with client
6980 authentication enabled may be affected if they use custom verification
6983 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6984 independently by Hanno Böck.
6989 * PKCS7 crash with missing EnvelopedContent
6991 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6992 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6993 with missing content and trigger a NULL pointer dereference on parsing.
6995 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6996 structures from untrusted sources are affected. OpenSSL clients and
6997 servers are not affected.
6999 This issue was reported to OpenSSL by Michal Zalewski (Google).
7004 * CMS verify infinite loop with unknown hash function
7006 When verifying a signedData message the CMS code can enter an infinite loop
7007 if presented with an unknown hash function OID. This can be used to perform
7008 denial of service against any system which verifies signedData messages using
7010 This issue was reported to OpenSSL by Johannes Bauer.
7015 * Race condition handling NewSessionTicket
7017 If a NewSessionTicket is received by a multi-threaded client when attempting to
7018 reuse a previous ticket then a race condition can occur potentially leading to
7019 a double free of the ticket data.
7024 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7026 * Segmentation fault in ASN1_TYPE_cmp fix
7028 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7029 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7030 certificate signature algorithm consistency this can be used to crash any
7031 certificate verification operation and exploited in a DoS attack. Any
7032 application which performs certificate verification is vulnerable including
7033 OpenSSL clients and servers which enable client authentication.
7038 * ASN.1 structure reuse memory corruption fix
7040 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7041 memory corruption via an invalid write. Such reuse is and has been
7042 strongly discouraged and is believed to be rare.
7044 Applications that parse structures containing CHOICE or ANY DEFINED BY
7045 components may be affected. Certificate parsing (d2i_X509 and related
7046 functions) are however not affected. OpenSSL clients and servers are
7052 * PKCS7 NULL pointer dereferences fix
7054 The PKCS#7 parsing code does not handle missing outer ContentInfo
7055 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7056 missing content and trigger a NULL pointer dereference on parsing.
7058 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7059 otherwise parse PKCS#7 structures from untrusted sources are
7060 affected. OpenSSL clients and servers are not affected.
7062 This issue was reported to OpenSSL by Michal Zalewski (Google).
7067 * DoS via reachable assert in SSLv2 servers fix
7069 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7070 servers that both support SSLv2 and enable export cipher suites by sending
7071 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7073 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7074 (OpenSSL development team).
7079 * Use After Free following d2i_ECPrivatekey error fix
7081 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7082 could cause a use after free condition. This, in turn, could cause a double
7083 free in several private key parsing functions (such as d2i_PrivateKey
7084 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7085 for applications that receive EC private keys from untrusted
7086 sources. This scenario is considered rare.
7088 This issue was discovered by the BoringSSL project and fixed in their
7094 * X509_to_X509_REQ NULL pointer deref fix
7096 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7097 the certificate key is invalid. This function is rarely used in practice.
7099 This issue was discovered by Brian Carpenter.
7104 * Removed the export ciphers from the DEFAULT ciphers
7108 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7110 * Build fixes for the Windows and OpenVMS platforms
7112 *Matt Caswell and Richard Levitte*
7114 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7116 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7117 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7118 dereference. This could lead to a Denial Of Service attack. Thanks to
7119 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7124 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7125 dtls1_buffer_record function under certain conditions. In particular this
7126 could occur if an attacker sent repeated DTLS records with the same
7127 sequence number but for the next epoch. The memory leak could be exploited
7128 by an attacker in a Denial of Service attack through memory exhaustion.
7129 Thanks to Chris Mueller for reporting this issue.
7134 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7135 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7136 method would be set to NULL which could later result in a NULL pointer
7137 dereference. Thanks to Frank Schmirler for reporting this issue.
7142 * Abort handshake if server key exchange message is omitted for ephemeral
7145 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7146 reporting this issue.
7151 * Remove non-export ephemeral RSA code on client and server. This code
7152 violated the TLS standard by allowing the use of temporary RSA keys in
7153 non-export ciphersuites and could be used by a server to effectively
7154 downgrade the RSA key length used to a value smaller than the server
7155 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7156 INRIA or reporting this issue.
7161 * Fixed issue where DH client certificates are accepted without verification.
7162 An OpenSSL server will accept a DH certificate for client authentication
7163 without the certificate verify message. This effectively allows a client to
7164 authenticate without the use of a private key. This only affects servers
7165 which trust a client certificate authority which issues certificates
7166 containing DH keys: these are extremely rare and hardly ever encountered.
7167 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7173 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7174 results on some platforms, including x86_64. This bug occurs at random
7175 with a very low probability, and is not known to be exploitable in any
7176 way, though its exact impact is difficult to determine. Thanks to Pieter
7177 Wuille (Blockstream) who reported this issue and also suggested an initial
7178 fix. Further analysis was conducted by the OpenSSL development team and
7179 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7180 the OpenSSL core team.
7185 * Fix various certificate fingerprint issues.
7187 By using non-DER or invalid encodings outside the signed portion of a
7188 certificate the fingerprint can be changed without breaking the signature.
7189 Although no details of the signed portion of the certificate can be changed
7190 this can cause problems with some applications: e.g. those using the
7191 certificate fingerprint for blacklists.
7193 1. Reject signatures with non zero unused bits.
7195 If the BIT STRING containing the signature has non zero unused bits reject
7196 the signature. All current signature algorithms require zero unused bits.
7198 2. Check certificate algorithm consistency.
7200 Check the AlgorithmIdentifier inside TBS matches the one in the
7201 certificate signature. NB: this will result in signature failure
7202 errors for some broken certificates.
7204 Thanks to Konrad Kraszewski from Google for reporting this issue.
7206 3. Check DSA/ECDSA signatures use DER.
7208 Reencode DSA/ECDSA signatures and compare with the original received
7209 signature. Return an error if there is a mismatch.
7211 This will reject various cases including garbage after signature
7212 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7213 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7214 (negative or with leading zeroes).
7216 Further analysis was conducted and fixes were developed by Stephen Henson
7217 of the OpenSSL core team.
7223 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7225 * Session Ticket Memory Leak.
7227 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7228 integrity of that ticket is first verified. In the event of a session
7229 ticket integrity check failing, OpenSSL will fail to free memory
7230 causing a memory leak. By sending a large number of invalid session
7231 tickets an attacker could exploit this issue in a Denial Of Service
7237 * Build option no-ssl3 is incomplete.
7239 When OpenSSL is configured with "no-ssl3" as a build option, servers
7240 could accept and complete a SSL 3.0 handshake, and clients could be
7241 configured to send them.
7244 *Akamai and the OpenSSL team*
7246 * Add support for TLS_FALLBACK_SCSV.
7247 Client applications doing fallback retries should call
7248 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7251 *Adam Langley, Bodo Moeller*
7253 * Add additional DigestInfo checks.
7255 Reencode DigestInto in DER and check against the original when
7256 verifying RSA signature: this will reject any improperly encoded
7257 DigestInfo structures.
7259 Note: this is a precautionary measure and no attacks are currently known.
7263 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7265 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7266 to a denial of service attack. A malicious server can crash the client
7267 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7268 ciphersuite and sending carefully crafted handshake messages.
7270 Thanks to Felix Gröbert (Google) for discovering and researching this
7276 * By sending carefully crafted DTLS packets an attacker could cause openssl
7277 to leak memory. This can be exploited through a Denial of Service attack.
7278 Thanks to Adam Langley for discovering and researching this issue.
7283 * An attacker can force openssl to consume large amounts of memory whilst
7284 processing DTLS handshake messages. This can be exploited through a
7285 Denial of Service attack.
7286 Thanks to Adam Langley for discovering and researching this issue.
7291 * An attacker can force an error condition which causes openssl to crash
7292 whilst processing DTLS packets due to memory being freed twice. This
7293 can be exploited through a Denial of Service attack.
7294 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7300 * If a multithreaded client connects to a malicious server using a resumed
7301 session and the server sends an ec point format extension it could write
7302 up to 255 bytes to freed memory.
7304 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7310 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7311 X509_name_oneline, X509_name_print_ex et al. to leak some information
7312 from the stack. Applications may be affected if they echo pretty printing
7313 output to the attacker.
7315 Thanks to Ivan Fratric (Google) for discovering this issue.
7318 *Emilia Käsper, and Steve Henson*
7320 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7321 for corner cases. (Certain input points at infinity could lead to
7322 bogus results, with non-infinity inputs mapped to infinity too.)
7326 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7328 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7329 handshake can force the use of weak keying material in OpenSSL
7330 SSL/TLS clients and servers.
7332 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7333 researching this issue. [CVE-2014-0224][]
7335 *KIKUCHI Masashi, Steve Henson*
7337 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7338 OpenSSL DTLS client the code can be made to recurse eventually crashing
7341 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7344 *Imre Rad, Steve Henson*
7346 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7347 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7348 client or server. This is potentially exploitable to run arbitrary
7349 code on a vulnerable client or server.
7351 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7353 *Jüri Aedla, Steve Henson*
7355 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7356 are subject to a denial of service attack.
7358 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7359 this issue. [CVE-2014-3470][]
7361 *Felix Gröbert, Ivan Fratric, Steve Henson*
7363 * Harmonize version and its documentation. -f flag is used to display
7366 *mancha <mancha1@zoho.com>*
7368 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7369 in i2d_ECPrivateKey.
7371 *mancha <mancha1@zoho.com>*
7373 * Fix some double frees. These are not thought to be exploitable.
7375 *mancha <mancha1@zoho.com>*
7377 * Fix for the attack described in the paper "Recovering OpenSSL
7378 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7379 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7380 <http://eprint.iacr.org/2014/140>
7382 Thanks to Yuval Yarom and Naomi Benger for discovering this
7383 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7385 *Yuval Yarom and Naomi Benger*
7387 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7389 * Keep original DTLS digest and encryption contexts in retransmission
7390 structures so we can use the previous session parameters if they need
7391 to be resent. [CVE-2013-6450][]
7395 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7396 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7397 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7398 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7399 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7400 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7402 *Rob Stradling, Adam Langley*
7404 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7406 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7408 This addresses the flaw in CBC record processing discovered by
7409 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7410 at: <http://www.isg.rhul.ac.uk/tls/>
7412 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7413 Security Group at Royal Holloway, University of London
7414 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7415 Emilia Käsper for the initial patch.
7418 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7420 * Return an error when checking OCSP signatures when key is NULL.
7421 This fixes a DoS attack. [CVE-2013-0166][]
7425 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7426 the right response is stapled. Also change SSL_get_certificate()
7427 so it returns the certificate actually sent.
7428 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7429 (This is a backport)
7431 *Rob Stradling <rob.stradling@comodo.com>*
7433 * Fix possible deadlock when decoding public keys.
7437 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7439 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7442 * Sanity check record length before skipping explicit IV in DTLS
7445 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7446 fuzzing as a service testing platform.
7451 * Initialise tkeylen properly when encrypting CMS messages.
7452 Thanks to Solar Designer of Openwall for reporting this issue.
7456 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7458 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7459 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7460 in CRYPTO_realloc_clean.
7462 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7463 issue and to Adam Langley <agl@chromium.org> for fixing it.
7466 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7468 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7470 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7471 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7472 content decryption and always return the same error. Note: this attack
7473 needs on average 2^20 messages so it only affects automated senders. The
7474 old behaviour can be re-enabled in the CMS code by setting the
7475 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7476 an MMA defence is not necessary.
7477 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7478 this issue. [CVE-2012-0884][]
7482 * Fix CVE-2011-4619: make sure we really are receiving a
7483 client hello before rejecting multiple SGC restarts. Thanks to
7484 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7488 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7490 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7491 Thanks to Antonio Martin, Enterprise Secure Access Research and
7492 Development, Cisco Systems, Inc. for discovering this bug and
7493 preparing a fix. [CVE-2012-0050][]
7497 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7499 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7500 of the Vaudenay padding oracle attack on CBC mode encryption
7501 which enables an efficient plaintext recovery attack against
7502 the OpenSSL implementation of DTLS. Their attack exploits timing
7503 differences arising during decryption processing. A research
7504 paper describing this attack can be found at:
7505 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7506 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7507 Security Group at Royal Holloway, University of London
7508 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7509 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7510 for preparing the fix. [CVE-2011-4108][]
7512 *Robin Seggelmann, Michael Tuexen*
7514 * Clear bytes used for block padding of SSL 3.0 records.
7517 *Adam Langley (Google)*
7519 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7520 Kadianakis <desnacked@gmail.com> for discovering this issue and
7521 Adam Langley for preparing the fix. [CVE-2011-4619][]
7523 *Adam Langley (Google)*
7525 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7527 *Andrey Kulikov <amdeich@gmail.com>*
7529 * Prevent malformed RFC3779 data triggering an assertion failure.
7530 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7531 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7533 *Rob Austein <sra@hactrn.net>*
7535 * Improved PRNG seeding for VOS.
7537 *Paul Green <Paul.Green@stratus.com>*
7539 * Fix ssl_ciph.c set-up race.
7541 *Adam Langley (Google)*
7543 * Fix spurious failures in ecdsatest.c.
7545 *Emilia Käsper (Google)*
7547 * Fix the BIO_f_buffer() implementation (which was mixing different
7548 interpretations of the `..._len` fields).
7550 *Adam Langley (Google)*
7552 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7553 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7554 threads won't reuse the same blinding coefficients.
7556 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7557 lock to call BN_BLINDING_invert_ex, and avoids one use of
7558 BN_BLINDING_update for each BN_BLINDING structure (previously,
7559 the last update always remained unused).
7561 *Emilia Käsper (Google)*
7563 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7565 *Bob Buckholz (Google)*
7567 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7569 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7570 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7572 *Kaspar Brand <ossl@velox.ch>*
7574 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7575 for multi-threaded use of ECDH. [CVE-2011-3210][]
7577 *Adam Langley (Google)*
7579 * Fix x509_name_ex_d2i memory leak on bad inputs.
7583 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7584 signature public key algorithm by using OID xref utilities instead.
7585 Before this you could only use some ECC ciphersuites with SHA1 only.
7589 * Add protection against ECDSA timing attacks as mentioned in the paper
7590 by Billy Bob Brumley and Nicola Tuveri, see:
7591 <http://eprint.iacr.org/2011/232.pdf>
7593 *Billy Bob Brumley and Nicola Tuveri*
7595 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7597 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7599 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7601 * Fix bug in string printing code: if *any* escaping is enabled we must
7602 escape the escape character (backslash) or the resulting string is
7607 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7609 * Disable code workaround for ancient and obsolete Netscape browsers
7610 and servers: an attacker can use it in a ciphersuite downgrade attack.
7611 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7615 * Fixed J-PAKE implementation error, originally discovered by
7616 Sebastien Martini, further info and confirmation from Stefan
7617 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7621 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7623 * Fix extension code to avoid race conditions which can result in a buffer
7624 overrun vulnerability: resumed sessions must not be modified as they can
7625 be shared by multiple threads. CVE-2010-3864
7629 * Fix WIN32 build system to correctly link an ENGINE directory into
7634 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7636 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7639 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7641 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7643 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7644 context. The operation can be customised via the ctrl mechanism in
7645 case ENGINEs want to include additional functionality.
7649 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7653 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7654 output hashes compatible with older versions of OpenSSL.
7656 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7658 * Fix compression algorithm handling: if resuming a session use the
7659 compression algorithm of the resumed session instead of determining
7660 it from client hello again. Don't allow server to change algorithm.
7664 * Add load_crls() function to apps tidying load_certs() too. Add option
7665 to verify utility to allow additional CRLs to be included.
7669 * Update OCSP request code to permit adding custom headers to the request:
7670 some responders need this.
7674 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7677 *Julia Lawall <julia@diku.dk>*
7679 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7680 needlessly dereferenced structures, used obsolete functions and
7681 didn't handle all updated verify codes correctly.
7685 * Disable MD2 in the default configuration.
7689 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7690 indicate the initial BIO being pushed or popped. This makes it possible
7691 to determine whether the BIO is the one explicitly called or as a result
7692 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7693 it handles reference counts correctly and doesn't zero out the I/O bio
7694 when it is not being explicitly popped. WARNING: applications which
7695 included workarounds for the old buggy behaviour will need to be modified
7696 or they could free up already freed BIOs.
7700 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7701 renaming to all platforms (within the 0.9.8 branch, this was
7702 done conditionally on Netware platforms to avoid a name clash).
7704 *Guenter <lists@gknw.net>*
7706 * Add ECDHE and PSK support to DTLS.
7708 *Michael Tuexen <tuexen@fh-muenster.de>*
7710 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7715 * Add "missing" function EVP_MD_flags() (without this the only way to
7716 retrieve a digest flags is by accessing the structure directly. Update
7717 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7718 or cipher is registered as in the "from" argument. Print out all
7719 registered digests in the dgst usage message instead of manually
7720 attempting to work them out.
7724 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7725 this allows the use of compression and extensions. Change default cipher
7726 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7727 by default unless an application cipher string requests it.
7731 * Alter match criteria in PKCS12_parse(). It used to try to use local
7732 key ids to find matching certificates and keys but some PKCS#12 files
7733 don't follow the (somewhat unwritten) rules and this strategy fails.
7734 Now just gather all certificates together and the first private key
7735 then look for the first certificate that matches the key.
7739 * Support use of registered digest and cipher names for dgst and cipher
7740 commands instead of having to add each one as a special case. So now
7747 openssl dgst -sha256 foo
7749 and this works for ENGINE based algorithms too.
7753 * Update Gost ENGINE to support parameter files.
7755 *Victor B. Wagner <vitus@cryptocom.ru>*
7757 * Support GeneralizedTime in ca utility.
7759 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7761 * Enhance the hash format used for certificate directory links. The new
7762 form uses the canonical encoding (meaning equivalent names will work
7763 even if they aren't identical) and uses SHA1 instead of MD5. This form
7764 is incompatible with the older format and as a result c_rehash should
7765 be used to rebuild symbolic links.
7769 * Make PKCS#8 the default write format for private keys, replacing the
7770 traditional format. This form is standardised, more secure and doesn't
7771 include an implicit MD5 dependency.
7775 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7776 committed to OpenSSL should pass this lot as a minimum.
7780 * Add session ticket override functionality for use by EAP-FAST.
7782 *Jouni Malinen <j@w1.fi>*
7784 * Modify HMAC functions to return a value. Since these can be implemented
7785 in an ENGINE errors can occur.
7789 * Type-checked OBJ_bsearch_ex.
7793 * Type-checked OBJ_bsearch. Also some constification necessitated
7794 by type-checking. Still to come: TXT_DB, bsearch(?),
7795 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7800 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7801 seconds to a tm structure directly, instead of going through OS
7802 specific date routines. This avoids any issues with OS routines such
7803 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7804 and X509_time_adj_ex() to cover the extended range. The existing
7805 X509_time_adj() is still usable and will no longer have any date issues.
7809 * Delta CRL support. New use deltas option which will attempt to locate
7810 and search any appropriate delta CRLs available.
7812 This work was sponsored by Google.
7816 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7817 code and add additional score elements. Validate alternate CRL paths
7818 as part of the CRL checking and indicate a new error "CRL path validation
7819 error" in this case. Applications wanting additional details can use
7820 the verify callback and check the new "parent" field. If this is not
7821 NULL CRL path validation is taking place. Existing applications won't
7822 see this because it requires extended CRL support which is off by
7825 This work was sponsored by Google.
7829 * Support for freshest CRL extension.
7831 This work was sponsored by Google.
7835 * Initial indirect CRL support. Currently only supported in the CRLs
7836 passed directly and not via lookup. Process certificate issuer
7837 CRL entry extension and lookup CRL entries by bother issuer name
7838 and serial number. Check and process CRL issuer entry in IDP extension.
7840 This work was sponsored by Google.
7844 * Add support for distinct certificate and CRL paths. The CRL issuer
7845 certificate is validated separately in this case. Only enabled if
7846 an extended CRL support flag is set: this flag will enable additional
7847 CRL functionality in future.
7849 This work was sponsored by Google.
7853 * Add support for policy mappings extension.
7855 This work was sponsored by Google.
7859 * Fixes to pathlength constraint, self issued certificate handling,
7860 policy processing to align with RFC3280 and PKITS tests.
7862 This work was sponsored by Google.
7866 * Support for name constraints certificate extension. DN, email, DNS
7867 and URI types are currently supported.
7869 This work was sponsored by Google.
7873 * To cater for systems that provide a pointer-based thread ID rather
7874 than numeric, deprecate the current numeric thread ID mechanism and
7875 replace it with a structure and associated callback type. This
7876 mechanism allows a numeric "hash" to be extracted from a thread ID in
7877 either case, and on platforms where pointers are larger than 'long',
7878 mixing is done to help ensure the numeric 'hash' is usable even if it
7879 can't be guaranteed unique. The default mechanism is to use "&errno"
7880 as a pointer-based thread ID to distinguish between threads.
7882 Applications that want to provide their own thread IDs should now use
7883 CRYPTO_THREADID_set_callback() to register a callback that will call
7884 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7886 Note that ERR_remove_state() is now deprecated, because it is tied
7887 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7888 to free the current thread's error state should be replaced by
7889 ERR_remove_thread_state(NULL).
7891 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7892 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7893 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7894 application was previously providing a numeric thread callback that
7895 was inappropriate for distinguishing threads, then uniqueness might
7896 have been obtained with &errno that happened immediately in the
7897 intermediate development versions of OpenSSL; this is no longer the
7898 case, the numeric thread callback will now override the automatic use
7901 *Geoff Thorpe, with help from Bodo Moeller*
7903 * Initial support for different CRL issuing certificates. This covers a
7904 simple case where the self issued certificates in the chain exist and
7905 the real CRL issuer is higher in the existing chain.
7907 This work was sponsored by Google.
7911 * Removed effectively defunct crypto/store from the build.
7915 * Revamp of STACK to provide stronger type-checking. Still to come:
7916 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7917 ASN1_STRING, CONF_VALUE.
7921 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7922 RAM on SSL connections. This option can save about 34k per idle SSL.
7926 * Revamp of LHASH to provide stronger type-checking. Still to come:
7927 STACK, TXT_DB, bsearch, qsort.
7931 * Initial support for Cryptographic Message Syntax (aka CMS) based
7932 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7933 support for data, signedData, compressedData, digestedData and
7934 encryptedData, envelopedData types included. Scripts to check against
7935 RFC4134 examples draft and interop and consistency checks of many
7936 content types and variants.
7940 * Add options to enc utility to support use of zlib compression BIO.
7944 * Extend mk1mf to support importing of options and assembly language
7945 files from Configure script, currently only included in VC-WIN32.
7946 The assembly language rules can now optionally generate the source
7947 files from the associated perl scripts.
7951 * Implement remaining functionality needed to support GOST ciphersuites.
7952 Interop testing has been performed using CryptoPro implementations.
7954 *Victor B. Wagner <vitus@cryptocom.ru>*
7956 * s390x assembler pack.
7960 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7965 * Implement Opaque PRF Input TLS extension as specified in
7966 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7967 official specification yet and no extension type assignment by
7968 IANA exists, this extension (for now) will have to be explicitly
7969 enabled when building OpenSSL by providing the extension number
7970 to use. For example, specify an option
7972 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7974 to the "config" or "Configure" script to enable the extension,
7975 assuming extension number 0x9527 (which is a completely arbitrary
7976 and unofficial assignment based on the MD5 hash of the Internet
7977 Draft). Note that by doing so, you potentially lose
7978 interoperability with other TLS implementations since these might
7979 be using the same extension number for other purposes.
7981 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7982 opaque PRF input value to use in the handshake. This will create
7983 an internal copy of the length-'len' string at 'src', and will
7984 return non-zero for success.
7986 To get more control and flexibility, provide a callback function
7989 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7990 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7994 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7997 Callback function 'cb' will be called in handshakes, and is
7998 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7999 Argument 'arg' is for application purposes (the value as given to
8000 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8001 be provided to the callback function). The callback function
8002 has to return non-zero to report success: usually 1 to use opaque
8003 PRF input just if possible, or 2 to enforce use of the opaque PRF
8004 input. In the latter case, the library will abort the handshake
8005 if opaque PRF input is not successfully negotiated.
8007 Arguments 'peerinput' and 'len' given to the callback function
8008 will always be NULL and 0 in the case of a client. A server will
8009 see the client's opaque PRF input through these variables if
8010 available (NULL and 0 otherwise). Note that if the server
8011 provides an opaque PRF input, the length must be the same as the
8012 length of the client's opaque PRF input.
8014 Note that the callback function will only be called when creating
8015 a new session (session resumption can resume whatever was
8016 previously negotiated), and will not be called in SSL 2.0
8017 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8018 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8019 for applications that need to enforce opaque PRF input.
8023 * Update ssl code to support digests other than SHA1+MD5 for handshake
8026 *Victor B. Wagner <vitus@cryptocom.ru>*
8028 * Add RFC4507 support to OpenSSL. This includes the corrections in
8029 RFC4507bis. The encrypted ticket format is an encrypted encoded
8030 SSL_SESSION structure, that way new session features are automatically
8033 If a client application caches session in an SSL_SESSION structure
8034 support is transparent because tickets are now stored in the encoded
8037 The SSL_CTX structure automatically generates keys for ticket
8038 protection in servers so again support should be possible
8039 with no application modification.
8041 If a client or server wishes to disable RFC4507 support then the option
8042 SSL_OP_NO_TICKET can be set.
8044 Add a TLS extension debugging callback to allow the contents of any client
8045 or server extensions to be examined.
8047 This work was sponsored by Google.
8051 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8052 OpenSSL should now compile cleanly on gcc 4.2
8054 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8056 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8057 support including streaming MAC support: this is required for GOST
8058 ciphersuite support.
8060 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8062 * Add option -stream to use PKCS#7 streaming in smime utility. New
8063 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8064 to output in BER and PEM format.
8068 * Experimental support for use of HMAC via EVP_PKEY interface. This
8069 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8070 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8071 ENGINE support for HMAC keys which are unextractable. New -mac and
8072 -macopt options to dgst utility.
8076 * New option -sigopt to dgst utility. Update dgst to use
8077 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8078 alternative signing parameters such as X9.31 or PSS in the dgst
8083 * Change ssl_cipher_apply_rule(), the internal function that does
8084 the work each time a ciphersuite string requests enabling
8085 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8086 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8087 the order of disabled ciphersuites such that those ciphersuites
8088 that most recently went from enabled to disabled not only stay
8089 in order with respect to each other, but also have higher priority
8090 than other disabled ciphersuites the next time ciphersuites are
8093 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8094 the same ciphersuites as with "HIGH" alone, but in a specific
8095 order where the PSK ciphersuites come first (since they are the
8096 most recently disabled ciphersuites when "HIGH" is parsed).
8098 Also, change ssl_create_cipher_list() (using this new
8099 functionality) such that between otherwise identical
8100 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8105 * Change ssl_create_cipher_list() so that it automatically
8106 arranges the ciphersuites in reasonable order before starting
8107 to process the rule string. Thus, the definition for "DEFAULT"
8108 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8109 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8110 This makes it much easier to arrive at a reasonable default order
8111 in applications for which anonymous ciphers are OK (meaning
8112 that you can't actually use DEFAULT).
8114 *Bodo Moeller; suggested by Victor Duchovni*
8116 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8117 processing) into multiple integers instead of setting
8118 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8119 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8120 (These masks as well as the individual bit definitions are hidden
8121 away into the non-exported interface ssl/ssl_locl.h, so this
8122 change to the definition of the SSL_CIPHER structure shouldn't
8123 affect applications.) This give us more bits for each of these
8124 categories, so there is no longer a need to coagulate AES128 and
8125 AES256 into a single algorithm bit, and to coagulate Camellia128
8126 and Camellia256 into a single algorithm bit, which has led to all
8129 Thus, among other things, the kludge introduced in 0.9.7m and
8130 0.9.8e for masking out AES256 independently of AES128 or masking
8131 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8133 With the change, we also introduce new ciphersuite aliases that
8134 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8139 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8140 Use the leftmost N bytes of the signature input if the input is
8141 larger than the prime q (with N being the size in bytes of q).
8145 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8146 it yet and it is largely untested.
8150 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8154 * Initial incomplete changes to avoid need for function casts in OpenSSL
8155 some compilers (gcc 4.2 and later) reject their use. Safestack is
8156 reimplemented. Update ASN1 to avoid use of legacy functions.
8160 * Win32/64 targets are linked with Winsock2.
8164 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8165 to external functions. This can be used to increase CRL handling
8166 efficiency especially when CRLs are very large by (for example) storing
8167 the CRL revoked certificates in a database.
8171 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8172 new CRLs added to a directory can be used. New command line option
8173 -verify_return_error to s_client and s_server. This causes real errors
8174 to be returned by the verify callback instead of carrying on no matter
8175 what. This reflects the way a "real world" verify callback would behave.
8179 * GOST engine, supporting several GOST algorithms and public key formats.
8180 Kindly donated by Cryptocom.
8184 * Partial support for Issuing Distribution Point CRL extension. CRLs
8185 partitioned by DP are handled but no indirect CRL or reason partitioning
8186 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8187 selected via a scoring technique which handles IDP and AKID in CRLs.
8191 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8192 will ultimately be used for all verify operations: this will remove the
8193 X509_STORE dependency on certificate verification and allow alternative
8194 lookup methods. X509_STORE based implementations of these two callbacks.
8198 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8199 Modify get_crl() to find a valid (unexpired) CRL if possible.
8203 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8204 this would be called X509_CRL_cmp() but that name is already used by
8205 a function that just compares CRL issuer names. Cache several CRL
8206 extensions in X509_CRL structure and cache CRLDP in X509.
8210 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8211 this maps equivalent X509_NAME structures into a consistent structure.
8212 Name comparison can then be performed rapidly using memcmp().
8216 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8221 * Allow digests to supply their own micalg string for S/MIME type using
8222 the ctrl EVP_MD_CTRL_MICALG.
8226 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8227 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8228 ctrl. It can then customise the structure before and/or after signing
8233 * New function OBJ_add_sigid() to allow application defined signature OIDs
8234 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8235 to free up any added signature OIDs.
8239 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8240 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8241 digest and cipher tables. New options added to openssl utility:
8242 list-message-digest-algorithms and list-cipher-algorithms.
8246 * Change the array representation of binary polynomials: the list
8247 of degrees of non-zero coefficients is now terminated with -1.
8248 Previously it was terminated with 0, which was also part of the
8249 value; thus, the array representation was not applicable to
8250 polynomials where t^0 has coefficient zero. This change makes
8251 the array representation useful in a more general context.
8255 * Various modifications and fixes to SSL/TLS cipher string
8256 handling. For ECC, the code now distinguishes between fixed ECDH
8257 with RSA certificates on the one hand and with ECDSA certificates
8258 on the other hand, since these are separate ciphersuites. The
8259 unused code for Fortezza ciphersuites has been removed.
8261 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8262 (not "ECDHE"). For consistency with the code for DH
8263 certificates, use of ECDH certificates is now considered ECDH
8264 authentication, not RSA or ECDSA authentication (the latter is
8265 merely the CA's signing algorithm and not actively used in the
8268 The temporary ciphersuite alias "ECCdraft" is no longer
8269 available, and ECC ciphersuites are no longer excluded from "ALL"
8270 and "DEFAULT". The following aliases now exist for RFC 4492
8271 ciphersuites, most of these by analogy with the DH case:
8273 kECDHr - ECDH cert, signed with RSA
8274 kECDHe - ECDH cert, signed with ECDSA
8275 kECDH - ECDH cert (signed with either RSA or ECDSA)
8276 kEECDH - ephemeral ECDH
8277 ECDH - ECDH cert or ephemeral ECDH
8283 AECDH - anonymous ECDH
8284 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8288 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8289 Use correct micalg parameters depending on digest(s) in signed message.
8293 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8294 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8298 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8299 an engine to register a method. Add ENGINE lookups for methods and
8300 functional reference processing.
8304 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8305 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8310 * New -resign option to smime utility. This adds one or more signers
8311 to an existing PKCS#7 signedData structure. Also -md option to use an
8312 alternative message digest algorithm for signing.
8316 * Tidy up PKCS#7 routines and add new functions to make it easier to
8317 create PKCS7 structures containing multiple signers. Update smime
8318 application to support multiple signers.
8322 * New -macalg option to pkcs12 utility to allow setting of an alternative
8327 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8328 Reorganize PBE internals to lookup from a static table using NIDs,
8329 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8330 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8331 PRF which will be automatically used with PBES2.
8335 * Replace the algorithm specific calls to generate keys in "req" with the
8340 * Update PKCS#7 enveloped data routines to use new API. This is now
8341 supported by any public key method supporting the encrypt operation. A
8342 ctrl is added to allow the public key algorithm to examine or modify
8343 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8348 * Add a ctrl to asn1 method to allow a public key algorithm to express
8349 a default digest type to use. In most cases this will be SHA1 but some
8350 algorithms (such as GOST) need to specify an alternative digest. The
8351 return value indicates how strong the preference is 1 means optional and
8352 2 is mandatory (that is it is the only supported type). Modify
8353 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8354 use the default md. Update openssl utilities to use the default digest
8355 type for signing if it is not explicitly indicated.
8359 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8360 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8361 signing method from the key type. This effectively removes the link
8362 between digests and public key types.
8366 * Add an OID cross reference table and utility functions. Its purpose is to
8367 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8368 rsaEncryption. This will allow some of the algorithm specific hackery
8369 needed to use the correct OID to be removed.
8373 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8374 structures for PKCS7_sign(). They are now set up by the relevant public
8379 * Add provisional EC pkey method with support for ECDSA and ECDH.
8383 * Add support for key derivation (agreement) in the API, DH method and
8388 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8389 public and private key formats. As a side effect these add additional
8390 command line functionality not previously available: DSA signatures can be
8391 generated and verified using pkeyutl and DH key support and generation in
8398 *Oliver Tappe <zooey@hirschkaefer.de>*
8400 * New make target "install_html_docs" installs HTML renditions of the
8403 *Oliver Tappe <zooey@hirschkaefer.de>*
8405 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8406 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8407 support key and parameter generation and add initial key generation
8408 functionality for RSA.
8412 * Add functions for main EVP_PKEY_method operations. The undocumented
8413 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8414 `EVP_PKEY_{encrypt,decrypt}_old`.
8418 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8419 key API, doesn't do much yet.
8423 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8424 public key algorithms. New option to openssl utility:
8425 "list-public-key-algorithms" to print out info.
8429 * Implement the Supported Elliptic Curves Extension for
8430 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8434 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8435 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8439 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8440 utilities such as rsa, dsa, dsaparam etc except they process any key
8445 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8446 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8447 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8452 * Initial support for pluggable public key ASN1.
8453 De-spaghettify the public key ASN1 handling. Move public and private
8454 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8455 algorithm specific handling to a single module within the relevant
8456 algorithm directory. Add functions to allow (near) opaque processing
8457 of public and private key structures.
8461 * Implement the Supported Point Formats Extension for
8462 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8466 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8467 for the psk identity [hint] and the psk callback functions to the
8468 SSL_SESSION, SSL and SSL_CTX structure.
8471 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8475 SSL_CTX_use_psk_identity_hint
8476 SSL_get_psk_identity_hint
8477 SSL_get_psk_identity
8478 SSL_use_psk_identity_hint
8480 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8482 * Add RFC 3161 compliant time stamp request creation, response generation
8483 and response verification functionality.
8485 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8487 * Add initial support for TLS extensions, specifically for the server_name
8488 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8489 have new members for a host name. The SSL data structure has an
8490 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8491 stored in that context to allow for session resumption, even after the
8492 SSL has been switched to a new SSL_CTX in reaction to a client's
8493 server_name extension.
8495 New functions (subject to change):
8497 SSL_get_servername()
8498 SSL_get_servername_type()
8501 New CTRL codes and macros (subject to change):
8503 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8504 - SSL_CTX_set_tlsext_servername_callback()
8505 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8506 - SSL_CTX_set_tlsext_servername_arg()
8507 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8509 openssl s_client has a new '-servername ...' option.
8511 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8512 '-key2 ...', '-servername_fatal' (subject to change). This allows
8513 testing the HostName extension for a specific single host name ('-cert'
8514 and '-key' remain fallbacks for handshakes without HostName
8515 negotiation). If the unrecognized_name alert has to be sent, this by
8516 default is a warning; it becomes fatal with the '-servername_fatal'
8519 *Peter Sylvester, Remy Allais, Christophe Renou*
8521 * Whirlpool hash implementation is added.
8525 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8526 bn(64,32). Because of instruction set limitations it doesn't have
8527 any negative impact on performance. This was done mostly in order
8528 to make it possible to share assembler modules, such as bn_mul_mont
8529 implementations, between 32- and 64-bit builds without hassle.
8533 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8534 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8539 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8540 dedicated Montgomery multiplication procedure, is introduced.
8541 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8542 "64-bit" performance on certain 32-bit targets.
8546 * New option SSL_OP_NO_COMP to disable use of compression selectively
8547 in SSL structures. New SSL ctrl to set maximum send fragment size.
8548 Save memory by setting the I/O buffer sizes dynamically instead of
8549 using the maximum available value.
8553 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8554 in addition to the text details.
8558 * Very, very preliminary EXPERIMENTAL support for printing of general
8559 ASN1 structures. This currently produces rather ugly output and doesn't
8560 handle several customised structures at all.
8564 * Integrated support for PVK file format and some related formats such
8565 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8566 these in the 'rsa' and 'dsa' utilities.
8570 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8574 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8575 place for the (very old) "NETSCAPE" format certificates which are now
8576 handled using new ASN1 code equivalents.
8580 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8581 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8582 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8586 * Modify CRL distribution points extension code to print out previously
8587 unsupported fields. Enhance extension setting code to allow setting of
8592 * Add print and set support for Issuing Distribution Point CRL extension.
8596 * Change 'Configure' script to enable Camellia by default.
8603 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8605 * When rejecting SSL/TLS records due to an incorrect version number, never
8606 update s->server with a new major version number. As of
8607 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8608 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8609 the previous behavior could result in a read attempt at NULL when
8610 receiving specific incorrect SSL/TLS records once record payload
8611 protection is active. [CVE-2010-0740][]
8613 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8615 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8616 could be crashed if the relevant tables were not present (e.g. chrooted).
8618 *Tomas Hoger <thoger@redhat.com>*
8620 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8622 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8624 *Martin Olsson, Neel Mehta*
8626 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8627 accommodate for stack sorting, always a write lock!).
8631 * On some versions of WIN32 Heap32Next is very slow. This can cause
8632 excessive delays in the RAND_poll(): over a minute. As a workaround
8633 include a time check in the inner Heap32Next loop too.
8637 * The code that handled flushing of data in SSL/TLS originally used the
8638 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8639 the problem outlined in PR#1949. The fix suggested there however can
8640 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8641 of Apache). So instead simplify the code to flush unconditionally.
8642 This should be fine since flushing with no data to flush is a no op.
8646 * Handle TLS versions 2.0 and later properly and correctly use the
8647 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8648 off ancient servers have a habit of sticking around for a while...
8652 * Modify compression code so it frees up structures without using the
8653 ex_data callbacks. This works around a problem where some applications
8654 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8655 restarting) then use compression (e.g. SSL with compression) later.
8656 This results in significant per-connection memory leaks and
8657 has caused some security issues including CVE-2008-1678 and
8662 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8663 change when encrypting or decrypting.
8667 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8668 connect and renegotiate with servers which do not support RI.
8669 Until RI is more widely deployed this option is enabled by default.
8673 * Add "missing" ssl ctrls to clear options and mode.
8677 * If client attempts to renegotiate and doesn't support RI respond with
8678 a no_renegotiation alert as required by RFC5746. Some renegotiating
8679 TLS clients will continue a connection gracefully when they receive
8680 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8681 waiting for a server hello which it will never receive. Now we treat a
8682 received no_renegotiation alert as a fatal error. This is because
8683 applications requesting a renegotiation might well expect it to succeed
8684 and would have no code in place to handle the server denying it so the
8685 only safe thing to do is to terminate the connection.
8689 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8690 peer supports secure renegotiation and 0 otherwise. Print out peer
8691 renegotiation support in s_client/s_server.
8695 * Replace the highly broken and deprecated SPKAC certification method with
8696 the updated NID creation version. This should correctly handle UTF8.
8700 * Implement RFC5746. Re-enable renegotiation but require the extension
8701 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8702 turns out to be a bad idea. It has been replaced by
8703 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8704 SSL_CTX_set_options(). This is really not recommended unless you
8705 know what you are doing.
8707 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8709 * Fixes to stateless session resumption handling. Use initial_ctx when
8710 issuing and attempting to decrypt tickets in case it has changed during
8711 servername handling. Use a non-zero length session ID when attempting
8712 stateless session resumption: this makes it possible to determine if
8713 a resumption has occurred immediately after receiving server hello
8714 (several places in OpenSSL subtly assume this) instead of later in
8719 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8720 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8721 fixes for a few places where the return code is not checked
8724 *Julia Lawall <julia@diku.dk>*
8726 * Add --strict-warnings option to Configure script to include devteam
8727 warnings in other configurations.
8731 * Add support for --libdir option and LIBDIR variable in makefiles. This
8732 makes it possible to install openssl libraries in locations which
8733 have names other than "lib", for example "/usr/lib64" which some
8736 *Steve Henson, based on patch from Jeremy Utley*
8738 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8739 X690 8.9.12 and can produce some misleading textual output of OIDs.
8741 *Steve Henson, reported by Dan Kaminsky*
8743 * Delete MD2 from algorithm tables. This follows the recommendation in
8744 several standards that it is not used in new applications due to
8745 several cryptographic weaknesses. For binary compatibility reasons
8746 the MD2 API is still compiled in by default.
8750 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8755 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8756 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8759 *Guenter <lists@gknw.net>*
8761 * Fix the server certificate chain building code to use X509_verify_cert(),
8762 it used to have an ad-hoc builder which was unable to cope with anything
8763 other than a simple chain.
8765 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8767 * Don't check self signed certificate signatures in X509_verify_cert()
8768 by default (a flag can override this): it just wastes time without
8769 adding any security. As a useful side effect self signed root CAs
8770 with non-FIPS digests are now usable in FIPS mode.
8774 * In dtls1_process_out_of_seq_message() the check if the current message
8775 is already buffered was missing. For every new message was memory
8776 allocated, allowing an attacker to perform an denial of service attack
8777 with sending out of seq handshake messages until there is no memory
8778 left. Additionally every future message was buffered, even if the
8779 sequence number made no sense and would be part of another handshake.
8780 So only messages with sequence numbers less than 10 in advance will be
8781 buffered. [CVE-2009-1378][]
8783 *Robin Seggelmann, discovered by Daniel Mentz*
8785 * Records are buffered if they arrive with a future epoch to be
8786 processed after finishing the corresponding handshake. There is
8787 currently no limitation to this buffer allowing an attacker to perform
8788 a DOS attack with sending records with future epochs until there is no
8789 memory left. This patch adds the pqueue_size() function to determine
8790 the size of a buffer and limits the record buffer to 100 entries.
8793 *Robin Seggelmann, discovered by Daniel Mentz*
8795 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8796 parent structure is freed. [CVE-2009-1379][]
8800 * Handle non-blocking I/O properly in SSL_shutdown() call.
8802 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8804 * Add `2.5.4.*` OIDs
8806 *Ilya O. <vrghost@gmail.com>*
8808 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8810 * Disable renegotiation completely - this fixes a severe security
8811 problem [CVE-2009-3555][] at the cost of breaking all
8812 renegotiation. Renegotiation can be re-enabled by setting
8813 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8814 run-time. This is really not recommended unless you know what
8819 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8821 * Don't set val to NULL when freeing up structures, it is freed up by
8822 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8823 zeroing past the valid field. [CVE-2009-0789][]
8825 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8827 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8828 checked correctly. This would allow some invalid signed attributes to
8829 appear to verify correctly. [CVE-2009-0591][]
8831 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8833 * Reject UniversalString and BMPString types with invalid lengths. This
8834 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8835 a legal length. [CVE-2009-0590][]
8839 * Set S/MIME signing as the default purpose rather than setting it
8840 unconditionally. This allows applications to override it at the store
8845 * Permit restricted recursion of ASN1 strings. This is needed in practice
8846 to handle some structures.
8850 * Improve efficiency of mem_gets: don't search whole buffer each time
8853 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8855 * New -hex option for openssl rand.
8859 * Print out UTF8String and NumericString when parsing ASN1.
8863 * Support NumericString type for name components.
8867 * Allow CC in the environment to override the automatically chosen
8868 compiler. Note that nothing is done to ensure flags work with the
8873 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8875 * Properly check EVP_VerifyFinal() and similar return values
8878 *Ben Laurie, Bodo Moeller, Google Security Team*
8880 * Enable TLS extensions by default.
8884 * Allow the CHIL engine to be loaded, whether the application is
8885 multithreaded or not. (This does not release the developer from the
8886 obligation to set up the dynamic locking callbacks.)
8888 *Sander Temme <sander@temme.net>*
8890 * Use correct exit code if there is an error in dgst command.
8892 *Steve Henson; problem pointed out by Roland Dirlewanger*
8894 * Tweak Configure so that you need to say "experimental-jpake" to enable
8895 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8899 * Add experimental JPAKE support, including demo authentication in
8900 s_client and s_server.
8904 * Set the comparison function in v3_addr_canonize().
8906 *Rob Austein <sra@hactrn.net>*
8908 * Add support for XMPP STARTTLS in s_client.
8910 *Philip Paeps <philip@freebsd.org>*
8912 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8913 to ensure that even with this option, only ciphersuites in the
8914 server's preference list will be accepted. (Note that the option
8915 applies only when resuming a session, so the earlier behavior was
8916 just about the algorithm choice for symmetric cryptography.)
8920 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8922 * Fix NULL pointer dereference if a DTLS server received
8923 ChangeCipherSpec as first record [CVE-2009-1386][].
8927 * Fix a state transition in s3_srvr.c and d1_srvr.c
8928 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8932 * The fix in 0.9.8c that supposedly got rid of unsafe
8933 double-checked locking was incomplete for RSA blinding,
8934 addressing just one layer of what turns out to have been
8935 doubly unsafe triple-checked locking.
8937 So now fix this for real by retiring the MONT_HELPER macro
8938 in crypto/rsa/rsa_eay.c.
8940 *Bodo Moeller; problem pointed out by Marius Schilder*
8942 * Various precautionary measures:
8944 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8946 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8947 (NB: This would require knowledge of the secret session ticket key
8948 to exploit, in which case you'd be SOL either way.)
8950 - Change bn_nist.c so that it will properly handle input BIGNUMs
8951 outside the expected range.
8953 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8956 *Neel Mehta, Bodo Moeller*
8958 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8959 the load fails. Useful for distros.
8961 *Ben Laurie and the FreeBSD team*
8963 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8967 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8971 * Expand ENGINE to support engine supplied SSL client certificate functions.
8973 This work was sponsored by Logica.
8977 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8978 keystores. Support for SSL/TLS client authentication too.
8979 Not compiled unless enable-capieng specified to Configure.
8981 This work was sponsored by Logica.
8985 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8986 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8987 attribute creation routines such as certificate requests and PKCS#12
8992 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8994 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8995 handshake which could lead to a client crash as found using the
8996 Codenomicon TLS test suite [CVE-2008-1672][]
8998 *Steve Henson, Mark Cox*
9000 * Fix double free in TLS server name extensions which could lead to
9001 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
9005 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9007 Clear the error queue to ensure that error entries left from
9008 older function calls do not interfere with the correct operation.
9010 *Lutz Jaenicke, Erik de Castro Lopo*
9012 * Remove root CA certificates of commercial CAs:
9014 The OpenSSL project does not recommend any specific CA and does not
9015 have any policy with respect to including or excluding any CA.
9016 Therefore it does not make any sense to ship an arbitrary selection
9017 of root CA certificates with the OpenSSL software.
9021 * RSA OAEP patches to fix two separate invalid memory reads.
9022 The first one involves inputs when 'lzero' is greater than
9023 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9024 before the beginning of from). The second one involves inputs where
9025 the 'db' section contains nothing but zeroes (there is a one-byte
9026 invalid read after the end of 'db').
9028 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9030 * Partial backport from 0.9.9-dev:
9032 Introduce bn_mul_mont (dedicated Montgomery multiplication
9033 procedure) as a candidate for BIGNUM assembler implementation.
9034 While 0.9.9-dev uses assembler for various architectures, only
9035 x86_64 is available by default here in the 0.9.8 branch, and
9036 32-bit x86 is available through a compile-time setting.
9038 To try the 32-bit x86 assembler implementation, use Configure
9039 option "enable-montasm" (which exists only for this backport).
9041 As "enable-montasm" for 32-bit x86 disclaims code stability
9042 anyway, in this constellation we activate additional code
9043 backported from 0.9.9-dev for further performance improvements,
9044 namely BN_from_montgomery_word. (To enable this otherwise,
9045 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9047 *Andy Polyakov (backport partially by Bodo Moeller)*
9049 * Add TLS session ticket callback. This allows an application to set
9050 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9051 values. This is useful for key rollover for example where several key
9052 sets may exist with different names.
9056 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9057 This was broken until now in 0.9.8 releases, such that the only way
9058 a registered ENGINE could be used (assuming it initialises
9059 successfully on the host) was to explicitly set it as the default
9060 for the relevant algorithms. This is in contradiction with 0.9.7
9061 behaviour and the documentation. With this fix, when an ENGINE is
9062 registered into a given algorithm's table of implementations, the
9063 'uptodate' flag is reset so that auto-discovery will be used next
9064 time a new context for that algorithm attempts to select an
9067 *Ian Lister (tweaked by Geoff Thorpe)*
9069 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9070 implementation in the following ways:
9072 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9075 Lack of BER streaming support means one pass streaming processing is
9076 only supported if data is detached: setting the streaming flag is
9077 ignored for embedded content.
9079 CMS support is disabled by default and must be explicitly enabled
9080 with the enable-cms configuration option.
9084 * Update the GMP engine glue to do direct copies between BIGNUM and
9085 mpz_t when openssl and GMP use the same limb size. Otherwise the
9086 existing "conversion via a text string export" trick is still used.
9088 *Paul Sheer <paulsheer@gmail.com>*
9090 * Zlib compression BIO. This is a filter BIO which compressed and
9091 uncompresses any data passed through it.
9095 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9096 RFC3394 compatible AES key wrapping.
9100 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9101 sets string data without copying. X509_ALGOR_set0() and
9102 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9103 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9104 from an X509_ATTRIBUTE structure optionally checking it occurs only
9105 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9110 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9111 to get the expected BN_FLG_CONSTTIME behavior.
9113 *Bodo Moeller (Google)*
9117 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9118 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9119 - added some more tests to do_tests.pl
9120 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9121 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9122 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9123 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9124 - various changes to netware.pl to enable gcc-cross builds on Win32
9126 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9127 - various changes to fix missing prototype warnings
9128 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9129 - added AES, WHIRLPOOL and CPUID assembler code to build files
9130 - added missing AES assembler make rules to mk1mf.pl
9131 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9133 *Guenter Knauf <eflash@gmx.net>*
9135 * Implement certificate status request TLS extension defined in RFC3546.
9136 A client can set the appropriate parameters and receive the encoded
9137 OCSP response via a callback. A server can query the supplied parameters
9138 and set the encoded OCSP response in the callback. Add simplified examples
9139 to s_client and s_server.
9143 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9146 + Binary incompatibility of ssl_ctx_st structure
9147 + DTLS interoperation with non-compliant servers
9148 + Don't call get_session_cb() without proposed session
9149 + Fix ia64 assembler code
9151 *Andy Polyakov, Steve Henson*
9153 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9155 * DTLS Handshake overhaul. There were longstanding issues with
9156 OpenSSL DTLS implementation, which were making it impossible for
9157 RFC 4347 compliant client to communicate with OpenSSL server.
9158 Unfortunately just fixing these incompatibilities would "cut off"
9159 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9160 server keeps tolerating non RFC compliant syntax. The opposite is
9161 not true, 0.9.8f client can not communicate with earlier server.
9162 This update even addresses CVE-2007-4995.
9166 * Changes to avoid need for function casts in OpenSSL: some compilers
9167 (gcc 4.2 and later) reject their use.
9168 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9171 * Add RFC4507 support to OpenSSL. This includes the corrections in
9172 RFC4507bis. The encrypted ticket format is an encrypted encoded
9173 SSL_SESSION structure, that way new session features are automatically
9176 If a client application caches session in an SSL_SESSION structure
9177 support is transparent because tickets are now stored in the encoded
9180 The SSL_CTX structure automatically generates keys for ticket
9181 protection in servers so again support should be possible
9182 with no application modification.
9184 If a client or server wishes to disable RFC4507 support then the option
9185 SSL_OP_NO_TICKET can be set.
9187 Add a TLS extension debugging callback to allow the contents of any client
9188 or server extensions to be examined.
9190 This work was sponsored by Google.
9194 * Add initial support for TLS extensions, specifically for the server_name
9195 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9196 have new members for a host name. The SSL data structure has an
9197 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9198 stored in that context to allow for session resumption, even after the
9199 SSL has been switched to a new SSL_CTX in reaction to a client's
9200 server_name extension.
9202 New functions (subject to change):
9204 SSL_get_servername()
9205 SSL_get_servername_type()
9208 New CTRL codes and macros (subject to change):
9210 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9211 - SSL_CTX_set_tlsext_servername_callback()
9212 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9213 - SSL_CTX_set_tlsext_servername_arg()
9214 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9216 openssl s_client has a new '-servername ...' option.
9218 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9219 '-key2 ...', '-servername_fatal' (subject to change). This allows
9220 testing the HostName extension for a specific single host name ('-cert'
9221 and '-key' remain fallbacks for handshakes without HostName
9222 negotiation). If the unrecognized_name alert has to be sent, this by
9223 default is a warning; it becomes fatal with the '-servername_fatal'
9226 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9228 * Add AES and SSE2 assembly language support to VC++ build.
9232 * Mitigate attack on final subtraction in Montgomery reduction.
9236 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9237 (which previously caused an internal error).
9241 * Squeeze another 10% out of IGE mode when in != out.
9245 * AES IGE mode speedup.
9247 *Dean Gaudet (Google)*
9249 * Add the Korean symmetric 128-bit cipher SEED (see
9250 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9251 add SEED ciphersuites from RFC 4162:
9253 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9254 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9255 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9256 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9258 To minimize changes between patchlevels in the OpenSSL 0.9.8
9259 series, SEED remains excluded from compilation unless OpenSSL
9260 is configured with 'enable-seed'.
9262 *KISA, Bodo Moeller*
9264 * Mitigate branch prediction attacks, which can be practical if a
9265 single processor is shared, allowing a spy process to extract
9266 information. For detailed background information, see
9267 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9268 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9269 and Necessary Software Countermeasures"). The core of the change
9270 are new versions BN_div_no_branch() and
9271 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9272 respectively, which are slower, but avoid the security-relevant
9273 conditional branches. These are automatically called by BN_div()
9274 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9275 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9276 remove a conditional branch.
9278 BN_FLG_CONSTTIME is the new name for the previous
9279 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9280 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9281 in the exponent causes BN_mod_exp_mont() to use the alternative
9282 implementation in BN_mod_exp_mont_consttime().) The old name
9283 remains as a deprecated alias.
9285 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9286 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9287 constant-time implementations for more than just exponentiation.
9288 Here too the old name is kept as a deprecated alias.
9290 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9291 the BN_BLINDING structure gets an independent copy of the
9292 modulus. This means that the previous `BIGNUM *m` argument to
9293 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9294 essentially becomes `const BIGNUM *m`, although we can't actually
9295 change this in the header file before 0.9.9. It allows
9296 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9297 enable BN_FLG_CONSTTIME.
9299 *Matthew D Wood (Intel Corp)*
9301 * In the SSL/TLS server implementation, be strict about session ID
9302 context matching (which matters if an application uses a single
9303 external cache for different purposes). Previously,
9304 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9305 set. This did ensure strict client verification, but meant that,
9306 with applications using a single external cache for quite
9307 different requirements, clients could circumvent ciphersuite
9308 restrictions for a given session ID context by starting a session
9309 in a different context.
9313 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9314 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9315 authentication-only ciphersuites.
9319 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9320 not complete and could lead to a possible single byte overflow
9321 [CVE-2007-5135][] [Ben Laurie]
9323 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9325 * Since AES128 and AES256 (and similarly Camellia128 and
9326 Camellia256) share a single mask bit in the logic of
9327 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9328 kludge to work properly if AES128 is available and AES256 isn't
9329 (or if Camellia128 is available and Camellia256 isn't).
9333 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9334 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9335 When a point or a seed is encoded in a BIT STRING, we need to
9336 prevent the removal of trailing zero bits to get the proper DER
9337 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9338 of a NamedBitList, for which trailing 0 bits need to be removed.)
9342 * Have SSL/TLS server implementation tolerate "mismatched" record
9343 protocol version while receiving ClientHello even if the
9344 ClientHello is fragmented. (The server can't insist on the
9345 particular protocol version it has chosen before the ServerHello
9346 message has informed the client about his choice.)
9350 * Add RFC 3779 support.
9352 *Rob Austein for ARIN, Ben Laurie*
9354 * Load error codes if they are not already present instead of using a
9355 static variable. This allows them to be cleanly unloaded and reloaded.
9356 Improve header file function name parsing.
9360 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9361 or CAPABILITY handshake as required by RFCs.
9365 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9367 * Introduce limits to prevent malicious keys being able to
9368 cause a denial of service. [CVE-2006-2940][]
9370 *Steve Henson, Bodo Moeller*
9372 * Fix ASN.1 parsing of certain invalid structures that can result
9373 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9375 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9376 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9378 * Fix SSL client code which could crash if connecting to a
9379 malicious SSLv2 server. [CVE-2006-4343][]
9381 *Tavis Ormandy and Will Drewry, Google Security Team*
9383 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9384 match only those. Before that, "AES256-SHA" would be interpreted
9385 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9386 the same strength classification in 0.9.7h) as we currently only
9387 have a single AES bit in the ciphersuite description bitmap.
9388 That change, however, also applied to ciphersuite strings such as
9389 "RC4-MD5" that intentionally matched multiple ciphersuites --
9390 namely, SSL 2.0 ciphersuites in addition to the more common ones
9391 from SSL 3.0/TLS 1.0.
9393 So we change the selection algorithm again: Naming an explicit
9394 ciphersuite selects this one ciphersuite, and any other similar
9395 ciphersuite (same bitmap) from *other* protocol versions.
9396 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9397 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9399 Since SSL 2.0 does not have any ciphersuites for which the
9400 128/256 bit distinction would be relevant, this works for now.
9401 The proper fix will be to use different bits for AES128 and
9402 AES256, which would have avoided the problems from the beginning;
9403 however, bits are scarce, so we can only do this in a new release
9404 (not just a patchlevel) when we can change the SSL_CIPHER
9405 definition to split the single 'unsigned long mask' bitmap into
9406 multiple values to extend the available space.
9410 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9412 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9413 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9415 * Add AES IGE and biIGE modes.
9419 * Change the Unix randomness entropy gathering to use poll() when
9420 possible instead of select(), since the latter has some
9421 undesirable limitations.
9423 *Darryl Miles via Richard Levitte and Bodo Moeller*
9425 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9426 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9427 cannot be implicitly activated as part of, e.g., the "AES" alias.
9428 However, please upgrade to OpenSSL 0.9.9[-dev] for
9429 non-experimental use of the ECC ciphersuites to get TLS extension
9430 support, which is required for curve and point format negotiation
9431 to avoid potential handshake problems.
9435 * Disable rogue ciphersuites:
9437 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9438 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9439 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9441 The latter two were purportedly from
9442 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9445 Also deactivate the remaining ciphersuites from
9446 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9447 unofficial, and the ID has long expired.
9451 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9452 dual-core machines) and other potential thread-safety issues.
9456 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9457 versions), which is now available for royalty-free use
9458 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9459 Also, add Camellia TLS ciphersuites from RFC 4132.
9461 To minimize changes between patchlevels in the OpenSSL 0.9.8
9462 series, Camellia remains excluded from compilation unless OpenSSL
9463 is configured with 'enable-camellia'.
9467 * Disable the padding bug check when compression is in use. The padding
9468 bug check assumes the first packet is of even length, this is not
9469 necessarily true if compression is enabled and can result in false
9470 positives causing handshake failure. The actual bug test is ancient
9471 code so it is hoped that implementations will either have fixed it by
9472 now or any which still have the bug do not support compression.
9476 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9478 * When applying a cipher rule check to see if string match is an explicit
9479 cipher suite and only match that one cipher suite if it is.
9483 * Link in manifests for VC++ if needed.
9485 *Austin Ziegler <halostatue@gmail.com>*
9487 * Update support for ECC-based TLS ciphersuites according to
9488 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9489 TLS extensions, which are supported starting with the 0.9.9
9490 branch, not in the OpenSSL 0.9.8 branch).
9494 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9495 opaque EVP_CIPHER_CTX handling.
9499 * Fixes and enhancements to zlib compression code. We now only use
9500 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9501 to conform with the standards mentioned here:
9502 <http://www.zlib.net/DLL_FAQ.txt>
9503 Static zlib linking now works on Windows and the new --with-zlib-include
9504 --with-zlib-lib options to Configure can be used to supply the location
9505 of the headers and library. Gracefully handle case where zlib library
9510 * Several fixes and enhancements to the OID generation code. The old code
9511 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9512 handle numbers larger than ULONG_MAX, truncated printing and had a
9513 non standard OBJ_obj2txt() behaviour.
9517 * Add support for building of engines under engine/ as shared libraries
9518 under VC++ build system.
9522 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9523 Hopefully, we will not see any false combination of paths any more.
9527 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9529 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9530 (part of SSL_OP_ALL). This option used to disable the
9531 countermeasure against man-in-the-middle protocol-version
9532 rollback in the SSL 2.0 server implementation, which is a bad
9533 idea. [CVE-2005-2969][]
9535 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9536 for Information Security, National Institute of Advanced Industrial
9537 Science and Technology [AIST], Japan)*
9539 * Add two function to clear and return the verify parameter flags.
9543 * Keep cipherlists sorted in the source instead of sorting them at
9544 runtime, thus removing the need for a lock.
9548 * Avoid some small subgroup attacks in Diffie-Hellman.
9550 *Nick Mathewson and Ben Laurie*
9552 * Add functions for well-known primes.
9556 * Extended Windows CE support.
9558 *Satoshi Nakamura and Andy Polyakov*
9560 * Initialize SSL_METHOD structures at compile time instead of during
9561 runtime, thus removing the need for a lock.
9565 * Make PKCS7_decrypt() work even if no certificate is supplied by
9566 attempting to decrypt each encrypted key in turn. Add support to
9571 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9573 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9576 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9580 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9581 key into the same file any more.
9585 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9589 * Add -utf8 command line and config file option to 'ca'.
9591 *Stefan <stf@udoma.org*
9593 * Removed the macro des_crypt(), as it seems to conflict with some
9594 libraries. Use DES_crypt().
9598 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9599 involves renaming the source and generated shared-libs for
9600 both. The engines will accept the corrected or legacy ids
9601 ('ncipher' and '4758_cca' respectively) when binding. NB,
9602 this only applies when building 'shared'.
9604 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9606 * Add attribute functions to EVP_PKEY structure. Modify
9607 PKCS12_create() to recognize a CSP name attribute and
9608 use it. Make -CSP option work again in pkcs12 utility.
9612 * Add new functionality to the bn blinding code:
9613 - automatic re-creation of the BN_BLINDING parameters after
9614 a fixed number of uses (currently 32)
9615 - add new function for parameter creation
9616 - introduce flags to control the update behaviour of the
9617 BN_BLINDING parameters
9618 - hide BN_BLINDING structure
9619 Add a second BN_BLINDING slot to the RSA structure to improve
9620 performance when a single RSA object is shared among several
9625 * Add support for DTLS.
9627 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9629 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9630 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9634 * Remove buggy and incomplete DH cert support from
9635 ssl/ssl_rsa.c and ssl/s3_both.c
9639 * Use SHA-1 instead of MD5 as the default digest algorithm for
9640 the apps/openssl applications.
9644 * Compile clean with "-Wall -Wmissing-prototypes
9645 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9646 DEBUG_SAFESTACK must also be set.
9650 * Change ./Configure so that certain algorithms can be disabled by default.
9651 The new counterpiece to "no-xxx" is "enable-xxx".
9653 The patented RC5 and MDC2 algorithms will now be disabled unless
9654 "enable-rc5" and "enable-mdc2", respectively, are specified.
9656 (IDEA remains enabled despite being patented. This is because IDEA
9657 is frequently required for interoperability, and there is no license
9658 fee for non-commercial use. As before, "no-idea" can be used to
9659 avoid this algorithm.)
9663 * Add processing of proxy certificates (see RFC 3820). This work was
9664 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9665 EGEE (Enabling Grids for E-science in Europe).
9669 * RC4 performance overhaul on modern architectures/implementations, such
9670 as Intel P4, IA-64 and AMD64.
9674 * New utility extract-section.pl. This can be used specify an alternative
9675 section number in a pod file instead of having to treat each file as
9676 a separate case in Makefile. This can be done by adding two lines to the
9679 =for comment openssl_section:XXX
9681 The blank line is mandatory.
9685 * New arguments -certform, -keyform and -pass for s_client and s_server
9686 to allow alternative format key and certificate files and passphrase
9691 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9692 update associated structures and add various utility functions.
9694 Add new policy related verify parameters, include policy checking in
9695 standard verify code. Enhance 'smime' application with extra parameters
9696 to support policy checking and print out.
9700 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9701 Nehemiah processors. These extensions support AES encryption in hardware
9702 as well as RNG (though RNG support is currently disabled).
9704 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9706 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9710 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9712 *Andy Polyakov and a number of other people*
9714 * Improved PowerPC platform support. Most notably BIGNUM assembler
9715 implementation contributed by IBM.
9717 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9719 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9720 exponent rather than 'unsigned long'. There is a corresponding change to
9721 the new 'rsa_keygen' element of the RSA_METHOD structure.
9723 *Jelte Jansen, Geoff Thorpe*
9725 * Functionality for creating the initial serial number file is now
9726 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9728 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9729 number file to 1, which is bound to cause problems. To avoid
9730 the problems while respecting compatibility between different 0.9.7
9731 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9732 CA.pl for serial number initialization. With the new release 0.9.8,
9733 we can fix the problem directly in the 'ca' utility.)
9737 * Reduced header interdependencies by declaring more opaque objects in
9738 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9739 give fewer recursive includes, which could break lazy source code - so
9740 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9741 developers should define this symbol when building and using openssl to
9742 ensure they track the recommended behaviour, interfaces, [etc], but
9743 backwards-compatible behaviour prevails when this isn't defined.
9747 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9751 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9752 This will generate a random key of the appropriate length based on the
9753 cipher context. The EVP_CIPHER can provide its own random key generation
9754 routine to support keys of a specific form. This is used in the des and
9755 3des routines to generate a key of the correct parity. Update S/MIME
9756 code to use new functions and hence generate correct parity DES keys.
9757 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9758 valid (weak or incorrect parity).
9762 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9763 as looking them up. This is useful when the verified structure may contain
9764 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9765 present unless the new PKCS7_NO_CRL flag is asserted.
9769 * Extend ASN1 oid configuration module. It now additionally accepts the
9772 shortName = some long name, 1.2.3.4
9776 * Reimplemented the BN_CTX implementation. There is now no more static
9777 limitation on the number of variables it can handle nor the depth of the
9778 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9779 information can now expand as required, and rather than having a single
9780 static array of bignums, BN_CTX now uses a linked-list of such arrays
9781 allowing it to expand on demand whilst maintaining the usefulness of
9782 BN_CTX's "bundling".
9786 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9787 to allow all RSA operations to function using a single BN_CTX.
9791 * Preliminary support for certificate policy evaluation and checking. This
9792 is initially intended to pass the tests outlined in "Conformance Testing
9793 of Relying Party Client Certificate Path Processing Logic" v1.07.
9797 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9798 remained unused and not that useful. A variety of other little bignum
9799 tweaks and fixes have also been made continuing on from the audit (see
9804 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9805 associated ASN1, EVP and SSL functions and old ASN1 macros.
9809 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9810 and this should never fail. So the return value from the use of
9811 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9812 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9816 * BN_CTX_get() should return zero-valued bignums, providing the same
9817 initialised value as BN_new().
9819 *Geoff Thorpe, suggested by Ulf Möller*
9821 * Support for inhibitAnyPolicy certificate extension.
9825 * An audit of the BIGNUM code is underway, for which debugging code is
9826 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9827 is considered valid when processing BIGNUMs, and causes execution to
9828 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9829 further steps are taken to deliberately pollute unused data in BIGNUM
9830 structures to try and expose faulty code further on. For now, openssl will
9831 (in its default mode of operation) continue to tolerate the inconsistent
9832 forms that it has tolerated in the past, but authors and packagers should
9833 consider trying openssl and their own applications when compiled with
9834 these debugging symbols defined. It will help highlight potential bugs in
9835 their own code, and will improve the test coverage for OpenSSL itself. At
9836 some point, these tighter rules will become openssl's default to improve
9837 maintainability, though the assert()s and other overheads will remain only
9838 in debugging configurations. See bn.h for more details.
9840 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9842 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9843 that can only be obtained through BN_CTX_new() (which implicitly
9844 initialises it). The presence of this function only made it possible
9845 to overwrite an existing structure (and cause memory leaks).
9849 * Because of the callback-based approach for implementing LHASH as a
9850 template type, lh_insert() adds opaque objects to hash-tables and
9851 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9852 to clean up those corresponding objects before destroying the hash table
9853 (and losing the object pointers). So some over-zealous constifications in
9854 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9855 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9856 prototyped to have "const" restrictions on the object pointers they are
9857 given (and so aren't required to cast them away any more).
9861 * The tmdiff.h API was so ugly and minimal that our own timing utility
9862 (speed) prefers to use its own implementation. The two implementations
9863 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9864 its object type properly exposed (MS_TM) instead of casting to/from
9865 `char *`. This may still change yet if someone realises MS_TM and
9867 aren't necessarily the greatest nomenclatures - but this is what was used
9868 internally to the implementation so I've used that for now.
9872 * Ensure that deprecated functions do not get compiled when
9873 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9874 the self-tests were still using deprecated key-generation functions so
9875 these have been updated also.
9879 * Reorganise PKCS#7 code to separate the digest location functionality
9880 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9881 New function PKCS7_set_digest() to set the digest type for PKCS#7
9882 digestedData type. Add additional code to correctly generate the
9883 digestedData type and add support for this type in PKCS7 initialization
9888 * New function PKCS7_set0_type_other() this initializes a PKCS7
9889 structure of type "other".
9893 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9894 sure the loop does correctly stop and breaking ("division by zero")
9895 modulus operations are not performed. The (pre-generated) prime
9896 table crypto/bn/bn_prime.h was already correct, but it could not be
9897 re-generated on some platforms because of the "division by zero"
9898 situation in the script.
9900 *Ralf S. Engelschall*
9902 * Update support for ECC-based TLS ciphersuites according to
9903 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9904 SHA-1 now is only used for "small" curves (where the
9905 representation of a field element takes up to 24 bytes); for
9906 larger curves, the field element resulting from ECDH is directly
9907 used as premaster secret.
9909 *Douglas Stebila (Sun Microsystems Laboratories)*
9911 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9912 curve secp160r1 to the tests.
9914 *Douglas Stebila (Sun Microsystems Laboratories)*
9916 * Add the possibility to load symbols globally with DSO.
9918 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9920 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9921 control of the error stack.
9925 * Add support for STORE in ENGINE.
9929 * Add the STORE type. The intention is to provide a common interface
9930 to certificate and key stores, be they simple file-based stores, or
9931 HSM-type store, or LDAP stores, or...
9932 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9936 * Add a generic structure called OPENSSL_ITEM. This can be used to
9937 pass a list of arguments to any function as well as provide a way
9938 for a function to pass data back to the caller.
9942 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9943 works like BUF_strdup() but can be used to duplicate a portion of
9944 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9949 * Add the function sk_find_ex() which works like sk_find(), but will
9950 return an index to an element even if an exact match couldn't be
9951 found. The index is guaranteed to point at the element where the
9952 searched-for key would be inserted to preserve sorting order.
9956 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9957 takes an extra flags argument for optional functionality. Currently,
9958 the following flags are defined:
9960 OBJ_BSEARCH_VALUE_ON_NOMATCH
9961 This one gets OBJ_bsearch_ex() to return a pointer to the first
9962 element where the comparing function returns a negative or zero
9965 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9966 This one gets OBJ_bsearch_ex() to return a pointer to the first
9967 element where the comparing function returns zero. This is useful
9968 if there are more than one element where the comparing function
9973 * Make it possible to create self-signed certificates with 'openssl ca'
9974 in such a way that the self-signed certificate becomes part of the
9975 CA database and uses the same mechanisms for serial number generation
9976 as all other certificate signing. The new flag '-selfsign' enables
9977 this functionality. Adapt CA.sh and CA.pl.in.
9981 * Add functionality to check the public key of a certificate request
9982 against a given private. This is useful to check that a certificate
9983 request can be signed by that key (self-signing).
9987 * Make it possible to have multiple active certificates with the same
9988 subject in the CA index file. This is done only if the keyword
9989 'unique_subject' is set to 'no' in the main CA section (default
9990 if 'CA_default') of the configuration file. The value is saved
9991 with the database itself in a separate index attribute file,
9992 named like the index file with '.attr' appended to the name.
9996 * Generate multi-valued AVAs using '+' notation in config files for
10001 * Support for nameConstraints certificate extension.
10005 * Support for policyConstraints certificate extension.
10009 * Support for policyMappings certificate extension.
10013 * Make sure the default DSA_METHOD implementation only uses its
10014 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10015 and change its own handlers to be NULL so as to remove unnecessary
10016 indirection. This lets alternative implementations fallback to the
10017 default implementation more easily.
10021 * Support for directoryName in GeneralName related extensions
10026 * Make it possible to link applications using Makefile.shared.
10027 Make that possible even when linking against static libraries!
10031 * Support for single pass processing for S/MIME signing. This now
10032 means that S/MIME signing can be done from a pipe, in addition
10033 cleartext signing (multipart/signed type) is effectively streaming
10034 and the signed data does not need to be all held in memory.
10036 This is done with a new flag PKCS7_STREAM. When this flag is set
10037 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10038 is done after the data is output (and digests calculated) in
10039 SMIME_write_PKCS7().
10043 * Add full support for -rpath/-R, both in shared libraries and
10044 applications, at least on the platforms where it's known how
10049 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10050 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10051 will now compute a table of multiples of the generator that
10052 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10053 faster (notably in the case of a single point multiplication,
10054 scalar * generator).
10056 *Nils Larsch, Bodo Moeller*
10058 * IPv6 support for certificate extensions. The various extensions
10059 which use the IP:a.b.c.d can now take IPv6 addresses using the
10060 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10065 * Added an ENGINE that implements RSA by performing private key
10066 exponentiations with the GMP library. The conversions to and from
10067 GMP's mpz_t format aren't optimised nor are any montgomery forms
10068 cached, and on x86 it appears OpenSSL's own performance has caught up.
10069 However there are likely to be other architectures where GMP could
10070 provide a boost. This ENGINE is not built in by default, but it can be
10071 specified at Configure time and should be accompanied by the necessary
10072 linker additions, eg;
10073 ./config -DOPENSSL_USE_GMP -lgmp
10077 * "openssl engine" will not display ENGINE/DSO load failure errors when
10078 testing availability of engines with "-t" - the old behaviour is
10079 produced by increasing the feature's verbosity with "-tt".
10083 * ECDSA routines: under certain error conditions uninitialized BN objects
10084 could be freed. Solution: make sure initialization is performed early
10085 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10090 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10091 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10092 software implementations. For DSA and DH, parameter generation can
10093 also be overridden by providing the appropriate method callbacks.
10097 * Change the "progress" mechanism used in key-generation and
10098 primality testing to functions that take a new BN_GENCB pointer in
10099 place of callback/argument pairs. The new API functions have `_ex`
10100 postfixes and the older functions are reimplemented as wrappers for
10101 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10102 declarations of the old functions to help (graceful) attempts to
10103 migrate to the new functions. Also, the new key-generation API
10104 functions operate on a caller-supplied key-structure and return
10105 success/failure rather than returning a key or NULL - this is to
10106 help make "keygen" another member function of RSA_METHOD etc.
10108 Example for using the new callback interface:
10110 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10111 void *my_arg = ...;
10114 BN_GENCB_set(&my_cb, my_callback, my_arg);
10116 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10117 /* For the meaning of a, b in calls to my_callback(), see the
10118 * documentation of the function that calls the callback.
10119 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10120 * my_callback should return 1 if it wants BN_is_prime_ex()
10121 * to continue, or 0 to stop.
10126 * Change the ZLIB compression method to be stateful, and make it
10127 available to TLS with the number defined in
10128 draft-ietf-tls-compression-04.txt.
10132 * Add the ASN.1 structures and functions for CertificatePair, which
10133 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10135 CertificatePair ::= SEQUENCE {
10136 forward [0] Certificate OPTIONAL,
10137 reverse [1] Certificate OPTIONAL,
10138 -- at least one of the pair shall be present -- }
10140 Also implement the PEM functions to read and write certificate
10141 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10143 This needed to be defined, mostly for the sake of the LDAP
10144 attribute crossCertificatePair, but may prove useful elsewhere as
10149 * Make it possible to inhibit symlinking of shared libraries in
10150 Makefile.shared, for Cygwin's sake.
10154 * Extend the BIGNUM API by creating a function
10155 void BN_set_negative(BIGNUM *a, int neg);
10156 and a macro that behave like
10157 int BN_is_negative(const BIGNUM *a);
10159 to avoid the need to access 'a->neg' directly in applications.
10163 * Implement fast modular reduction for pseudo-Mersenne primes
10164 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10165 EC_GROUP_new_curve_GFp() will now automatically use this
10168 *Nils Larsch <nla@trustcenter.de>*
10170 * Add new lock type (CRYPTO_LOCK_BN).
10174 * Change the ENGINE framework to automatically load engines
10175 dynamically from specific directories unless they could be
10176 found to already be built in or loaded. Move all the
10177 current engines except for the cryptodev one to a new
10178 directory engines/.
10179 The engines in engines/ are built as shared libraries if
10180 the "shared" options was given to ./Configure or ./config.
10181 Otherwise, they are inserted in libcrypto.a.
10182 /usr/local/ssl/engines is the default directory for dynamic
10183 engines, but that can be overridden at configure time through
10184 the usual use of --prefix and/or --openssldir, and at run
10185 time with the environment variable OPENSSL_ENGINES.
10187 *Geoff Thorpe and Richard Levitte*
10189 * Add Makefile.shared, a helper makefile to build shared
10190 libraries. Adapt Makefile.org.
10194 * Add version info to Win32 DLLs.
10196 *Peter 'Luna' Runestig" <peter@runestig.com>*
10198 * Add new 'medium level' PKCS#12 API. Certificates and keys
10199 can be added using this API to created arbitrary PKCS#12
10200 files while avoiding the low level API.
10202 New options to PKCS12_create(), key or cert can be NULL and
10203 will then be omitted from the output file. The encryption
10204 algorithm NIDs can be set to -1 for no encryption, the mac
10205 iteration count can be set to 0 to omit the mac.
10207 Enhance pkcs12 utility by making the -nokeys and -nocerts
10208 options work when creating a PKCS#12 file. New option -nomac
10209 to omit the mac, NONE can be set for an encryption algorithm.
10210 New code is modified to use the enhanced PKCS12_create()
10211 instead of the low level API.
10215 * Extend ASN1 encoder to support indefinite length constructed
10216 encoding. This can output sequences tags and octet strings in
10217 this form. Modify pk7_asn1.c to support indefinite length
10218 encoding. This is experimental and needs additional code to
10219 be useful, such as an ASN1 bio and some enhanced streaming
10222 Extend template encode functionality so that tagging is passed
10223 down to the template encoder.
10227 * Let 'openssl req' fail if an argument to '-newkey' is not
10228 recognized instead of using RSA as a default.
10232 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10233 As these are not official, they are not included in "ALL";
10234 the "ECCdraft" ciphersuite group alias can be used to select them.
10236 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10238 * Add ECDH engine support.
10240 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10242 * Add ECDH in new directory crypto/ecdh/.
10244 *Douglas Stebila (Sun Microsystems Laboratories)*
10246 * Let BN_rand_range() abort with an error after 100 iterations
10247 without success (which indicates a broken PRNG).
10251 * Change BN_mod_sqrt() so that it verifies that the input value
10252 is really the square of the return value. (Previously,
10253 BN_mod_sqrt would show GIGO behaviour.)
10257 * Add named elliptic curves over binary fields from X9.62, SECG,
10258 and WAP/WTLS; add OIDs that were still missing.
10260 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10262 * Extend the EC library for elliptic curves over binary fields
10263 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10266 EC_GF2m_simple_method
10270 EC_GROUP_new_curve_GF2m
10271 EC_GROUP_set_curve_GF2m
10272 EC_GROUP_get_curve_GF2m
10273 EC_POINT_set_affine_coordinates_GF2m
10274 EC_POINT_get_affine_coordinates_GF2m
10275 EC_POINT_set_compressed_coordinates_GF2m
10277 Point compression for binary fields is disabled by default for
10278 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10281 As binary polynomials are represented as BIGNUMs, various members
10282 of the EC_GROUP and EC_POINT data structures can be shared
10283 between the implementations for prime fields and binary fields;
10284 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10285 are essentially identical to their `..._GFp` counterparts.
10286 (For simplicity, the `..._GFp` prefix has been dropped from
10287 various internal method names.)
10289 An internal 'field_div' method (similar to 'field_mul' and
10290 'field_sqr') has been added; this is used only for binary fields.
10292 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10294 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10295 through methods ('mul', 'precompute_mult').
10297 The generic implementations (now internally called 'ec_wNAF_mul'
10298 and 'ec_wNAF_precomputed_mult') remain the default if these
10299 methods are undefined.
10301 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10303 * New function EC_GROUP_get_degree, which is defined through
10304 EC_METHOD. For curves over prime fields, this returns the bit
10305 length of the modulus.
10307 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10309 * New functions EC_GROUP_dup, EC_POINT_dup.
10310 (These simply call ..._new and ..._copy).
10312 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10314 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10315 Polynomials are represented as BIGNUMs (where the sign bit is not
10316 used) in the following functions [macros]:
10319 BN_GF2m_sub [= BN_GF2m_add]
10320 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10321 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10322 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10324 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10325 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10326 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10327 BN_GF2m_cmp [= BN_ucmp]
10329 (Note that only the 'mod' functions are actually for fields GF(2^m).
10330 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10332 For some functions, an the irreducible polynomial defining a
10333 field can be given as an 'unsigned int[]' with strictly
10334 decreasing elements giving the indices of those bits that are set;
10335 i.e., p[] represents the polynomial
10336 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10338 p[0] > p[1] > ... > p[k] = 0.
10339 This applies to the following functions:
10342 BN_GF2m_mod_mul_arr
10343 BN_GF2m_mod_sqr_arr
10344 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10345 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10346 BN_GF2m_mod_exp_arr
10347 BN_GF2m_mod_sqrt_arr
10348 BN_GF2m_mod_solve_quad_arr
10352 Conversion can be performed by the following functions:
10357 bntest.c has additional tests for binary polynomial arithmetic.
10359 Two implementations for BN_GF2m_mod_div() are available.
10360 The default algorithm simply uses BN_GF2m_mod_inv() and
10361 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10362 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10363 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10365 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10367 * Add new error code 'ERR_R_DISABLED' that can be used when some
10368 functionality is disabled at compile-time.
10370 *Douglas Stebila <douglas.stebila@sun.com>*
10372 * Change default behaviour of 'openssl asn1parse' so that more
10373 information is visible when viewing, e.g., a certificate:
10375 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10376 mode the content of non-printable OCTET STRINGs is output in a
10377 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10378 avoid the appearance of a printable string.
10380 *Nils Larsch <nla@trustcenter.de>*
10382 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10384 EC_GROUP_set_asn1_flag()
10385 EC_GROUP_get_asn1_flag()
10386 EC_GROUP_set_point_conversion_form()
10387 EC_GROUP_get_point_conversion_form()
10388 These control ASN1 encoding details:
10389 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10390 has been set to OPENSSL_EC_NAMED_CURVE.
10391 - Points are encoded in uncompressed form by default; options for
10392 asn1_for are as for point2oct, namely
10393 POINT_CONVERSION_COMPRESSED
10394 POINT_CONVERSION_UNCOMPRESSED
10395 POINT_CONVERSION_HYBRID
10397 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10399 EC_GROUP_set_seed()
10400 EC_GROUP_get0_seed()
10401 EC_GROUP_get_seed_len()
10402 This is used only for ASN1 purposes (so far).
10404 *Nils Larsch <nla@trustcenter.de>*
10406 * Add 'field_type' member to EC_METHOD, which holds the NID
10407 of the appropriate field type OID. The new function
10408 EC_METHOD_get_field_type() returns this value.
10410 *Nils Larsch <nla@trustcenter.de>*
10413 EC_POINT_point2bn()
10414 EC_POINT_bn2point()
10415 EC_POINT_point2hex()
10416 EC_POINT_hex2point()
10417 providing useful interfaces to EC_POINT_point2oct() and
10418 EC_POINT_oct2point().
10420 *Nils Larsch <nla@trustcenter.de>*
10422 * Change internals of the EC library so that the functions
10423 EC_GROUP_set_generator()
10424 EC_GROUP_get_generator()
10425 EC_GROUP_get_order()
10426 EC_GROUP_get_cofactor()
10427 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10428 to methods, which would lead to unnecessary code duplication when
10429 adding different types of curves.
10431 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10433 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10434 arithmetic, and such that modified wNAFs are generated
10435 (which avoid length expansion in many cases).
10439 * Add a function EC_GROUP_check_discriminant() (defined via
10440 EC_METHOD) that verifies that the curve discriminant is non-zero.
10442 Add a function EC_GROUP_check() that makes some sanity tests
10443 on a EC_GROUP, its generator and order. This includes
10444 EC_GROUP_check_discriminant().
10446 *Nils Larsch <nla@trustcenter.de>*
10448 * Add ECDSA in new directory crypto/ecdsa/.
10450 Add applications 'openssl ecparam' and 'openssl ecdsa'
10451 (these are based on 'openssl dsaparam' and 'openssl dsa').
10453 ECDSA support is also included in various other files across the
10454 library. Most notably,
10455 - 'openssl req' now has a '-newkey ecdsa:file' option;
10456 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10457 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10458 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10459 them suitable for ECDSA where domain parameters must be
10460 extracted before the specific public key;
10461 - ECDSA engine support has been added.
10463 *Nils Larsch <nla@trustcenter.de>*
10465 * Include some named elliptic curves, and add OIDs from X9.62,
10466 SECG, and WAP/WTLS. Each curve can be obtained from the new
10468 EC_GROUP_new_by_curve_name(),
10469 and the list of available named curves can be obtained with
10470 EC_get_builtin_curves().
10471 Also add a 'curve_name' member to EC_GROUP objects, which can be
10473 EC_GROUP_set_curve_name()
10474 EC_GROUP_get_curve_name()
10476 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10478 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10479 was actually never needed) and in BN_mul(). The removal in BN_mul()
10480 required a small change in bn_mul_part_recursive() and the addition
10481 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10482 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10483 bn_sub_words() and bn_add_words() except they take arrays with
10488 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10490 * Cleanse PEM buffers before freeing them since they may contain
10493 *Benjamin Bennett <ben@psc.edu>*
10495 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10496 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10497 authentication-only ciphersuites.
10501 * Since AES128 and AES256 share a single mask bit in the logic of
10502 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10503 kludge to work properly if AES128 is available and AES256 isn't.
10507 * Expand security boundary to match 1.1.1 module.
10511 * Remove redundant features: hash file source, editing of test vectors
10512 modify fipsld to use external fips_premain.c signature.
10516 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10517 run algorithm test programs.
10521 * Make algorithm test programs more tolerant of whitespace.
10525 * Have SSL/TLS server implementation tolerate "mismatched" record
10526 protocol version while receiving ClientHello even if the
10527 ClientHello is fragmented. (The server can't insist on the
10528 particular protocol version it has chosen before the ServerHello
10529 message has informed the client about his choice.)
10533 * Load error codes if they are not already present instead of using a
10534 static variable. This allows them to be cleanly unloaded and reloaded.
10538 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10540 * Introduce limits to prevent malicious keys being able to
10541 cause a denial of service. [CVE-2006-2940][]
10543 *Steve Henson, Bodo Moeller*
10545 * Fix ASN.1 parsing of certain invalid structures that can result
10546 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10548 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10549 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10551 * Fix SSL client code which could crash if connecting to a
10552 malicious SSLv2 server. [CVE-2006-4343][]
10554 *Tavis Ormandy and Will Drewry, Google Security Team*
10556 * Change ciphersuite string processing so that an explicit
10557 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10558 will no longer include "AES128-SHA"), and any other similar
10559 ciphersuite (same bitmap) from *other* protocol versions (so that
10560 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10561 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10562 changes from 0.9.8b and 0.9.8d.
10566 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10568 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10569 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10571 * Change the Unix randomness entropy gathering to use poll() when
10572 possible instead of select(), since the latter has some
10573 undesirable limitations.
10575 *Darryl Miles via Richard Levitte and Bodo Moeller*
10577 * Disable rogue ciphersuites:
10579 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10580 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10581 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10583 The latter two were purportedly from
10584 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10587 Also deactivate the remaining ciphersuites from
10588 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10589 unofficial, and the ID has long expired.
10593 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10594 dual-core machines) and other potential thread-safety issues.
10598 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10600 * Adapt fipsld and the build system to link against the validated FIPS
10601 module in FIPS mode.
10605 * Fixes for VC++ 2005 build under Windows.
10609 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10610 from a Windows bash shell such as MSYS. It is autodetected from the
10611 "config" script when run from a VC++ environment. Modify standard VC++
10612 build to use fipscanister.o from the GNU make build.
10616 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10618 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10619 The value now differs depending on if you build for FIPS or not.
10620 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10621 safely run with a non-FIPSed libcrypto, as it may crash because of
10622 the difference induced by this change.
10626 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10628 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10629 (part of SSL_OP_ALL). This option used to disable the
10630 countermeasure against man-in-the-middle protocol-version
10631 rollback in the SSL 2.0 server implementation, which is a bad
10632 idea. [CVE-2005-2969][]
10634 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10635 for Information Security, National Institute of Advanced Industrial
10636 Science and Technology [AIST, Japan)]*
10638 * Minimal support for X9.31 signatures and PSS padding modes. This is
10639 mainly for FIPS compliance and not fully integrated at this stage.
10643 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10644 the exponentiation using a fixed-length exponent. (Otherwise,
10645 the information leaked through timing could expose the secret key
10646 after many signatures; cf. Bleichenbacher's attack on DSA with
10651 * Make a new fixed-window mod_exp implementation the default for
10652 RSA, DSA, and DH private-key operations so that the sequence of
10653 squares and multiplies and the memory access pattern are
10654 independent of the particular secret key. This will mitigate
10655 cache-timing and potential related attacks.
10657 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10658 and this is automatically used by BN_mod_exp_mont() if the new flag
10659 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10660 will use this BN flag for private exponents unless the flag
10661 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10662 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10664 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10666 * Change the client implementation for SSLv23_method() and
10667 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10668 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10669 (Previously, the SSL 2.0 backwards compatible Client Hello
10670 message format would be used even with SSL_OP_NO_SSLv2.)
10674 * Add support for smime-type MIME parameter in S/MIME messages which some
10679 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10680 a threadsafe manner. Modify rsa code to use new function and add calls
10681 to dsa and dh code (which had race conditions before).
10685 * Include the fixed error library code in the C error file definitions
10686 instead of fixing them up at runtime. This keeps the error code
10687 structures constant.
10691 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10693 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10696 * Fixes for newer kerberos headers. NB: the casts are needed because
10697 the 'length' field is signed on one version and unsigned on another
10698 with no (?) obvious way to tell the difference, without these VC++
10699 complains. Also the "definition" of FAR (blank) is no longer included
10700 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10701 some needed definitions.
10705 * Undo Cygwin change.
10709 * Added support for proxy certificates according to RFC 3820.
10710 Because they may be a security thread to unaware applications,
10711 they must be explicitly allowed in run-time. See
10712 docs/HOWTO/proxy_certificates.txt for further information.
10716 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10718 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10719 server and client random values. Previously
10720 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10721 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10723 This change has negligible security impact because:
10725 1. Server and client random values still have 24 bytes of pseudo random
10728 2. Server and client random values are sent in the clear in the initial
10731 3. The master secret is derived using the premaster secret (48 bytes in
10732 size for static RSA ciphersuites) as well as client server and random
10735 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10738 *Stephen Henson, reported by UK NISCC*
10740 * Use Windows randomness collection on Cygwin.
10744 * Fix hang in EGD/PRNGD query when communication socket is closed
10745 prematurely by EGD/PRNGD.
10747 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10749 * Prompt for pass phrases when appropriate for PKCS12 input format.
10753 * Back-port of selected performance improvements from development
10754 branch, as well as improved support for PowerPC platforms.
10758 * Add lots of checks for memory allocation failure, error codes to indicate
10759 failure and freeing up memory if a failure occurs.
10761 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10763 * Add new -passin argument to dgst.
10767 * Perform some character comparisons of different types in X509_NAME_cmp:
10768 this is needed for some certificates that re-encode DNs into UTF8Strings
10769 (in violation of RFC3280) and can't or won't issue name rollover
10774 * Make an explicit check during certificate validation to see that
10775 the CA setting in each certificate on the chain is correct. As a
10776 side effect always do the following basic checks on extensions,
10777 not just when there's an associated purpose to the check:
10779 - if there is an unhandled critical extension (unless the user
10780 has chosen to ignore this fault)
10781 - if the path length has been exceeded (if one is set at all)
10782 - that certain extensions fit the associated purpose (if one has
10787 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10789 * Avoid a race condition when CRLs are checked in a multi threaded
10790 environment. This would happen due to the reordering of the revoked
10791 entries during signature checking and serial number lookup. Now the
10792 encoding is cached and the serial number sort performed under a lock.
10793 Add new STACK function sk_is_sorted().
10797 * Add Delta CRL to the extension code.
10801 * Various fixes to s3_pkt.c so alerts are sent properly.
10803 *David Holmes <d.holmes@f5.com>*
10805 * Reduce the chances of duplicate issuer name and serial numbers (in
10806 violation of RFC3280) using the OpenSSL certificate creation utilities.
10807 This is done by creating a random 64 bit value for the initial serial
10808 number when a serial number file is created or when a self signed
10809 certificate is created using 'openssl req -x509'. The initial serial
10810 number file is created using 'openssl x509 -next_serial' in CA.pl
10811 rather than being initialized to 1.
10815 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10817 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10818 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10820 *Joe Orton, Steve Henson*
10822 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10825 *Joe Orton, Steve Henson*
10827 * Make it possible to have multiple active certificates with the same
10828 subject in the CA index file. This is done only if the keyword
10829 'unique_subject' is set to 'no' in the main CA section (default
10830 if 'CA_default') of the configuration file. The value is saved
10831 with the database itself in a separate index attribute file,
10832 named like the index file with '.attr' appended to the name.
10836 * X509 verify fixes. Disable broken certificate workarounds when
10837 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10838 keyUsage extension present. Don't accept CRLs with unhandled critical
10839 extensions: since verify currently doesn't process CRL extensions this
10840 rejects a CRL with *any* critical extensions. Add new verify error codes
10845 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10846 A clarification of RFC2560 will require the use of OCTET STRINGs and
10847 some implementations cannot handle the current raw format. Since OpenSSL
10848 copies and compares OCSP nonces as opaque blobs without any attempt at
10849 parsing them this should not create any compatibility issues.
10853 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10854 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10855 this HMAC (and other) operations are several times slower than OpenSSL
10860 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10862 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10864 * Use the correct content when signing type "other".
10868 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10870 * Fix various bugs revealed by running the NISCC test suite:
10872 Stop out of bounds reads in the ASN1 code when presented with
10873 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10875 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10877 If verify callback ignores invalid public key errors don't try to check
10878 certificate signature with the NULL public key.
10882 * New -ignore_err option in ocsp application to stop the server
10883 exiting on the first error in a request.
10887 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10888 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10893 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10894 extra data after the compression methods not only for TLS 1.0
10895 but also for SSL 3.0 (as required by the specification).
10897 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10899 * Change X509_certificate_type() to mark the key as exported/exportable
10900 when it's 512 *bits* long, not 512 bytes.
10904 * Change AES_cbc_encrypt() so it outputs exact multiple of
10905 blocks during encryption.
10909 * Various fixes to base64 BIO and non blocking I/O. On write
10910 flushes were not handled properly if the BIO retried. On read
10911 data was not being buffered properly and had various logic bugs.
10912 This also affects blocking I/O when the data being decoded is a
10917 * Various S/MIME bugfixes and compatibility changes:
10918 output correct application/pkcs7 MIME type if
10919 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10920 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10921 of files as .eml work). Correctly handle very long lines in MIME
10926 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10928 * Countermeasure against the Klima-Pokorny-Rosa extension of
10929 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10930 a protocol version number mismatch like a decryption error
10931 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10935 * Turn on RSA blinding by default in the default implementation
10936 to avoid a timing attack. Applications that don't want it can call
10937 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10938 They would be ill-advised to do so in most cases.
10940 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10942 * Change RSA blinding code so that it works when the PRNG is not
10943 seeded (in this case, the secret RSA exponent is abused as
10944 an unpredictable seed -- if it is not unpredictable, there
10945 is no point in blinding anyway). Make RSA blinding thread-safe
10946 by remembering the creator's thread ID in rsa->blinding and
10947 having all other threads use local one-time blinding factors
10948 (this requires more computation than sharing rsa->blinding, but
10949 avoids excessive locking; and if an RSA object is not shared
10950 between threads, blinding will still be very fast).
10954 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10955 ENGINE as defaults for all supported algorithms irrespective of
10956 the 'flags' parameter. 'flags' is now honoured, so applications
10957 should make sure they are passing it correctly.
10961 * Target "mingw" now allows native Windows code to be generated in
10962 the Cygwin environment as well as with the MinGW compiler.
10966 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10968 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10969 via timing by performing a MAC computation even if incorrect
10970 block cipher padding has been found. This is a countermeasure
10971 against active attacks where the attacker has to distinguish
10972 between bad padding and a MAC verification error. [CVE-2003-0078][]
10974 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10975 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10976 Martin Vuagnoux (EPFL, Ilion)*
10978 * Make the no-err option work as intended. The intention with no-err
10979 is not to have the whole error stack handling routines removed from
10980 libcrypto, it's only intended to remove all the function name and
10981 reason texts, thereby removing some of the footprint that may not
10982 be interesting if those errors aren't displayed anyway.
10984 NOTE: it's still possible for any application or module to have its
10985 own set of error texts inserted. The routines are there, just not
10986 used by default when no-err is given.
10990 * Add support for FreeBSD on IA64.
10992 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10994 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10995 Kerberos function mit_des_cbc_cksum(). Before this change,
10996 the value returned by DES_cbc_cksum() was like the one from
10997 mit_des_cbc_cksum(), except the bytes were swapped.
10999 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11001 * Allow an application to disable the automatic SSL chain building.
11002 Before this a rather primitive chain build was always performed in
11003 ssl3_output_cert_chain(): an application had no way to send the
11004 correct chain if the automatic operation produced an incorrect result.
11006 Now the chain builder is disabled if either:
11008 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11010 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11012 The reasoning behind this is that an application would not want the
11013 auto chain building to take place if extra chain certificates are
11014 present and it might also want a means of sending no additional
11015 certificates (for example the chain has two certificates and the
11020 * Add the possibility to build without the ENGINE framework.
11022 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11024 * Under Win32 gmtime() can return NULL: check return value in
11025 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11029 * DSA routines: under certain error conditions uninitialized BN objects
11030 could be freed. Solution: make sure initialization is performed early
11031 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11032 Nils Larsch <nla@trustcenter.de> via PR#459)
11036 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11037 checked on reconnect on the client side, therefore session resumption
11038 could still fail with a "ssl session id is different" error. This
11039 behaviour is masked when SSL_OP_ALL is used due to
11040 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11041 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11042 followup to PR #377.
11046 * IA-32 assembler support enhancements: unified ELF targets, support
11047 for SCO/Caldera platforms, fix for Cygwin shared build.
11051 * Add support for FreeBSD on sparc64. As a consequence, support for
11052 FreeBSD on non-x86 processors is separate from x86 processors on
11053 the config script, much like the NetBSD support.
11055 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11057 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11059 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11062 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11063 code (06) was taken as the first octet of the session ID and the last
11064 octet was ignored consequently. As a result SSLv2 client side session
11065 caching could not have worked due to the session ID mismatch between
11067 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11072 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11073 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11078 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11079 seems that in spite of existing for more than a year, many application
11080 author have done nothing to provide the necessary callbacks, which
11081 means that this particular engine will not work properly anywhere.
11082 This is a very unfortunate situation which forces us, in the name
11083 of usability, to give the hw_ncipher.c a static lock, which is part
11085 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11086 appear in 0.9.8 or later. We EXPECT application authors to have
11087 dealt properly with this when 0.9.8 is released (unless we actually
11088 make such changes in the libcrypto locking code that changes will
11089 have to be made anyway).
11093 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11094 octets have been read, EOF or an error occurs. Without this change
11095 some truncated ASN1 structures will not produce an error.
11099 * Disable Heimdal support, since it hasn't been fully implemented.
11100 Still give the possibility to force the use of Heimdal, but with
11101 warnings and a request that patches get sent to openssl-dev.
11105 * Add the VC-CE target, introduce the WINCE sysname, and add
11106 INSTALL.WCE and appropriate conditionals to make it build.
11108 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11110 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11111 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11112 edit numbers of the version.
11114 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11116 * Introduce safe string copy and catenation functions
11117 (BUF_strlcpy() and BUF_strlcat()).
11119 *Ben Laurie (CHATS) and Richard Levitte*
11121 * Avoid using fixed-size buffers for one-line DNs.
11123 *Ben Laurie (CHATS)*
11125 * Add BUF_MEM_grow_clean() to avoid information leakage when
11126 resizing buffers containing secrets, and use where appropriate.
11128 *Ben Laurie (CHATS)*
11130 * Avoid using fixed size buffers for configuration file location.
11132 *Ben Laurie (CHATS)*
11134 * Avoid filename truncation for various CA files.
11136 *Ben Laurie (CHATS)*
11138 * Use sizeof in preference to magic numbers.
11140 *Ben Laurie (CHATS)*
11142 * Avoid filename truncation in cert requests.
11144 *Ben Laurie (CHATS)*
11146 * Add assertions to check for (supposedly impossible) buffer
11149 *Ben Laurie (CHATS)*
11151 * Don't cache truncated DNS entries in the local cache (this could
11152 potentially lead to a spoofing attack).
11154 *Ben Laurie (CHATS)*
11156 * Fix various buffers to be large enough for hex/decimal
11157 representations in a platform independent manner.
11159 *Ben Laurie (CHATS)*
11161 * Add CRYPTO_realloc_clean() to avoid information leakage when
11162 resizing buffers containing secrets, and use where appropriate.
11164 *Ben Laurie (CHATS)*
11166 * Add BIO_indent() to avoid much slightly worrying code to do
11169 *Ben Laurie (CHATS)*
11171 * Convert sprintf()/BIO_puts() to BIO_printf().
11173 *Ben Laurie (CHATS)*
11175 * buffer_gets() could terminate with the buffer only half
11178 *Ben Laurie (CHATS)*
11180 * Add assertions to prevent user-supplied crypto functions from
11181 overflowing internal buffers by having large block sizes, etc.
11183 *Ben Laurie (CHATS)*
11185 * New OPENSSL_assert() macro (similar to assert(), but enabled
11188 *Ben Laurie (CHATS)*
11190 * Eliminate unused copy of key in RC4.
11192 *Ben Laurie (CHATS)*
11194 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11196 *Ben Laurie (CHATS)*
11198 * Fix off-by-one error in EGD path.
11200 *Ben Laurie (CHATS)*
11202 * If RANDFILE path is too long, ignore instead of truncating.
11204 *Ben Laurie (CHATS)*
11206 * Eliminate unused and incorrectly sized X.509 structure
11209 *Ben Laurie (CHATS)*
11211 * Eliminate unused and dangerous function knumber().
11213 *Ben Laurie (CHATS)*
11215 * Eliminate unused and dangerous structure, KSSL_ERR.
11217 *Ben Laurie (CHATS)*
11219 * Protect against overlong session ID context length in an encoded
11220 session object. Since these are local, this does not appear to be
11223 *Ben Laurie (CHATS)*
11225 * Change from security patch (see 0.9.6e below) that did not affect
11226 the 0.9.6 release series:
11228 Remote buffer overflow in SSL3 protocol - an attacker could
11229 supply an oversized master key in Kerberos-enabled versions.
11232 *Ben Laurie (CHATS)*
11234 * Change the SSL kerb5 codes to match RFC 2712.
11238 * Make -nameopt work fully for req and add -reqopt switch.
11240 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11242 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11244 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11246 * Make sure tests can be performed even if the corresponding algorithms
11247 have been removed entirely. This was also the last step to make
11248 OpenSSL compilable with DJGPP under all reasonable conditions.
11250 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11252 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11253 to allow version independent disabling of normally unselected ciphers,
11254 which may be activated as a side-effect of selecting a single cipher.
11256 (E.g., cipher list string "RSA" enables ciphersuites that are left
11257 out of "ALL" because they do not provide symmetric encryption.
11258 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11260 *Lutz Jaenicke, Bodo Moeller*
11262 * Add appropriate support for separate platform-dependent build
11263 directories. The recommended way to make a platform-dependent
11264 build directory is the following (tested on Linux), maybe with
11267 # Place yourself outside of the OpenSSL source tree. In
11268 # this example, the environment variable OPENSSL_SOURCE
11269 # is assumed to contain the absolute OpenSSL source directory.
11270 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11271 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11272 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11273 mkdir -p `dirname $F`
11274 ln -s $OPENSSL_SOURCE/$F $F
11277 To be absolutely sure not to disturb the source tree, a "make clean"
11278 is a good thing. If it isn't successful, don't worry about it,
11279 it probably means the source directory is very clean.
11283 * Make sure any ENGINE control commands make local copies of string
11284 pointers passed to them whenever necessary. Otherwise it is possible
11285 the caller may have overwritten (or deallocated) the original string
11286 data when a later ENGINE operation tries to use the stored values.
11288 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11290 * Improve diagnostics in file reading and command-line digests.
11292 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11294 * Add AES modes CFB and OFB to the object database. Correct an
11295 error in AES-CFB decryption.
11299 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11300 allows existing EVP_CIPHER_CTX structures to be reused after
11301 calling `EVP_*Final()`. This behaviour is used by encryption
11302 BIOs and some applications. This has the side effect that
11303 applications must explicitly clean up cipher contexts with
11304 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11308 * Check the values of dna and dnb in bn_mul_recursive before calling
11309 bn_mul_comba (a non zero value means the a or b arrays do not contain
11310 n2 elements) and fallback to bn_mul_normal if either is not zero.
11314 * Fix escaping of non-ASCII characters when using the -subj option
11315 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11319 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11320 form for "surname", serialNumber has no short form.
11321 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11322 therefore remove "mail" short name for "internet 7".
11323 The OID for unique identifiers in X509 certificates is
11324 x500UniqueIdentifier, not uniqueIdentifier.
11325 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11329 * Add an "init" command to the ENGINE config module and auto initialize
11330 ENGINEs. Without any "init" command the ENGINE will be initialized
11331 after all ctrl commands have been executed on it. If init=1 the
11332 ENGINE is initialized at that point (ctrls before that point are run
11333 on the uninitialized ENGINE and after on the initialized one). If
11334 init=0 then the ENGINE will not be initialized at all.
11338 * Fix the 'app_verify_callback' interface so that the user-defined
11339 argument is actually passed to the callback: In the
11340 SSL_CTX_set_cert_verify_callback() prototype, the callback
11341 declaration has been changed from
11344 int (*cb)(X509_STORE_CTX *,void *);
11345 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11346 i=s->ctx->app_verify_callback(&ctx)
11347 has been changed into
11348 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11350 To update applications using SSL_CTX_set_cert_verify_callback(),
11351 a dummy argument can be added to their callback functions.
11353 *D. K. Smetters <smetters@parc.xerox.com>*
11355 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11357 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11359 * Add and OPENSSL_LOAD_CONF define which will cause
11360 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11361 This allows older applications to transparently support certain
11362 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11363 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11364 load the config file and OPENSSL_add_all_algorithms_conf() which will
11365 always load it have also been added.
11369 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11370 Adjust NIDs and EVP layer.
11372 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11374 * Config modules support in openssl utility.
11376 Most commands now load modules from the config file,
11377 though in a few (such as version) this isn't done
11378 because it couldn't be used for anything.
11380 In the case of ca and req the config file used is
11381 the same as the utility itself: that is the -config
11382 command line option can be used to specify an
11387 * Move default behaviour from OPENSSL_config(). If appname is NULL
11388 use "openssl_conf" if filename is NULL use default openssl config file.
11392 * Add an argument to OPENSSL_config() to allow the use of an alternative
11393 config section name. Add a new flag to tolerate a missing config file
11394 and move code to CONF_modules_load_file().
11398 * Support for crypto accelerator cards from Accelerated Encryption
11399 Processing, www.aep.ie. (Use engine 'aep')
11400 The support was copied from 0.9.6c [engine] and adapted/corrected
11401 to work with the new engine framework.
11403 *AEP Inc. and Richard Levitte*
11405 * Support for SureWare crypto accelerator cards from Baltimore
11406 Technologies. (Use engine 'sureware')
11407 The support was copied from 0.9.6c [engine] and adapted
11408 to work with the new engine framework.
11412 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11413 make the newer ENGINE framework commands for the CHIL engine work.
11415 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11417 * Make it possible to produce shared libraries on ReliantUNIX.
11419 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11421 * Add the configuration target debug-linux-ppro.
11422 Make 'openssl rsa' use the general key loading routines
11423 implemented in apps.c, and make those routines able to
11424 handle the key format FORMAT_NETSCAPE and the variant
11427 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11429 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11431 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11433 * Add -keyform to rsautl, and document -engine.
11435 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11437 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11438 BIO_R_NO_SUCH_FILE error code rather than the generic
11439 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11443 * Add new functions
11444 ERR_peek_last_error
11445 ERR_peek_last_error_line
11446 ERR_peek_last_error_line_data.
11447 These are similar to
11449 ERR_peek_error_line
11450 ERR_peek_error_line_data,
11451 but report on the latest error recorded rather than the first one
11452 still in the error queue.
11454 *Ben Laurie, Bodo Moeller*
11456 * default_algorithms option in ENGINE config module. This allows things
11458 default_algorithms = ALL
11459 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11463 * Preliminary ENGINE config module.
11467 * New experimental application configuration code.
11471 * Change the AES code to follow the same name structure as all other
11472 symmetric ciphers, and behave the same way. Move everything to
11473 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11475 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11477 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11479 *Ben Laurie and Theo de Raadt*
11481 * Add option to output public keys in req command.
11483 *Massimiliano Pala madwolf@openca.org*
11485 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11486 (up to about 10% better than before for P-192 and P-224).
11490 * New functions/macros
11492 SSL_CTX_set_msg_callback(ctx, cb)
11493 SSL_CTX_set_msg_callback_arg(ctx, arg)
11494 SSL_set_msg_callback(ssl, cb)
11495 SSL_set_msg_callback_arg(ssl, arg)
11497 to request calling a callback function
11499 void cb(int write_p, int version, int content_type,
11500 const void *buf, size_t len, SSL *ssl, void *arg)
11502 whenever a protocol message has been completely received
11503 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11504 protocol version according to which the SSL library interprets
11505 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11506 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11507 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11508 specification (change_cipher_spec(20), alert(21), handshake(22)).
11509 'buf' and 'len' point to the actual message, 'ssl' to the
11510 SSL object, and 'arg' is the application-defined value set by
11511 SSL[_CTX]_set_msg_callback_arg().
11513 'openssl s_client' and 'openssl s_server' have new '-msg' options
11514 to enable a callback that displays all protocol messages.
11518 * Change the shared library support so shared libraries are built as
11519 soon as the corresponding static library is finished, and thereby get
11520 openssl and the test programs linked against the shared library.
11521 This still only happens when the keyword "shard" has been given to
11522 the configuration scripts.
11524 NOTE: shared library support is still an experimental thing, and
11525 backward binary compatibility is still not guaranteed.
11527 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11529 * Add support for Subject Information Access extension.
11531 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11533 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11534 additional bytes when new memory had to be allocated, not just
11535 when reusing an existing buffer.
11539 * New command line and configuration option 'utf8' for the req command.
11540 This allows field values to be specified as UTF8 strings.
11544 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11545 runs for the former and machine-readable output for the latter.
11549 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11550 of the e-mail address in the DN (i.e., it will go into a certificate
11551 extension only). The new configuration file option 'email_in_dn = no'
11552 has the same effect.
11554 *Massimiliano Pala madwolf@openca.org*
11556 * Change all functions with names starting with `des_` to be starting
11557 with `DES_` instead. Add wrappers that are compatible with libdes,
11558 but are named `_ossl_old_des_*`. Finally, add macros that map the
11559 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11560 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11561 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11564 Since we provide two compatibility mappings, the user needs to
11565 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11566 compatibility is desired. The default (i.e., when that macro
11567 isn't defined) is OpenSSL 0.9.6c compatibility.
11569 There are also macros that enable and disable the support of old
11570 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11571 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11572 are defined, the default will apply: to support the old des routines.
11574 In either case, one must include openssl/des.h to get the correct
11575 definitions. Do not try to just include openssl/des_old.h, that
11578 NOTE: This is a major break of an old API into a new one. Software
11579 authors are encouraged to switch to the `DES_` style functions. Some
11580 time in the future, des_old.h and the libdes compatibility functions
11581 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11582 default), and then completely removed.
11586 * Test for certificates which contain unsupported critical extensions.
11587 If such a certificate is found during a verify operation it is
11588 rejected by default: this behaviour can be overridden by either
11589 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11590 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11591 X509_supported_extension() has also been added which returns 1 if a
11592 particular extension is supported.
11596 * Modify the behaviour of EVP cipher functions in similar way to digests
11597 to retain compatibility with existing code.
11601 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11602 compatibility with existing code. In particular the 'ctx' parameter does
11603 not have to be to be initialized before the call to EVP_DigestInit() and
11604 it is tidied up after a call to EVP_DigestFinal(). New function
11605 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11606 EVP_MD_CTX_copy() changed to not require the destination to be
11607 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11608 requires the destination to be valid.
11610 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11611 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11615 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11616 so that complete 'Handshake' protocol structures are kept in memory
11617 instead of overwriting 'msg_type' and 'length' with 'body' data.
11621 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11623 *Massimo Santin via Richard Levitte*
11625 * Major restructuring to the underlying ENGINE code. This includes
11626 reduction of linker bloat, separation of pure "ENGINE" manipulation
11627 (initialisation, etc) from functionality dealing with implementations
11628 of specific crypto interfaces. This change also introduces integrated
11629 support for symmetric ciphers and digest implementations - so ENGINEs
11630 can now accelerate these by providing EVP_CIPHER and EVP_MD
11631 implementations of their own. This is detailed in
11632 [crypto/engine/README.md](crypto/engine/README.md)
11633 as it couldn't be adequately described here. However, there are a few
11634 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11635 were changed in the original introduction of ENGINE code have now
11636 reverted back - the hooking from this code to ENGINE is now a good
11637 deal more passive and at run-time, operations deal directly with
11638 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11639 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11640 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11641 they were not being used by the framework as there is no concept of a
11642 BIGNUM_METHOD and they could not be generalised to the new
11643 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11644 ENGINE_cpy() has been removed as it cannot be consistently defined in
11649 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11653 * Change mkdef.pl to sort symbols that get the same entry number,
11654 and make sure the automatically generated functions `ERR_load_*`
11655 become part of libeay.num as well.
11659 * New function SSL_renegotiate_pending(). This returns true once
11660 renegotiation has been requested (either SSL_renegotiate() call
11661 or HelloRequest/ClientHello received from the peer) and becomes
11662 false once a handshake has been completed.
11663 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11664 sends a HelloRequest, but does not ensure that a handshake takes
11665 place. SSL_renegotiate_pending() is useful for checking if the
11666 client has followed the request.)
11670 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11671 By default, clients may request session resumption even during
11672 renegotiation (if session ID contexts permit); with this option,
11673 session resumption is possible only in the first handshake.
11675 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11676 more bits available for options that should not be part of
11677 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11681 * Add some demos for certificate and certificate request creation.
11685 * Make maximum certificate chain size accepted from the peer application
11686 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11687 "Douglas E. Engert" <deengert@anl.gov>.
11691 * Add support for shared libraries for Unixware-7
11692 (Boyd Lynn Gerber <gerberb@zenez.com>).
11696 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11697 be done prior to destruction. Use this to unload error strings from
11698 ENGINEs that load their own error strings. NB: This adds two new API
11699 functions to "get" and "set" this destroy handler in an ENGINE.
11703 * Alter all existing ENGINE implementations (except "openssl" and
11704 "openbsd") to dynamically instantiate their own error strings. This
11705 makes them more flexible to be built both as statically-linked ENGINEs
11706 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11707 Also, add stub code to each that makes building them as self-contained
11708 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11712 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11713 implementations into applications that are completely implemented in
11714 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11715 commands that can be used to configure what shared-library to load and
11716 to control aspects of the way it is handled. Also, made an update to
11717 the [README-Engine.md](README-Engine.md) file
11718 that brings its information up-to-date and
11719 provides some information and instructions on the "dynamic" ENGINE
11720 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11724 * Make it possible to unload ranges of ERR strings with a new
11725 "ERR_unload_strings" function.
11729 * Add a copy() function to EVP_MD.
11733 * Make EVP_MD routines take a context pointer instead of just the
11734 md_data void pointer.
11738 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11739 that the digest can only process a single chunk of data
11740 (typically because it is provided by a piece of
11741 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11742 is only going to provide a single chunk of data, and hence the
11743 framework needn't accumulate the data for oneshot drivers.
11747 * As with "ERR", make it possible to replace the underlying "ex_data"
11748 functions. This change also alters the storage and management of global
11749 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11750 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11751 index counters. The API functions that use this state have been changed
11752 to take a "class_index" rather than pointers to the class's local STACK
11753 and counter, and there is now an API function to dynamically create new
11754 classes. This centralisation allows us to (a) plug a lot of the
11755 thread-safety problems that existed, and (b) makes it possible to clean
11756 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11757 such data would previously have always leaked in application code and
11758 workarounds were in place to make the memory debugging turn a blind eye
11759 to it. Application code that doesn't use this new function will still
11760 leak as before, but their memory debugging output will announce it now
11761 rather than letting it slide.
11763 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11764 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11765 has a return value to indicate success or failure.
11769 * Make it possible to replace the underlying "ERR" functions such that the
11770 global state (2 LHASH tables and 2 locks) is only used by the "default"
11771 implementation. This change also adds two functions to "get" and "set"
11772 the implementation prior to it being automatically set the first time
11773 any other ERR function takes place. Ie. an application can call "get",
11774 pass the return value to a module it has just loaded, and that module
11775 can call its own "set" function using that value. This means the
11776 module's "ERR" operations will use (and modify) the error state in the
11777 application and not in its own statically linked copy of OpenSSL code.
11781 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11782 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11783 the operation, and provides a more encapsulated way for external code
11784 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11785 to use these functions rather than manually incrementing the counts.
11787 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11791 * Add EVP test program.
11795 * Add symmetric cipher support to ENGINE. Expect the API to change!
11799 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11800 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11801 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11802 These allow a CRL to be built without having to access X509_CRL fields
11803 directly. Modify 'ca' application to use new functions.
11807 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11808 bug workarounds. Rollback attack detection is a security feature.
11809 The problem will only arise on OpenSSL servers when TLSv1 is not
11810 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11811 Software authors not wanting to support TLSv1 will have special reasons
11812 for their choice and can explicitly enable this option.
11814 *Bodo Moeller, Lutz Jaenicke*
11816 * Rationalise EVP so it can be extended: don't include a union of
11817 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11818 (similar to those existing for EVP_CIPHER_CTX).
11823 EVP_MD_CTX_init(&md); /* new function call */
11824 EVP_DigestInit(&md, EVP_sha1());
11825 EVP_DigestUpdate(&md, in, len);
11826 EVP_DigestFinal(&md, out, NULL);
11827 EVP_MD_CTX_cleanup(&md); /* new function call */
11831 * Make DES key schedule conform to the usual scheme, as well as
11832 correcting its structure. This means that calls to DES functions
11833 now have to pass a pointer to a des_key_schedule instead of a
11834 plain des_key_schedule (which was actually always a pointer
11837 des_key_schedule ks;
11839 des_set_key_checked(..., &ks);
11840 des_ncbc_encrypt(..., &ks, ...);
11842 (Note that a later change renames 'des_...' into 'DES_...'.)
11846 * Initial reduction of linker bloat: the use of some functions, such as
11847 PEM causes large amounts of unused functions to be linked in due to
11848 poor organisation. For example pem_all.c contains every PEM function
11849 which has a knock on effect of linking in large amounts of (unused)
11850 ASN1 code. Grouping together similar functions and splitting unrelated
11851 functions prevents this.
11855 * Cleanup of EVP macros.
11859 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11860 correct `_ecb suffix`.
11864 * Add initial OCSP responder support to ocsp application. The
11865 revocation information is handled using the text based index
11866 use by the ca application. The responder can either handle
11867 requests generated internally, supplied in files (for example
11868 via a CGI script) or using an internal minimal server.
11872 * Add configuration choices to get zlib compression for TLS.
11876 * Changes to Kerberos SSL for RFC 2712 compliance:
11877 1. Implemented real KerberosWrapper, instead of just using
11878 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11879 2. Implemented optional authenticator field of KerberosWrapper.
11881 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11882 and authenticator structs; see crypto/krb5/.
11884 Generalized Kerberos calls to support multiple Kerberos libraries.
11885 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11886 via Richard Levitte*
11888 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11889 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11890 values for each of the key sizes rather than having just
11891 parameters (and 'speed' generating keys each time).
11895 * Speed up EVP routines.
11898 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11899 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11900 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11901 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11903 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11904 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11905 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11908 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11910 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11914 * Added the OS2-EMX target.
11916 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11918 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11919 to support NCONF routines in extension code. New function CONF_set_nconf()
11920 to allow functions which take an NCONF to also handle the old LHASH
11921 structure: this means that the old CONF compatible routines can be
11922 retained (in particular wrt extensions) without having to duplicate the
11923 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11927 * Enhance the general user interface with mechanisms for inner control
11928 and with possibilities to have yes/no kind of prompts.
11932 * Change all calls to low level digest routines in the library and
11933 applications to use EVP. Add missing calls to HMAC_cleanup() and
11934 don't assume HMAC_CTX can be copied using memcpy().
11936 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11938 * Add the possibility to control engines through control names but with
11939 arbitrary arguments instead of just a string.
11940 Change the key loaders to take a UI_METHOD instead of a callback
11941 function pointer. NOTE: this breaks binary compatibility with earlier
11942 versions of OpenSSL [engine].
11943 Adapt the nCipher code for these new conditions and add a card insertion
11948 * Enhance the general user interface with mechanisms to better support
11949 dialog box interfaces, application-defined prompts, the possibility
11950 to use defaults (for example default passwords from somewhere else)
11951 and interrupts/cancellations.
11955 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11956 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11960 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11961 tidy up some unnecessarily weird code in 'sk_new()').
11963 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11965 * Change the key loading routines for ENGINEs to use the same kind
11966 callback (pem_password_cb) as all other routines that need this
11971 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11972 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11973 than this minimum value is recommended.
11977 * New random seeder for OpenVMS, using the system process statistics
11978 that are easily reachable.
11982 * Windows apparently can't transparently handle global
11983 variables defined in DLLs. Initialisations such as:
11985 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11987 won't compile. This is used by the any applications that need to
11988 declare their own ASN1 modules. This was fixed by adding the option
11989 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11990 needed for static libraries under Win32.
11994 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11995 setting of purpose and trust fields. New X509_STORE trust and
11996 purpose functions and tidy up setting in other SSL functions.
12000 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12001 structure. These are inherited by X509_STORE_CTX when it is
12002 initialised. This allows various defaults to be set in the
12003 X509_STORE structure (such as flags for CRL checking and custom
12004 purpose or trust settings) for functions which only use X509_STORE_CTX
12005 internally such as S/MIME.
12007 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12008 trust settings if they are not set in X509_STORE. This allows X509_STORE
12009 purposes and trust (in S/MIME for example) to override any set by default.
12011 Add command line options for CRL checking to smime, s_client and s_server
12016 * Initial CRL based revocation checking. If the CRL checking flag(s)
12017 are set then the CRL is looked up in the X509_STORE structure and
12018 its validity and signature checked, then if the certificate is found
12019 in the CRL the verify fails with a revoked error.
12021 Various new CRL related callbacks added to X509_STORE_CTX structure.
12023 Command line options added to 'verify' application to support this.
12025 This needs some additional work, such as being able to handle multiple
12026 CRLs with different times, extension based lookup (rather than just
12027 by subject name) and ultimately more complete V2 CRL extension
12032 * Add a general user interface API (crypto/ui/). This is designed
12033 to replace things like des_read_password and friends (backward
12034 compatibility functions using this new API are provided).
12035 The purpose is to remove prompting functions from the DES code
12036 section as well as provide for prompting through dialog boxes in
12037 a window system and the like.
12041 * Add "ex_data" support to ENGINE so implementations can add state at a
12042 per-structure level rather than having to store it globally.
12046 * Make it possible for ENGINE structures to be copied when retrieved by
12047 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12048 This causes the "original" ENGINE structure to act like a template,
12049 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12050 operational state can be localised to each ENGINE structure, despite the
12051 fact they all share the same "methods". New ENGINE structures returned in
12052 this case have no functional references and the return value is the single
12053 structural reference. This matches the single structural reference returned
12054 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12059 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12060 needs to match any other type at all we need to manually clear the
12065 * Changes to the "openssl engine" utility to include;
12066 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12067 about an ENGINE's available control commands.
12068 - executing control commands from command line arguments using the
12069 '-pre' and '-post' switches. '-post' is only used if '-t' is
12070 specified and the ENGINE is successfully initialised. The syntax for
12071 the individual commands are colon-separated, for example;
12072 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12076 * New dynamic control command support for ENGINEs. ENGINEs can now
12077 declare their own commands (numbers), names (strings), descriptions,
12078 and input types for run-time discovery by calling applications. A
12079 subset of these commands are implicitly classed as "executable"
12080 depending on their input type, and only these can be invoked through
12081 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12082 can be based on user input, config files, etc). The distinction is
12083 that "executable" commands cannot return anything other than a boolean
12084 result and can only support numeric or string input, whereas some
12085 discoverable commands may only be for direct use through
12086 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12087 pointers, or other custom uses. The "executable" commands are to
12088 support parameterisations of ENGINE behaviour that can be
12089 unambiguously defined by ENGINEs and used consistently across any
12090 OpenSSL-based application. Commands have been added to all the
12091 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12092 control over shared-library paths without source code alterations.
12096 * Changed all ENGINE implementations to dynamically allocate their
12097 ENGINEs rather than declaring them statically. Apart from this being
12098 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12099 this also allows the implementations to compile without using the
12100 internal engine_int.h header.
12104 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12105 'const' value. Any code that should be able to modify a RAND_METHOD
12106 should already have non-const pointers to it (ie. they should only
12107 modify their own ones).
12111 * Made a variety of little tweaks to the ENGINE code.
12112 - "atalla" and "ubsec" string definitions were moved from header files
12113 to C code. "nuron" string definitions were placed in variables
12114 rather than hard-coded - allowing parameterisation of these values
12115 later on via ctrl() commands.
12116 - Removed unused "#if 0"'d code.
12117 - Fixed engine list iteration code so it uses ENGINE_free() to release
12118 structural references.
12119 - Constified the RAND_METHOD element of ENGINE structures.
12120 - Constified various get/set functions as appropriate and added
12121 missing functions (including a catch-all ENGINE_cpy that duplicates
12122 all ENGINE values onto a new ENGINE except reference counts/state).
12123 - Removed NULL parameter checks in get/set functions. Setting a method
12124 or function to NULL is a way of cancelling out a previously set
12125 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12126 and doesn't justify the extra error symbols and code.
12127 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12128 flags from engine_int.h to engine.h.
12129 - Changed prototypes for ENGINE handler functions (init(), finish(),
12130 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12134 * Implement binary inversion algorithm for BN_mod_inverse in addition
12135 to the algorithm using long division. The binary algorithm can be
12136 used only if the modulus is odd. On 32-bit systems, it is faster
12137 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12138 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12139 up to 450 bits. In 64-bit environments, the binary algorithm
12140 appears to be advantageous for much longer moduli; here we use it
12141 for moduli up to 2048 bits.
12145 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12146 could not support the combine flag in choice fields.
12150 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12151 extensions from a certificate request to the certificate.
12155 * Allow multiple 'certopt' and 'nameopt' options to be separated
12156 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12157 file: this allows the display of the certificate about to be
12158 signed to be customised, to allow certain fields to be included
12159 or excluded and extension details. The old system didn't display
12160 multicharacter strings properly, omitted fields not in the policy
12161 and couldn't display additional details such as extensions.
12165 * Function EC_POINTs_mul for multiple scalar multiplication
12166 of an arbitrary number of elliptic curve points
12167 \sum scalars[i]*points[i],
12168 optionally including the generator defined for the EC_GROUP:
12169 scalar*generator + \sum scalars[i]*points[i].
12171 EC_POINT_mul is a simple wrapper function for the typical case
12172 that the point list has just one item (besides the optional
12177 * First EC_METHODs for curves over GF(p):
12179 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12180 operations and provides various method functions that can also
12181 operate with faster implementations of modular arithmetic.
12183 EC_GFp_mont_method() reuses most functions that are part of
12184 EC_GFp_simple_method, but uses Montgomery arithmetic.
12186 *Bodo Moeller; point addition and point doubling
12187 implementation directly derived from source code provided by
12188 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12190 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12191 crypto/ec/ec_lib.c):
12193 Curves are EC_GROUP objects (with an optional group generator)
12194 based on EC_METHODs that are built into the library.
12196 Points are EC_POINT objects based on EC_GROUP objects.
12198 Most of the framework would be able to handle curves over arbitrary
12199 finite fields, but as there are no obvious types for fields other
12200 than GF(p), some functions are limited to that for now.
12204 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12205 that the file contains a complete HTTP response.
12209 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12210 change the def and num file printf format specifier from "%-40sXXX"
12211 to "%-39s XXX". The latter will always guarantee a space after the
12212 field while the former will cause them to run together if the field
12213 is 40 of more characters long.
12217 * Constify the cipher and digest 'method' functions and structures
12218 and modify related functions to take constant EVP_MD and EVP_CIPHER
12223 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12224 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12228 * Modify `EVP_Digest*()` routines so they now return values. Although the
12229 internal software routines can never fail additional hardware versions
12234 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12236 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12237 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12240 ERR_R_NESTED_ASN1_ERROR
12242 ERR_R_MISSING_ASN1_EOS
12243 were 4 .. 9, conflicting with
12244 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12246 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12247 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12249 Add new error code 'ERR_R_INTERNAL_ERROR'.
12253 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12258 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12259 sets the subject name for a new request or supersedes the
12260 subject name in a given request. Formats that can be parsed are
12261 'CN=Some Name, OU=myOU, C=IT'
12263 'CN=Some Name/OU=myOU/C=IT'.
12265 Add options '-batch' and '-verbose' to 'openssl req'.
12267 *Massimiliano Pala <madwolf@hackmasters.net>*
12269 * Introduce the possibility to access global variables through
12270 functions on platform were that's the best way to handle exporting
12271 global variables in shared libraries. To enable this functionality,
12272 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12273 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12274 is normally done by Configure or something similar).
12276 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12277 in the source file (foo.c) like this:
12279 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12280 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12282 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12283 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12285 OPENSSL_DECLARE_GLOBAL(int,foo);
12286 #define foo OPENSSL_GLOBAL_REF(foo)
12287 OPENSSL_DECLARE_GLOBAL(double,bar);
12288 #define bar OPENSSL_GLOBAL_REF(bar)
12290 The #defines are very important, and therefore so is including the
12291 header file everywhere where the defined globals are used.
12293 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12294 of ASN.1 items, but that structure is a bit different.
12296 The largest change is in util/mkdef.pl which has been enhanced with
12297 better and easier to understand logic to choose which symbols should
12298 go into the Windows .def files as well as a number of fixes and code
12299 cleanup (among others, algorithm keywords are now sorted
12300 lexicographically to avoid constant rewrites).
12304 * In BN_div() keep a copy of the sign of 'num' before writing the
12305 result to 'rm' because if rm==num the value will be overwritten
12306 and produce the wrong result if 'num' is negative: this caused
12307 problems with BN_mod() and BN_nnmod().
12311 * Function OCSP_request_verify(). This checks the signature on an
12312 OCSP request and verifies the signer certificate. The signer
12313 certificate is just checked for a generic purpose and OCSP request
12318 * Add OCSP_check_validity() function to check the validity of OCSP
12319 responses. OCSP responses are prepared in real time and may only
12320 be a few seconds old. Simply checking that the current time lies
12321 between thisUpdate and nextUpdate max reject otherwise valid responses
12322 caused by either OCSP responder or client clock inaccuracy. Instead
12323 we allow thisUpdate and nextUpdate to fall within a certain period of
12324 the current time. The age of the response can also optionally be
12325 checked. Two new options -validity_period and -status_age added to
12330 * If signature or public key algorithm is unrecognized print out its
12331 OID rather that just UNKNOWN.
12335 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12336 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12337 ID to be generated from the issuer certificate alone which can then be
12338 passed to OCSP_id_issuer_cmp().
12342 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12343 ASN1 modules to export functions returning ASN1_ITEM pointers
12344 instead of the ASN1_ITEM structures themselves. This adds several
12345 new macros which allow the underlying ASN1 function/structure to
12346 be accessed transparently. As a result code should not use ASN1_ITEM
12347 references directly (such as &X509_it) but instead use the relevant
12348 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12349 use of the new ASN1 code on platforms where exporting structures
12350 is problematical (for example in shared libraries) but exporting
12351 functions returning pointers to structures is not.
12355 * Add support for overriding the generation of SSL/TLS session IDs.
12356 These callbacks can be registered either in an SSL_CTX or per SSL.
12357 The purpose of this is to allow applications to control, if they wish,
12358 the arbitrary values chosen for use as session IDs, particularly as it
12359 can be useful for session caching in multiple-server environments. A
12360 command-line switch for testing this (and any client code that wishes
12361 to use such a feature) has been added to "s_server".
12363 *Geoff Thorpe, Lutz Jaenicke*
12365 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12366 of the form `#if defined(...) || defined(...) || ...` and
12367 `#if !defined(...) && !defined(...) && ...`. This also avoids
12368 the growing number of special cases it was previously handling.
12372 * Make all configuration macros available for application by making
12373 sure they are available in opensslconf.h, by giving them names starting
12374 with `OPENSSL_` to avoid conflicts with other packages and by making
12375 sure e_os2.h will cover all platform-specific cases together with
12377 Additionally, it is now possible to define configuration/platform-
12378 specific names (called "system identities"). In the C code, these
12379 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12380 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12381 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12386 * New option -set_serial to 'req' and 'x509' this allows the serial
12387 number to use to be specified on the command line. Previously self
12388 signed certificates were hard coded with serial number 0 and the
12389 CA options of 'x509' had to use a serial number in a file which was
12394 * New options to 'ca' utility to support V2 CRL entry extensions.
12395 Currently CRL reason, invalidity date and hold instruction are
12396 supported. Add new CRL extensions to V3 code and some new objects.
12400 * New function EVP_CIPHER_CTX_set_padding() this is used to
12401 disable standard block padding (aka PKCS#5 padding) in the EVP
12402 API, which was previously mandatory. This means that the data is
12403 not padded in any way and so the total length much be a multiple
12404 of the block size, otherwise an error occurs.
12408 * Initial (incomplete) OCSP SSL support.
12412 * New function OCSP_parse_url(). This splits up a URL into its host,
12413 port and path components: primarily to parse OCSP URLs. New -url
12414 option to ocsp utility.
12418 * New nonce behavior. The return value of OCSP_check_nonce() now
12419 reflects the various checks performed. Applications can decide
12420 whether to tolerate certain situations such as an absent nonce
12421 in a response when one was present in a request: the ocsp application
12422 just prints out a warning. New function OCSP_add1_basic_nonce()
12423 this is to allow responders to include a nonce in a response even if
12424 the request is nonce-less.
12428 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12429 skipped when using openssl x509 multiple times on a single input file,
12430 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12434 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12435 set string type: to handle setting ASN1_TIME structures. Fix ca
12436 utility to correctly initialize revocation date of CRLs.
12440 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12441 the clients preferred ciphersuites and rather use its own preferences.
12442 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12443 Internet Explorer by ensuring unchanged hash method during stepup.
12444 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12448 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12449 to aes and add a new 'exist' option to print out symbols that don't
12454 * Additional options to ocsp utility to allow flags to be set and
12455 additional certificates supplied.
12459 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12460 OCSP client a number of certificate to only verify the response
12465 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12466 handle the new API. Currently only ECB, CBC modes supported. Add new
12469 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12470 Encryption Standard (AES) Ciphersuites for Transport Layer
12471 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12472 not enabled by default and were not part of the "ALL" ciphersuite
12473 alias because they were not yet official; they could be
12474 explicitly requested by specifying the "AESdraft" ciphersuite
12475 group alias. In the final release of OpenSSL 0.9.7, the group
12476 alias is called "AES" and is part of "ALL".)
12478 *Ben Laurie, Steve Henson, Bodo Moeller*
12480 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12481 request to response.
12485 * Functions for OCSP responders. OCSP_request_onereq_count(),
12486 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12487 extract information from a certificate request. OCSP_response_create()
12488 creates a response and optionally adds a basic response structure.
12489 OCSP_basic_add1_status() adds a complete single response to a basic
12490 response and returns the OCSP_SINGLERESP structure just added (to allow
12491 extensions to be included for example). OCSP_basic_add1_cert() adds a
12492 certificate to a basic response and OCSP_basic_sign() signs a basic
12493 response with various flags. New helper functions ASN1_TIME_check()
12494 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12495 (converts ASN1_TIME to GeneralizedTime).
12499 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12500 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12501 structure from a certificate. X509_pubkey_digest() digests the public_key
12502 contents: this is used in various key identifiers.
12506 * Make sk_sort() tolerate a NULL argument.
12508 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12510 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12511 passed by the function are trusted implicitly. If any of them signed the
12512 response then it is assumed to be valid and is not verified.
12516 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12517 to data. This was previously part of the PKCS7 ASN1 code. This
12518 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12519 *Steve Henson, reported by Kenneth R. Robinette
12520 <support@securenetterm.com>*
12522 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12523 routines: without these tracing memory leaks is very painful.
12524 Fix leaks in PKCS12 and PKCS7 routines.
12528 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12529 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12530 effectively meant GeneralizedTime would never be used. Now it
12531 is initialised to -1 but X509_time_adj() now has to check the value
12532 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12533 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12534 *Steve Henson, reported by Kenneth R. Robinette
12535 <support@securenetterm.com>*
12537 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12538 result in a zero length in the ASN1_INTEGER structure which was
12539 not consistent with the structure when d2i_ASN1_INTEGER() was used
12540 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12541 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12542 where it did not print out a minus for negative ASN1_INTEGER.
12546 * Add summary printout to ocsp utility. The various functions which
12547 convert status values to strings have been renamed to:
12548 OCSP_response_status_str(), OCSP_cert_status_str() and
12549 OCSP_crl_reason_str() and are no longer static. New options
12550 to verify nonce values and to disable verification. OCSP response
12551 printout format cleaned up.
12555 * Add additional OCSP certificate checks. These are those specified
12556 in RFC2560. This consists of two separate checks: the CA of the
12557 certificate being checked must either be the OCSP signer certificate
12558 or the issuer of the OCSP signer certificate. In the latter case the
12559 OCSP signer certificate must contain the OCSP signing extended key
12560 usage. This check is performed by attempting to match the OCSP
12561 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12562 in the OCSP_CERTID structures of the response.
12566 * Initial OCSP certificate verification added to OCSP_basic_verify()
12567 and related routines. This uses the standard OpenSSL certificate
12568 verify routines to perform initial checks (just CA validity) and
12569 to obtain the certificate chain. Then additional checks will be
12570 performed on the chain. Currently the root CA is checked to see
12571 if it is explicitly trusted for OCSP signing. This is used to set
12572 a root CA as a global signing root: that is any certificate that
12573 chains to that CA is an acceptable OCSP signing certificate.
12577 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12578 extensions from a separate configuration file.
12579 As when reading extensions from the main configuration file,
12580 the '-extensions ...' option may be used for specifying the
12583 *Massimiliano Pala <madwolf@comune.modena.it>*
12585 * New OCSP utility. Allows OCSP requests to be generated or
12586 read. The request can be sent to a responder and the output
12587 parsed, outputted or printed in text form. Not complete yet:
12588 still needs to check the OCSP response validity.
12592 * New subcommands for 'openssl ca':
12593 `openssl ca -status <serial>` prints the status of the cert with
12594 the given serial number (according to the index file).
12595 `openssl ca -updatedb` updates the expiry status of certificates
12598 *Massimiliano Pala <madwolf@comune.modena.it>*
12600 * New '-newreq-nodes' command option to CA.pl. This is like
12601 '-newreq', but calls 'openssl req' with the '-nodes' option
12602 so that the resulting key is not encrypted.
12604 *Damien Miller <djm@mindrot.org>*
12606 * New configuration for the GNU Hurd.
12608 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12610 * Initial code to implement OCSP basic response verify. This
12611 is currently incomplete. Currently just finds the signer's
12612 certificate and verifies the signature on the response.
12616 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12617 value of OPENSSLDIR. This is available via the new '-d' option
12618 to 'openssl version', and is also included in 'openssl version -a'.
12622 * Allowing defining memory allocation callbacks that will be given
12623 file name and line number information in additional arguments
12624 (a `const char*` and an int). The basic functionality remains, as
12625 well as the original possibility to just replace malloc(),
12626 realloc() and free() by functions that do not know about these
12627 additional arguments. To register and find out the current
12628 settings for extended allocation functions, the following
12629 functions are provided:
12631 CRYPTO_set_mem_ex_functions
12632 CRYPTO_set_locked_mem_ex_functions
12633 CRYPTO_get_mem_ex_functions
12634 CRYPTO_get_locked_mem_ex_functions
12636 These work the same way as CRYPTO_set_mem_functions and friends.
12637 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12638 extended allocation function is enabled.
12639 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12640 a conventional allocation function is enabled.
12642 *Richard Levitte, Bodo Moeller*
12644 * Finish off removing the remaining LHASH function pointer casts.
12645 There should no longer be any prototype-casting required when using
12646 the LHASH abstraction, and any casts that remain are "bugs". See
12647 the callback types and macros at the head of lhash.h for details
12648 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12652 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12653 If /dev/[u]random devices are not available or do not return enough
12654 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12656 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12657 /etc/entropy will be queried once each in this sequence, querying stops
12658 when enough entropy was collected without querying more sockets.
12662 * Change the Unix RAND_poll() variant to be able to poll several
12663 random devices, as specified by DEVRANDOM, until a sufficient amount
12664 of data has been collected. We spend at most 10 ms on each file
12665 (select timeout) and read in non-blocking mode. DEVRANDOM now
12666 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12667 (previously it was just the string "/dev/urandom"), so on typical
12668 platforms the 10 ms delay will never occur.
12669 Also separate out the Unix variant to its own file, rand_unix.c.
12670 For VMS, there's a currently-empty rand_vms.c.
12674 * Move OCSP client related routines to ocsp_cl.c. These
12675 provide utility functions which an application needing
12676 to issue a request to an OCSP responder and analyse the
12677 response will typically need: as opposed to those which an
12678 OCSP responder itself would need which will be added later.
12680 OCSP_request_sign() signs an OCSP request with an API similar
12681 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12682 response. OCSP_response_get1_basic() extracts basic response
12683 from response. OCSP_resp_find_status(): finds and extracts status
12684 information from an OCSP_CERTID structure (which will be created
12685 when the request structure is built). These are built from lower
12686 level functions which work on OCSP_SINGLERESP structures but
12687 won't normally be used unless the application wishes to examine
12688 extensions in the OCSP response for example.
12690 Replace nonce routines with a pair of functions.
12691 OCSP_request_add1_nonce() adds a nonce value and optionally
12692 generates a random value. OCSP_check_nonce() checks the
12693 validity of the nonce in an OCSP response.
12697 * Change function OCSP_request_add() to OCSP_request_add0_id().
12698 This doesn't copy the supplied OCSP_CERTID and avoids the
12699 need to free up the newly created id. Change return type
12700 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12701 This can then be used to add extensions to the request.
12702 Deleted OCSP_request_new(), since most of its functionality
12703 is now in OCSP_REQUEST_new() (and the case insensitive name
12704 clash) apart from the ability to set the request name which
12705 will be added elsewhere.
12709 * Update OCSP API. Remove obsolete extensions argument from
12710 various functions. Extensions are now handled using the new
12711 OCSP extension code. New simple OCSP HTTP function which
12712 can be used to send requests and parse the response.
12716 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12717 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12718 uses the special reorder version of SET OF to sort the attributes
12719 and reorder them to match the encoded order. This resolves a long
12720 standing problem: a verify on a PKCS7 structure just after signing
12721 it used to fail because the attribute order did not match the
12722 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12723 it uses the received order. This is necessary to tolerate some broken
12724 software that does not order SET OF. This is handled by encoding
12725 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12726 to produce the required SET OF.
12730 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12731 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12732 files to get correct declarations of the ASN.1 item variables.
12736 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12737 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12738 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12739 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12740 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12741 ASN1_ITEM and no wrapper functions.
12745 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12746 replace the old function pointer based I/O routines. Change most of
12747 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12751 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12752 lines, recognize more "algorithms" that can be deselected, and make
12753 it complain about algorithm deselection that isn't recognised.
12757 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12758 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12759 to use new functions. Add NO_ASN1_OLD which can be set to remove
12760 some old style ASN1 functions: this can be used to determine if old
12761 code will still work when these eventually go away.
12765 * New extension functions for OCSP structures, these follow the
12766 same conventions as certificates and CRLs.
12770 * New function X509V3_add1_i2d(). This automatically encodes and
12771 adds an extension. Its behaviour can be customised with various
12772 flags to append, replace or delete. Various wrappers added for
12773 certificates and CRLs.
12777 * Fix to avoid calling the underlying ASN1 print routine when
12778 an extension cannot be parsed. Correct a typo in the
12779 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12783 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12784 entries for variables.
12788 * Add functionality to apps/openssl.c for detecting locking
12789 problems: As the program is single-threaded, all we have
12790 to do is register a locking callback using an array for
12791 storing which locks are currently held by the program.
12795 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12796 SSL_get_ex_data_X509_STORE_idx(), which is used in
12797 ssl_verify_cert_chain() and thus can be called at any time
12798 during TLS/SSL handshakes so that thread-safety is essential.
12799 Unfortunately, the ex_data design is not at all suited
12800 for multi-threaded use, so it probably should be abolished.
12804 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12806 *Broadcom, tweaked and integrated by Geoff Thorpe*
12808 * Move common extension printing code to new function
12809 X509V3_print_extensions(). Reorganise OCSP print routines and
12810 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12814 * New function X509_signature_print() to remove duplication in some
12819 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12820 set (this was treated exactly the same as SET OF previously). This
12821 is used to reorder the STACK representing the structure to match the
12822 encoding. This will be used to get round a problem where a PKCS7
12823 structure which was signed could not be verified because the STACK
12824 order did not reflect the encoded order.
12828 * Reimplement the OCSP ASN1 module using the new code.
12832 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12833 for its ASN1 operations. The old style function pointers still exist
12834 for now but they will eventually go away.
12838 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12839 completely replaces the old ASN1 functionality with a table driven
12840 encoder and decoder which interprets an ASN1_ITEM structure describing
12841 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12842 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12843 has also been converted to the new form.
12847 * Change BN_mod_exp_recp so that negative moduli are tolerated
12848 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12849 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12850 for negative moduli.
12854 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12855 of not touching the result's sign bit.
12859 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12864 * Changed the LHASH code to use prototypes for callbacks, and created
12865 macros to declare and implement thin (optionally static) functions
12866 that provide type-safety and avoid function pointer casting for the
12867 type-specific callbacks.
12871 * Added Kerberos Cipher Suites to be used with TLS, as written in
12873 *Veers Staats <staatsvr@asc.hpc.mil>,
12874 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12876 * Reformat the FAQ so the different questions and answers can be divided
12877 in sections depending on the subject.
12881 * Have the zlib compression code load ZLIB.DLL dynamically under
12886 * New function BN_mod_sqrt for computing square roots modulo a prime
12887 (using the probabilistic Tonelli-Shanks algorithm unless
12888 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12889 be handled deterministically).
12891 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12893 * Make BN_mod_inverse faster by explicitly handling small quotients
12894 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12895 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12899 * New function BN_kronecker.
12903 * Fix BN_gcd so that it works on negative inputs; the result is
12904 positive unless both parameters are zero.
12905 Previously something reasonably close to an infinite loop was
12906 possible because numbers could be growing instead of shrinking
12907 in the implementation of Euclid's algorithm.
12911 * Fix BN_is_word() and BN_is_one() macros to take into account the
12912 sign of the number in question.
12914 Fix BN_is_word(a,w) to work correctly for w == 0.
12916 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12917 because its test if the absolute value of 'a' equals 'w'.
12918 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12919 it exists mostly for use in the implementations of BN_is_zero(),
12920 BN_is_one(), and BN_is_word().
12924 * New function BN_swap.
12928 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12929 the exponentiation functions are more likely to produce reasonable
12930 results on negative inputs.
12934 * Change BN_mod_mul so that the result is always non-negative.
12935 Previously, it could be negative if one of the factors was negative;
12936 I don't think anyone really wanted that behaviour.
12940 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
12941 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
12942 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
12943 and add new functions:
12952 BN_mod_lshift1_quick
12954 BN_mod_lshift_quick
12956 These functions always generate non-negative results.
12958 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
12959 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
12961 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
12962 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
12963 be reduced modulo `m`.
12965 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12968 The following entry accidentally appeared in the CHANGES file
12969 distributed with OpenSSL 0.9.7. The modifications described in
12970 it do *not* apply to OpenSSL 0.9.7.
12972 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12973 was actually never needed) and in BN_mul(). The removal in BN_mul()
12974 required a small change in bn_mul_part_recursive() and the addition
12975 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12976 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12977 bn_sub_words() and bn_add_words() except they take arrays with
12983 * In 'openssl passwd', verify passwords read from the terminal
12984 unless the '-salt' option is used (which usually means that
12985 verification would just waste user's time since the resulting
12986 hash is going to be compared with some given password hash)
12987 or the new '-noverify' option is used.
12989 This is an incompatible change, but it does not affect
12990 non-interactive use of 'openssl passwd' (passwords on the command
12991 line, '-stdin' option, '-in ...' option) and thus should not
12992 cause any problems.
12996 * Remove all references to RSAref, since there's no more need for it.
13000 * Make DSO load along a path given through an environment variable
13001 (SHLIB_PATH) with shl_load().
13005 * Constify the ENGINE code as a result of BIGNUM constification.
13006 Also constify the RSA code and most things related to it. In a
13007 few places, most notable in the depth of the ASN.1 code, ugly
13008 casts back to non-const were required (to be solved at a later
13013 * Make it so the openssl application has all engines loaded by default.
13017 * Constify the BIGNUM routines a little more.
13021 * Add the following functions:
13023 ENGINE_load_cswift()
13025 ENGINE_load_atalla()
13026 ENGINE_load_nuron()
13027 ENGINE_load_builtin_engines()
13029 That way, an application can itself choose if external engines that
13030 are built-in in OpenSSL shall ever be used or not. The benefit is
13031 that applications won't have to be linked with libdl or other dso
13032 libraries unless it's really needed.
13034 Changed 'openssl engine' to load all engines on demand.
13035 Changed the engine header files to avoid the duplication of some
13036 declarations (they differed!).
13040 * 'openssl engine' can now list capabilities.
13044 * Better error reporting in 'openssl engine'.
13048 * Never call load_dh_param(NULL) in s_server.
13052 * Add engine application. It can currently list engines by name and
13053 identity, and test if they are actually available.
13057 * Improve RPM specification file by forcing symbolic linking and making
13058 sure the installed documentation is also owned by root.root.
13060 *Damien Miller <djm@mindrot.org>*
13062 * Give the OpenSSL applications more possibilities to make use of
13063 keys (public as well as private) handled by engines.
13067 * Add OCSP code that comes from CertCo.
13071 * Add VMS support for the Rijndael code.
13075 * Added untested support for Nuron crypto accelerator.
13079 * Add support for external cryptographic devices. This code was
13080 previously distributed separately as the "engine" branch.
13082 *Geoff Thorpe, Richard Levitte*
13084 * Rework the filename-translation in the DSO code. It is now possible to
13085 have far greater control over how a "name" is turned into a filename
13086 depending on the operating environment and any oddities about the
13087 different shared library filenames on each system.
13091 * Support threads on FreeBSD-elf in Configure.
13095 * Fix for SHA1 assembly problem with MASM: it produces
13096 warnings about corrupt line number information when assembling
13097 with debugging information. This is caused by the overlapping
13100 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13103 NCONF_get_number() has no error checking at all. As a replacement,
13104 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13105 promoted strongly. The old NCONF_get_number is kept around for
13106 binary backward compatibility.
13107 Make it possible for methods to load from something other than a BIO,
13108 by providing a function pointer that is given a name instead of a BIO.
13109 For example, this could be used to load configuration data from an
13114 * Fix for non blocking accept BIOs. Added new I/O special reason
13115 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13116 with non blocking I/O was not possible because no retry code was
13117 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13122 * Added the beginnings of Rijndael support.
13126 * Fix for bug in DirectoryString mask setting. Add support for
13127 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13128 to allow certificate printing to more controllable, additional
13129 'certopt' option to 'x509' to allow new printing options to be
13134 * Clean old EAY MD5 hack from e_os.h.
13138 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13140 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13141 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13143 *Joe Orton, Steve Henson*
13145 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13147 * Fix additional bug revealed by the NISCC test suite:
13149 Stop bug triggering large recursion when presented with
13150 certain ASN.1 tags [CVE-2003-0851][]
13154 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13156 * Fix various bugs revealed by running the NISCC test suite:
13158 Stop out of bounds reads in the ASN1 code when presented with
13159 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13161 If verify callback ignores invalid public key errors don't try to check
13162 certificate signature with the NULL public key.
13166 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13167 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13172 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13173 extra data after the compression methods not only for TLS 1.0
13174 but also for SSL 3.0 (as required by the specification).
13176 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13178 * Change X509_certificate_type() to mark the key as exported/exportable
13179 when it's 512 *bits* long, not 512 bytes.
13183 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13185 * Countermeasure against the Klima-Pokorny-Rosa extension of
13186 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13187 a protocol version number mismatch like a decryption error
13188 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13192 * Turn on RSA blinding by default in the default implementation
13193 to avoid a timing attack. Applications that don't want it can call
13194 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13195 They would be ill-advised to do so in most cases.
13197 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13199 * Change RSA blinding code so that it works when the PRNG is not
13200 seeded (in this case, the secret RSA exponent is abused as
13201 an unpredictable seed -- if it is not unpredictable, there
13202 is no point in blinding anyway). Make RSA blinding thread-safe
13203 by remembering the creator's thread ID in rsa->blinding and
13204 having all other threads use local one-time blinding factors
13205 (this requires more computation than sharing rsa->blinding, but
13206 avoids excessive locking; and if an RSA object is not shared
13207 between threads, blinding will still be very fast).
13211 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13213 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13214 via timing by performing a MAC computation even if incorrect
13215 block cipher padding has been found. This is a countermeasure
13216 against active attacks where the attacker has to distinguish
13217 between bad padding and a MAC verification error. [CVE-2003-0078][]
13219 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13220 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13221 Martin Vuagnoux (EPFL, Ilion)*
13223 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13225 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13226 memory from its contents. This is done with a counter that will
13227 place alternating values in each byte. This can be used to solve
13228 two issues: 1) the removal of calls to memset() by highly optimizing
13229 compilers, and 2) cleansing with other values than 0, since those can
13230 be read through on certain media, for example a swap space on disk.
13234 * Bugfix: client side session caching did not work with external caching,
13235 because the session->cipher setting was not restored when reloading
13236 from the external cache. This problem was masked, when
13237 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13238 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13242 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13243 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13245 *Zeev Lieber <zeev-l@yahoo.com>*
13247 * Undo an undocumented change introduced in 0.9.6e which caused
13248 repeated calls to OpenSSL_add_all_ciphers() and
13249 OpenSSL_add_all_digests() to be ignored, even after calling
13254 * Change the default configuration reader to deal with last line not
13255 being properly terminated.
13259 * Change X509_NAME_cmp() so it applies the special rules on handling
13260 DN values that are of type PrintableString, as well as RDNs of type
13261 emailAddress where the value has the type ia5String.
13263 *stefank@valicert.com via Richard Levitte*
13265 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13266 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13267 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13268 the bitwise-OR of the two for use by the majority of applications
13269 wanting this behaviour, and update the docs. The documented
13270 behaviour and actual behaviour were inconsistent and had been
13271 changing anyway, so this is more a bug-fix than a behavioural
13274 *Geoff Thorpe, diagnosed by Nadav Har'El*
13276 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13277 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13281 * Fix initialization code race conditions in
13282 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13283 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13284 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13285 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13286 ssl2_get_cipher_by_char(),
13287 ssl3_get_cipher_by_char().
13289 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13291 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13292 the cached sessions are flushed, as the remove_cb() might use ex_data
13293 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13294 (see [openssl.org #212]).
13296 *Geoff Thorpe, Lutz Jaenicke*
13298 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13299 length, instead of the encoding length to d2i_ASN1_OBJECT.
13303 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13305 * [In 0.9.6g-engine release:]
13306 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13308 *Lynn Gazis <lgazis@rainbow.com>*
13310 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13312 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13313 and get fix the header length calculation.
13314 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13315 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13317 * Use proper error handling instead of 'assertions' in buffer
13318 overflow checks added in 0.9.6e. This prevents DoS (the
13319 assertions could call abort()).
13321 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13323 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13325 * Add various sanity checks to asn1_get_length() to reject
13326 the ASN1 length bytes if they exceed sizeof(long), will appear
13327 negative or the content length exceeds the length of the
13330 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13332 * Fix cipher selection routines: ciphers without encryption had no flags
13333 for the cipher strength set and where therefore not handled correctly
13334 by the selection routines (PR #130).
13338 * Fix EVP_dsa_sha macro.
13343 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13344 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13345 that was added in OpenSSL 0.9.6d.
13347 As the countermeasure turned out to be incompatible with some
13348 broken SSL implementations, the new option is part of SSL_OP_ALL.
13349 SSL_OP_ALL is usually employed when compatibility with weird SSL
13350 implementations is desired (e.g. '-bugs' option to 's_client' and
13351 's_server'), so the new option is automatically set in many
13356 * Changes in security patch:
13358 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13359 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13360 Air Force Materiel Command, USAF, under agreement number
13363 * Add various sanity checks to asn1_get_length() to reject
13364 the ASN1 length bytes if they exceed sizeof(long), will appear
13365 negative or the content length exceeds the length of the
13366 supplied buffer. [CVE-2002-0659][]
13368 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13370 * Assertions for various potential buffer overflows, not known to
13371 happen in practice.
13373 *Ben Laurie (CHATS)*
13375 * Various temporary buffers to hold ASCII versions of integers were
13376 too small for 64 bit platforms. [CVE-2002-0655][]
13377 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13379 * Remote buffer overflow in SSL3 protocol - an attacker could
13380 supply an oversized session ID to a client. [CVE-2002-0656][]
13382 *Ben Laurie (CHATS)*
13384 * Remote buffer overflow in SSL2 protocol - an attacker could
13385 supply an oversized client master key. [CVE-2002-0656][]
13387 *Ben Laurie (CHATS)*
13389 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13391 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13392 encoded as NULL) with id-dsa-with-sha1.
13394 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13396 * Check various `X509_...()` return values in apps/req.c.
13398 *Nils Larsch <nla@trustcenter.de>*
13400 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13401 an end-of-file condition would erroneously be flagged, when the CRLF
13402 was just at the end of a processed block. The bug was discovered when
13403 processing data through a buffering memory BIO handing the data to a
13404 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13405 <ptsekov@syntrex.com> and Nedelcho Stanev.
13409 * Implement a countermeasure against a vulnerability recently found
13410 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13411 before application data chunks to avoid the use of known IVs
13412 with data potentially chosen by the attacker.
13416 * Fix length checks in ssl3_get_client_hello().
13420 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13421 to prevent ssl3_read_internal() from incorrectly assuming that
13422 ssl3_read_bytes() found application data while handshake
13423 processing was enabled when in fact s->s3->in_read_app_data was
13424 merely automatically cleared during the initial handshake.
13426 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13428 * Fix object definitions for Private and Enterprise: they were not
13429 recognized in their shortname (=lowercase) representation. Extend
13430 obj_dat.pl to issue an error when using undefined keywords instead
13431 of silently ignoring the problem (Svenning Sorensen
13432 <sss@sss.dnsalias.net>).
13436 * Fix DH_generate_parameters() so that it works for 'non-standard'
13437 generators, i.e. generators other than 2 and 5. (Previously, the
13438 code did not properly initialise the 'add' and 'rem' values to
13439 BN_generate_prime().)
13441 In the new general case, we do not insist that 'generator' is
13442 actually a primitive root: This requirement is rather pointless;
13443 a generator of the order-q subgroup is just as good, if not
13448 * Map new X509 verification errors to alerts. Discovered and submitted by
13449 Tom Wu <tom@arcot.com>.
13453 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13454 returning non-zero before the data has been completely received
13455 when using non-blocking I/O.
13457 *Bodo Moeller; problem pointed out by John Hughes*
13459 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13461 *Ben Laurie, Lutz Jaenicke*
13463 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13464 Yoram Zahavi <YoramZ@gilian.com>).
13468 * Add information about CygWin 1.3 and on, and preserve proper
13469 configuration for the versions before that.
13471 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13473 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13474 check whether we deal with a copy of a session and do not delete from
13475 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13476 <izhar@checkpoint.com>.
13480 * Do not store session data into the internal session cache, if it
13481 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13482 flag is set). Proposed by Aslam <aslam@funk.com>.
13486 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13491 * [In 0.9.6d-engine release:]
13492 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13494 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13496 * Add the configuration target linux-s390x.
13498 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13500 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13501 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13502 variable as an indication that a ClientHello message has been
13503 received. As the flag value will be lost between multiple
13504 invocations of ssl3_accept when using non-blocking I/O, the
13505 function may not be aware that a handshake has actually taken
13506 place, thus preventing a new session from being added to the
13509 To avoid this problem, we now set s->new_session to 2 instead of
13510 using a local variable.
13512 *Lutz Jaenicke, Bodo Moeller*
13514 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13515 if the SSL_R_LENGTH_MISMATCH error is detected.
13517 *Geoff Thorpe, Bodo Moeller*
13519 * New 'shared_ldflag' column in Configure platform table.
13523 * Fix EVP_CIPHER_mode macro.
13525 *"Dan S. Camper" <dan@bti.net>*
13527 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13528 type, we must throw them away by setting rr->length to 0.
13530 *D P Chang <dpc@qualys.com>*
13532 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13534 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13535 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13536 worked incorrectly for those cases where range = `10..._2` and
13537 `3*range` is two bits longer than range.)
13541 * Only add signing time to PKCS7 structures if it is not already
13546 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13547 OBJ_ld_ce should be OBJ_id_ce.
13548 Also some ip-pda OIDs in crypto/objects/objects.txt were
13549 incorrect (cf. RFC 3039).
13551 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13553 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13554 returns early because it has nothing to do.
13556 *Andy Schneider <andy.schneider@bjss.co.uk>*
13558 * [In 0.9.6c-engine release:]
13559 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13561 *Andy Schneider <andy.schneider@bjss.co.uk>*
13563 * [In 0.9.6c-engine release:]
13564 Add support for Cryptographic Appliance's keyserver technology.
13565 (Use engine 'keyclient')
13567 *Cryptographic Appliances and Geoff Thorpe*
13569 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13570 is called via tools/c89.sh because arguments have to be
13571 rearranged (all '-L' options must appear before the first object
13574 *Richard Shapiro <rshapiro@abinitio.com>*
13576 * [In 0.9.6c-engine release:]
13577 Add support for Broadcom crypto accelerator cards, backported
13580 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13582 * [In 0.9.6c-engine release:]
13583 Add support for SureWare crypto accelerator cards from
13584 Baltimore Technologies. (Use engine 'sureware')
13586 *Baltimore Technologies and Mark Cox*
13588 * [In 0.9.6c-engine release:]
13589 Add support for crypto accelerator cards from Accelerated
13590 Encryption Processing, www.aep.ie. (Use engine 'aep')
13592 *AEP Inc. and Mark Cox*
13594 * Add a configuration entry for gcc on UnixWare.
13596 *Gary Benson <gbenson@redhat.com>*
13598 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13599 messages are stored in a single piece (fixed-length part and
13600 variable-length part combined) and fix various bugs found on the way.
13604 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13605 instead. BIO_gethostbyname() does not know what timeouts are
13606 appropriate, so entries would stay in cache even when they have
13608 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13610 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13611 faced with a pathologically small ClientHello fragment that does
13612 not contain client_version: Instead of aborting with an error,
13613 simply choose the highest available protocol version (i.e.,
13614 TLS 1.0 unless it is disabled). In practice, ClientHello
13615 messages are never sent like this, but this change gives us
13616 strictly correct behaviour at least for TLS.
13620 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13621 never resets s->method to s->ctx->method when called from within
13622 one of the SSL handshake functions.
13624 *Bodo Moeller; problem pointed out by Niko Baric*
13626 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13627 (sent using the client's version number) if client_version is
13628 smaller than the protocol version in use. Also change
13629 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13630 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13631 the client will at least see that alert.
13635 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13640 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13641 client receives HelloRequest while in a handshake.
13643 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13645 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13646 should end in 'break', not 'goto end' which circumvents various
13647 cleanups done in state SSL_ST_OK. But session related stuff
13648 must be disabled for SSL_ST_OK in the case that we just sent a
13651 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13652 before just sending a HelloRequest.
13654 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13656 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13657 reveal whether illegal block cipher padding was found or a MAC
13658 verification error occurred. (Neither SSLerr() codes nor alerts
13659 are directly visible to potential attackers, but the information
13660 may leak via logfiles.)
13662 Similar changes are not required for the SSL 2.0 implementation
13663 because the number of padding bytes is sent in clear for SSL 2.0,
13664 and the extra bytes are just ignored. However ssl/s2_pkt.c
13665 failed to verify that the purported number of padding bytes is in
13670 * Add OpenUNIX-8 support including shared libraries
13671 (Boyd Lynn Gerber <gerberb@zenez.com>).
13675 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13676 'wristwatch attack' using huge encoding parameters (cf.
13677 James H. Manger's CRYPTO 2001 paper). Note that the
13678 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13679 encoding parameters and hence was not vulnerable.
13683 * BN_sqr() bug fix.
13685 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13687 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13688 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13689 followed by modular reduction.
13691 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13693 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13694 equivalent based on BN_pseudo_rand() instead of BN_rand().
13698 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13699 This function was broken, as the check for a new client hello message
13700 to handle SGC did not allow these large messages.
13701 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13705 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13709 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13710 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13714 * Rework the configuration and shared library support for Tru64 Unix.
13715 The configuration part makes use of modern compiler features and
13716 still retains old compiler behavior for those that run older versions
13717 of the OS. The shared library support part includes a variant that
13718 uses the RPATH feature, and is available through the special
13719 configuration target "alpha-cc-rpath", which will never be selected
13722 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13724 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13725 with the same message size as in ssl3_get_certificate_request().
13726 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13727 messages might inadvertently be reject as too long.
13729 *Petr Lampa <lampa@fee.vutbr.cz>*
13731 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13735 * Modified SSL library such that the verify_callback that has been set
13736 specifically for an SSL object with SSL_set_verify() is actually being
13737 used. Before the change, a verify_callback set with this function was
13738 ignored and the verify_callback() set in the SSL_CTX at the time of
13739 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13740 to allow the necessary settings.
13744 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13745 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13746 done automatically (in contradiction to the requirements of the C
13747 standard). This made problems when used from OpenSSH.
13751 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13752 dh->length and always used
13754 BN_rand_range(priv_key, dh->p).
13756 BN_rand_range() is not necessary for Diffie-Hellman, and this
13757 specific range makes Diffie-Hellman unnecessarily inefficient if
13758 dh->length (recommended exponent length) is much smaller than the
13759 length of dh->p. We could use BN_rand_range() if the order of
13760 the subgroup was stored in the DH structure, but we only have
13765 BN_rand(priv_key, l, ...)
13767 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13774 RSA_eay_public_encrypt
13775 RSA_eay_private_decrypt
13776 RSA_eay_private_encrypt (signing)
13777 RSA_eay_public_decrypt (signature verification)
13779 (default implementations for RSA_public_encrypt,
13780 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13781 always reject numbers >= n.
13785 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13786 to synchronize access to 'locking_thread'. This is necessary on
13787 systems where access to 'locking_thread' (an 'unsigned long'
13788 variable) is not atomic.
13792 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13793 *before* setting the 'crypto_lock_rand' flag. The previous code had
13794 a race condition if 0 is a valid thread ID.
13796 *Travis Vitek <vitek@roguewave.com>*
13798 * Add support for shared libraries under Irix.
13800 *Albert Chin-A-Young <china@thewrittenword.com>*
13802 * Add configuration option to build on Linux on both big-endian and
13803 little-endian MIPS.
13805 *Ralf Baechle <ralf@uni-koblenz.de>*
13807 * Add the possibility to create shared libraries on HP-UX.
13811 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13813 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13814 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13815 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13816 PRNG state recovery was possible based on the output of
13817 one PRNG request appropriately sized to gain knowledge on
13818 'md' followed by enough consecutive 1-byte PRNG requests
13819 to traverse all of 'state'.
13821 1. When updating 'md_local' (the current thread's copy of 'md')
13822 during PRNG output generation, hash all of the previous
13823 'md_local' value, not just the half used for PRNG output.
13825 2. Make the number of bytes from 'state' included into the hash
13826 independent from the number of PRNG bytes requested.
13828 The first measure alone would be sufficient to avoid
13829 Markku-Juhani's attack. (Actually it had never occurred
13830 to me that the half of 'md_local' used for chaining was the
13831 half from which PRNG output bytes were taken -- I had always
13832 assumed that the secret half would be used.) The second
13833 measure makes sure that additional data from 'state' is never
13834 mixed into 'md_local' in small portions; this heuristically
13835 further strengthens the PRNG.
13839 * Fix crypto/bn/asm/mips3.s.
13843 * When only the key is given to "enc", the IV is undefined. Print out
13844 an error message in this case.
13848 * Handle special case when X509_NAME is empty in X509 printing routines.
13852 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13853 positive and less than q.
13857 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13858 used: it isn't thread safe and the add_lock_callback should handle
13861 *Paul Rose <Paul.Rose@bridge.com>*
13863 * Verify that incoming data obeys the block size in
13864 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13870 *Ulf Möller, Bodo Möller*
13872 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13873 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13874 when fixing the server behaviour for backwards-compatible 'client
13875 hello' messages. (Note that the attack is impractical against
13876 SSL 3.0 and TLS 1.0 anyway because length and version checking
13877 means that the probability of guessing a valid ciphertext is
13878 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13881 Before 0.9.5, the countermeasure (hide the error by generating a
13882 random 'decryption result') did not work properly because
13883 ERR_clear_error() was missing, meaning that SSL_get_error() would
13884 detect the supposedly ignored error.
13886 Both problems are now fixed.
13890 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13891 (previously it was 1024).
13895 * Fix for compatibility mode trust settings: ignore trust settings
13896 unless some valid trust or reject settings are present.
13900 * Fix for blowfish EVP: its a variable length cipher.
13904 * Fix various bugs related to DSA S/MIME verification. Handle missing
13905 parameters in DSA public key structures and return an error in the
13906 DSA routines if parameters are absent.
13910 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13911 in the current directory if neither $RANDFILE nor $HOME was set.
13912 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13913 caused some confusion to Windows users who haven't defined $HOME.
13914 Thus RAND_file_name() is changed again: e_os.h can define a
13915 DEFAULT_HOME, which will be used if $HOME is not set.
13916 For Windows, we use "C:"; on other platforms, we still require
13917 environment variables.
13919 * Move 'if (!initialized) RAND_poll()' into regions protected by
13920 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13921 having multiple threads call RAND_poll() concurrently.
13925 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13926 combination of a flag and a thread ID variable.
13927 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13928 flag), *other* threads can enter ssleay_add_bytes without obeying
13929 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13930 that they do not hold after the first thread unsets add_do_not_lock).
13934 * Change bctest again: '-x' expressions are not available in all
13935 versions of 'test'.
13939 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13941 * Fix a couple of memory leaks in PKCS7_dataDecode()
13943 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13945 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13946 the default extension for executables, if any. Also, make the perl
13947 scripts that use symlink() to test if it really exists and use "cp"
13948 if it doesn't. All this made OpenSSL compilable and installable in
13953 * Fix for asn1_GetSequence() for indefinite length constructed data.
13954 If SEQUENCE is length is indefinite just set c->slen to the total
13955 amount of data available.
13957 *Steve Henson, reported by shige@FreeBSD.org*
13959 *This change does not apply to 0.9.7.*
13961 * Change bctest to avoid here-documents inside command substitution
13962 (workaround for FreeBSD /bin/sh bug).
13963 For compatibility with Ultrix, avoid shell functions (introduced
13964 in the bctest version that searches along $PATH).
13968 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13969 with des_encrypt() defined on some operating systems, like Solaris
13974 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13975 On the Importance of Eliminating Errors in Cryptographic
13976 Computations, J. Cryptology 14 (2001) 2, 101-119,
13977 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13981 * MIPS assembler BIGNUM division bug fix.
13985 * Disabled incorrect Alpha assembler code.
13989 * Fix PKCS#7 decode routines so they correctly update the length
13990 after reading an EOC for the EXPLICIT tag.
13994 *This change does not apply to 0.9.7.*
13996 * Fix bug in PKCS#12 key generation routines. This was triggered
13997 if a 3DES key was generated with a 0 initial byte. Include
13998 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13999 (but broken) behaviour.
14003 * Enhance bctest to search for a working bc along $PATH and print
14006 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14008 * Fix memory leaks in err.c: free err_data string if necessary;
14009 don't write to the wrong index in ERR_set_error_data.
14013 * Implement ssl23_peek (analogous to ssl23_read), which previously
14018 * Replace rdtsc with `_emit` statements for VC++ version 5.
14020 *Jeremy Cooper <jeremy@baymoo.org>*
14022 * Make it possible to reuse SSLv2 sessions.
14026 * In copy_email() check for >= 0 as a return value for
14027 X509_NAME_get_index_by_NID() since 0 is a valid index.
14029 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14031 * Avoid coredump with unsupported or invalid public keys by checking if
14032 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14033 PKCS7_verify() fails with non detached data.
14037 * Don't use getenv in library functions when run as setuid/setgid.
14038 New function OPENSSL_issetugid().
14042 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14043 due to incorrect handling of multi-threading:
14045 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14047 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14049 3. Count how many times MemCheck_off() has been called so that
14050 nested use can be treated correctly. This also avoids
14051 inband-signalling in the previous code (which relied on the
14052 assumption that thread ID 0 is impossible).
14056 * Add "-rand" option also to s_client and s_server.
14060 * Fix CPU detection on Irix 6.x.
14061 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14062 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14064 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14069 *This change does not apply to 0.9.7.*
14071 * Use the cached encoding of an X509_NAME structure rather than
14072 copying it. This is apparently the reason for the libsafe "errors"
14073 but the code is actually correct.
14077 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14078 Bleichenbacher's DSA attack.
14079 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14080 to be set and top=0 forces the highest bit to be set; top=-1 is new
14081 and leaves the highest bit random.
14083 *Ulf Moeller, Bodo Moeller*
14085 * In the `NCONF_...`-based implementations for `CONF_...` queries
14086 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14087 a temporary CONF structure with the data component set to NULL
14088 (which gives segmentation faults in lh_retrieve).
14089 Instead, use NULL for the CONF pointer in CONF_get_string and
14090 CONF_get_number (which may use environment variables) and directly
14091 return NULL from CONF_get_section.
14095 * Fix potential buffer overrun for EBCDIC.
14099 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14100 keyUsage if basicConstraints absent for a CA.
14104 * Make SMIME_write_PKCS7() write mail header values with a format that
14105 is more generally accepted (no spaces before the semicolon), since
14106 some programs can't parse those values properly otherwise. Also make
14107 sure BIO's that break lines after each write do not create invalid
14112 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14113 macros previously used would not encode an empty SEQUENCE OF
14114 and break the signature.
14118 *This change does not apply to 0.9.7.*
14120 * Zero the premaster secret after deriving the master secret in
14125 * Add some EVP_add_digest_alias registrations (as found in
14126 OpenSSL_add_all_digests()) to SSL_library_init()
14127 aka OpenSSL_add_ssl_algorithms(). This provides improved
14128 compatibility with peers using X.509 certificates
14129 with unconventional AlgorithmIdentifier OIDs.
14133 * Fix for Irix with NO_ASM.
14135 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14137 * ./config script fixes.
14139 *Ulf Moeller, Richard Levitte*
14141 * Fix 'openssl passwd -1'.
14145 * Change PKCS12_key_gen_asc() so it can cope with non null
14146 terminated strings whose length is passed in the passlen
14147 parameter, for example from PEM callbacks. This was done
14148 by adding an extra length parameter to asc2uni().
14150 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14152 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14153 call failed, free the DSA structure.
14157 * Fix to uni2asc() to cope with zero length Unicode strings.
14158 These are present in some PKCS#12 files.
14162 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14163 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14164 when writing a 32767 byte record.
14166 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14168 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14169 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14171 (RSA objects have a reference count access to which is protected
14172 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14173 so they are meant to be shared between threads.)
14174 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14175 "Reddie, Steven" <Steven.Reddie@ca.com>*
14177 * Fix a deadlock in CRYPTO_mem_leaks().
14181 * Use better test patterns in bntest.
14185 * rand_win.c fix for Borland C.
14189 * BN_rshift bugfix for n == 0.
14193 * Add a 'bctest' script that checks for some known 'bc' bugs
14194 so that 'make test' does not abort just because 'bc' is broken.
14198 * Store verify_result within SSL_SESSION also for client side to
14199 avoid potential security hole. (Re-used sessions on the client side
14200 always resulted in verify_result==X509_V_OK, not using the original
14201 result of the server certificate verification.)
14205 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14206 SSL3_RT_APPLICATION_DATA, return 0.
14207 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14212 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14213 releases, have been re-implemented by renaming the previous
14214 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14215 and ssl3_read_internal, respectively, and adding 'peek' parameters
14216 to them. The new ssl[23]_{read,peek} functions are calls to
14217 ssl[23]_read_internal with the 'peek' flag set appropriately.
14218 A 'peek' parameter has also been added to ssl3_read_bytes, which
14219 does the actual work for ssl3_read_internal.
14223 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14224 the method-specific "init()" handler. Also clean up ex_data after
14225 calling the method-specific "finish()" handler. Previously, this was
14226 happening the other way round.
14230 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14231 The previous value, 12, was not always sufficient for BN_mod_exp().
14235 * Make sure that shared libraries get the internal name engine with
14236 the full version number and not just 0. This should mark the
14237 shared libraries as not backward compatible. Of course, this should
14238 be changed again when we can guarantee backward binary compatibility.
14242 * Fix typo in get_cert_by_subject() in by_dir.c
14244 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14246 * Rework the system to generate shared libraries:
14248 - Make note of the expected extension for the shared libraries and
14249 if there is a need for symbolic links from for example libcrypto.so.0
14250 to libcrypto.so.0.9.7. There is extended info in Configure for
14253 - Make as few rebuilds of the shared libraries as possible.
14255 - Still avoid linking the OpenSSL programs with the shared libraries.
14257 - When installing, install the shared libraries separately from the
14262 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14264 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14265 and not in SSL_clear because the latter is also used by the
14266 accept/connect functions; previously, the settings made by
14267 SSL_set_read_ahead would be lost during the handshake.
14269 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14271 * Correct util/mkdef.pl to be selective about disabled algorithms.
14272 Previously, it would create entries for disabled algorithms no
14277 * Added several new manual pages for SSL_* function.
14281 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14283 * In ssl23_get_client_hello, generate an error message when faced
14284 with an initial SSL 3.0/TLS record that is too small to contain the
14285 first two bytes of the ClientHello message, i.e. client_version.
14286 (Note that this is a pathologic case that probably has never happened
14287 in real life.) The previous approach was to use the version number
14288 from the record header as a substitute; but our protocol choice
14289 should not depend on that one because it is not authenticated
14290 by the Finished messages.
14294 * More robust randomness gathering functions for Windows.
14296 *Jeffrey Altman <jaltman@columbia.edu>*
14298 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14299 not set then we don't setup the error code for issuer check errors
14300 to avoid possibly overwriting other errors which the callback does
14301 handle. If an application does set the flag then we assume it knows
14302 what it is doing and can handle the new informational codes
14307 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14308 a general "ANY" type, as such it should be able to decode anything
14309 including tagged types. However it didn't check the class so it would
14310 wrongly interpret tagged types in the same way as their universal
14311 counterpart and unknown types were just rejected. Changed so that the
14312 tagged and unknown types are handled in the same way as a SEQUENCE:
14313 that is the encoding is stored intact. There is also a new type
14314 "V_ASN1_OTHER" which is used when the class is not universal, in this
14315 case we have no idea what the actual type is so we just lump them all
14320 * On VMS, stdout may very well lead to a file that is written to
14321 in a record-oriented fashion. That means that every write() will
14322 write a separate record, which will be read separately by the
14323 programs trying to read from it. This can be very confusing.
14325 The solution is to put a BIO filter in the way that will buffer
14326 text until a linefeed is reached, and then write everything a
14327 line at a time, so every record written will be an actual line,
14328 not chunks of lines and not (usually doesn't happen, but I've
14329 seen it once) several lines in one record. BIO_f_linebuffer() is
14332 Currently, it's a VMS-only method, because that's where it has
14333 been tested well enough.
14337 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14338 it can return incorrect results.
14339 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14340 but it was in 0.9.6-beta[12].)
14344 * Disable the check for content being present when verifying detached
14345 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14346 include zero length content when signing messages.
14350 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14351 BIO_ctrl (for BIO pairs).
14355 * Add DSO method for VMS.
14359 * Bug fix: Montgomery multiplication could produce results with the
14364 * Add RPM specification openssl.spec and modify it to build three
14365 packages. The default package contains applications, application
14366 documentation and run-time libraries. The devel package contains
14367 include files, static libraries and function documentation. The
14368 doc package contains the contents of the doc directory. The original
14369 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14373 * Add a large number of documentation files for many SSL routines.
14375 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14377 * Add a configuration entry for Sony News 4.
14379 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14381 * Don't set the two most significant bits to one when generating a
14382 random number < q in the DSA library.
14386 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14387 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14388 the underlying transport is blocking) if a handshake took place.
14389 (The default behaviour is needed by applications such as s_client
14390 and s_server that use select() to determine when to use SSL_read;
14391 but for applications that know in advance when to expect data, it
14392 just makes things more complicated.)
14396 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14401 * Add a few more EBCDIC conditionals that make `req` and `x509`
14402 work better on such systems.
14404 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14406 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14407 Update PKCS12_parse() so it copies the friendlyName and the
14408 keyid to the certificates aux info.
14412 * Fix bug in PKCS7_verify() which caused an infinite loop
14413 if there was more than one signature.
14415 *Sven Uszpelkat <su@celocom.de>*
14417 * Major change in util/mkdef.pl to include extra information
14418 about each symbol, as well as presenting variables as well
14419 as functions. This change means that there's n more need
14420 to rebuild the .num files when some algorithms are excluded.
14424 * Allow the verify time to be set by an application,
14425 rather than always using the current time.
14429 * Phase 2 verify code reorganisation. The certificate
14430 verify code now looks up an issuer certificate by a
14431 number of criteria: subject name, authority key id
14432 and key usage. It also verifies self signed certificates
14433 by the same criteria. The main comparison function is
14434 X509_check_issued() which performs these checks.
14436 Lot of changes were necessary in order to support this
14437 without completely rewriting the lookup code.
14439 Authority and subject key identifier are now cached.
14441 The LHASH 'certs' is X509_STORE has now been replaced
14442 by a STACK_OF(X509_OBJECT). This is mainly because an
14443 LHASH can't store or retrieve multiple objects with
14444 the same hash value.
14446 As a result various functions (which were all internal
14447 use only) have changed to handle the new X509_STORE
14448 structure. This will break anything that messed round
14449 with X509_STORE internally.
14451 The functions X509_STORE_add_cert() now checks for an
14452 exact match, rather than just subject name.
14454 The X509_STORE API doesn't directly support the retrieval
14455 of multiple certificates matching a given criteria, however
14456 this can be worked round by performing a lookup first
14457 (which will fill the cache with candidate certificates)
14458 and then examining the cache for matches. This is probably
14459 the best we can do without throwing out X509_LOOKUP
14460 entirely (maybe later...).
14462 The X509_VERIFY_CTX structure has been enhanced considerably.
14464 All certificate lookup operations now go via a get_issuer()
14465 callback. Although this currently uses an X509_STORE it
14466 can be replaced by custom lookups. This is a simple way
14467 to bypass the X509_STORE hackery necessary to make this
14468 work and makes it possible to use more efficient techniques
14469 in future. A very simple version which uses a simple
14470 STACK for its trusted certificate store is also provided
14471 using X509_STORE_CTX_trusted_stack().
14473 The verify_cb() and verify() callbacks now have equivalents
14474 in the X509_STORE_CTX structure.
14476 X509_STORE_CTX also has a 'flags' field which can be used
14477 to customise the verify behaviour.
14481 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14482 excludes S/MIME capabilities.
14486 * When a certificate request is read in keep a copy of the
14487 original encoding of the signed data and use it when outputting
14488 again. Signatures then use the original encoding rather than
14489 a decoded, encoded version which may cause problems if the
14490 request is improperly encoded.
14494 * For consistency with other BIO_puts implementations, call
14495 buffer_write(b, ...) directly in buffer_puts instead of calling
14498 In BIO_puts, increment b->num_write as in BIO_write.
14500 *Peter.Sylvester@EdelWeb.fr*
14502 * Fix BN_mul_word for the case where the word is 0. (We have to use
14503 BN_zero, we may not return a BIGNUM with an array consisting of
14504 words set to zero.)
14508 * Avoid calling abort() from within the library when problems are
14509 detected, except if preprocessor symbols have been defined
14510 (such as REF_CHECK, BN_DEBUG etc.).
14514 * New openssl application 'rsautl'. This utility can be
14515 used for low level RSA operations. DER public key
14516 BIO/fp routines also added.
14520 * New Configure entry and patches for compiling on QNX 4.
14522 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14524 * A demo state-machine implementation was sponsored by
14525 Nuron (<http://www.nuron.com/>) and is now available in
14526 demos/state_machine.
14530 * New options added to the 'dgst' utility for signature
14531 generation and verification.
14535 * Unrecognized PKCS#7 content types are now handled via a
14536 catch all ASN1_TYPE structure. This allows unsupported
14537 types to be stored as a "blob" and an application can
14538 encode and decode it manually.
14542 * Fix various signed/unsigned issues to make a_strex.c
14543 compile under VC++.
14545 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14547 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14548 length if passed a buffer. ASN1_INTEGER_to_BN failed
14549 if passed a NULL BN and its argument was negative.
14551 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14553 * Modification to PKCS#7 encoding routines to output definite
14554 length encoding. Since currently the whole structures are in
14555 memory there's not real point in using indefinite length
14556 constructed encoding. However if OpenSSL is compiled with
14557 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14561 * Added BIO_vprintf() and BIO_vsnprintf().
14565 * Added more prefixes to parse for in the strings written
14566 through a logging bio, to cover all the levels that are available
14567 through syslog. The prefixes are now:
14569 PANIC, EMERG, EMR => LOG_EMERG
14570 ALERT, ALR => LOG_ALERT
14571 CRIT, CRI => LOG_CRIT
14572 ERROR, ERR => LOG_ERR
14573 WARNING, WARN, WAR => LOG_WARNING
14574 NOTICE, NOTE, NOT => LOG_NOTICE
14575 INFO, INF => LOG_INFO
14576 DEBUG, DBG => LOG_DEBUG
14578 and as before, if none of those prefixes are present at the
14579 beginning of the string, LOG_ERR is chosen.
14581 On Win32, the `LOG_*` levels are mapped according to this:
14583 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14584 LOG_WARNING => EVENTLOG_WARNING_TYPE
14585 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14589 * Made it possible to reconfigure with just the configuration
14590 argument "reconf" or "reconfigure". The command line arguments
14591 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14592 and are retrieved from there when reconfiguring.
14598 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14600 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14604 * The obj_dat.pl script was messing up the sorting of object
14605 names. The reason was that it compared the quoted version
14606 of strings as a result "OCSP" > "OCSP Signing" because
14607 " > SPACE. Changed script to store unquoted versions of
14608 names and add quotes on output. It was also omitting some
14609 names from the lookup table if they were given a default
14610 value (that is if SN is missing it is given the same
14611 value as LN and vice versa), these are now added on the
14612 grounds that if an object has a name we should be able to
14613 look it up. Finally added warning output when duplicate
14614 short or long names are found.
14618 * Changes needed for Tandem NSK.
14620 *Scott Uroff <scott@xypro.com>*
14622 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14623 RSA_padding_check_SSLv23(), special padding was never detected
14624 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14625 version rollback attacks was not effective.
14627 In s23_clnt.c, don't use special rollback-attack detection padding
14628 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14629 client; similarly, in s23_srvr.c, don't do the rollback check if
14630 SSL 2.0 is the only protocol enabled in the server.
14634 * Make it possible to get hexdumps of unprintable data with 'openssl
14635 asn1parse'. By implication, the functions ASN1_parse_dump() and
14636 BIO_dump_indent() are added.
14640 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14641 these print out strings and name structures based on various
14642 flags including RFC2253 support and proper handling of
14643 multibyte characters. Added options to the 'x509' utility
14644 to allow the various flags to be set.
14648 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14649 Also change the functions X509_cmp_current_time() and
14650 X509_gmtime_adj() work with an ASN1_TIME structure,
14651 this will enable certificates using GeneralizedTime in validity
14652 dates to be checked.
14656 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14657 negative public key encodings) on by default,
14658 NO_NEG_PUBKEY_BUG can be set to disable it.
14662 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14663 content octets. An i2c_ASN1_OBJECT is unnecessary because
14664 the encoding can be trivially obtained from the structure.
14668 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14669 not read locks (`CRYPTO_r_[un]lock`).
14673 * A first attempt at creating official support for shared
14674 libraries through configuration. I've kept it so the
14675 default is static libraries only, and the OpenSSL programs
14676 are always statically linked for now, but there are
14677 preparations for dynamic linking in place.
14678 This has been tested on Linux and Tru64.
14682 * Randomness polling function for Win9x, as described in:
14683 Peter Gutmann, Software Generation of Practically Strong
14688 * Fix so PRNG is seeded in req if using an already existing
14693 * New options to smime application. -inform and -outform
14694 allow alternative formats for the S/MIME message including
14695 PEM and DER. The -content option allows the content to be
14696 specified separately. This should allow things like Netscape
14697 form signing output easier to verify.
14701 * Fix the ASN1 encoding of tags using the 'long form'.
14705 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14706 STRING types. These convert content octets to and from the
14707 underlying type. The actual tag and length octets are
14708 already assumed to have been read in and checked. These
14709 are needed because all other string types have virtually
14710 identical handling apart from the tag. By having versions
14711 of the ASN1 functions that just operate on content octets
14712 IMPLICIT tagging can be handled properly. It also allows
14713 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14714 and ASN1_INTEGER are identical apart from the tag.
14718 * Change the handling of OID objects as follows:
14720 - New object identifiers are inserted in objects.txt, following
14721 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14722 - objects.pl is used to process obj_mac.num and create a new
14724 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14727 This is currently kind of a hack, and the perl code in objects.pl
14728 isn't very elegant, but it works as I intended. The simplest way
14729 to check that it worked correctly is to look in obj_dat.h and
14730 check the array nid_objs and make sure the objects haven't moved
14731 around (this is important!). Additions are OK, as well as
14732 consistent name changes.
14736 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14740 * Addition of the command line parameter '-rand file' to 'openssl req'.
14741 The given file adds to whatever has already been seeded into the
14742 random pool through the RANDFILE configuration file option or
14743 environment variable, or the default random state file.
14747 * mkstack.pl now sorts each macro group into lexical order.
14748 Previously the output order depended on the order the files
14749 appeared in the directory, resulting in needless rewriting
14754 * Patches to make OpenSSL compile under Win32 again. Mostly
14755 work arounds for the VC++ problem that it treats func() as
14756 func(void). Also stripped out the parts of mkdef.pl that
14757 added extra typesafe functions: these no longer exist.
14761 * Reorganisation of the stack code. The macros are now all
14762 collected in safestack.h . Each macro is defined in terms of
14763 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14764 DEBUG_SAFESTACK is now handled in terms of function casts,
14765 this has the advantage of retaining type safety without the
14766 use of additional functions. If DEBUG_SAFESTACK is not defined
14767 then the non typesafe macros are used instead. Also modified the
14768 mkstack.pl script to handle the new form. Needs testing to see
14769 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14770 the default if no major problems. Similar behaviour for ASN1_SET_OF
14771 and PKCS12_STACK_OF.
14775 * When some versions of IIS use the 'NET' form of private key the
14776 key derivation algorithm is different. Normally MD5(password) is
14777 used as a 128 bit RC4 key. In the modified case
14778 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14779 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14780 as the old Netscape_RSA functions except they have an additional
14781 'sgckey' parameter which uses the modified algorithm. Also added
14782 an -sgckey command line option to the rsa utility. Thanks to
14783 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14784 algorithm to openssl-dev.
14788 * The evp_local.h macros were using 'c.##kname' which resulted in
14789 invalid expansion on some systems (SCO 5.0.5 for example).
14790 Corrected to 'c.kname'.
14792 *Phillip Porch <root@theporch.com>*
14794 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14795 a STACK of email addresses from a certificate or request, these look
14796 in the subject name and the subject alternative name extensions and
14797 omit any duplicate addresses.
14801 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14802 This makes DSA verification about 2 % faster.
14806 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14807 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14808 plus overhead for 1024 bit moduli).
14809 This makes exponentiations about 0.5 % faster for 1024 bit
14810 exponents (as measured by "openssl speed rsa2048").
14814 * Rename memory handling macros to avoid conflicts with other
14816 Malloc => OPENSSL_malloc
14817 Malloc_locked => OPENSSL_malloc_locked
14818 Realloc => OPENSSL_realloc
14819 Free => OPENSSL_free
14823 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14824 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14828 * CygWin32 support.
14830 *John Jarvie <jjarvie@newsguy.com>*
14832 * The type-safe stack code has been rejigged. It is now only compiled
14833 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14834 by default all type-specific stack functions are "#define"d back to
14835 standard stack functions. This results in more streamlined output
14836 but retains the type-safety checking possibilities of the original
14841 * The STACK code has been cleaned up, and certain type declarations
14842 that didn't make a lot of sense have been brought in line. This has
14843 also involved a cleanup of sorts in safestack.h to more correctly
14844 map type-safe stack functions onto their plain stack counterparts.
14845 This work has also resulted in a variety of "const"ifications of
14846 lots of the code, especially `_cmp` operations which should normally
14847 be prototyped with "const" parameters anyway.
14851 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14852 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14853 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14854 where all of 'md' is used each time the PRNG is used, but 'state'
14855 is used only indexed by a cyclic counter. As entropy may not be
14856 well distributed from the beginning, 'md' is important as a
14857 chaining variable. However, the output function chains only half
14858 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14859 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14860 in all of 'state' being rewritten, with the new values depending
14861 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14865 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14866 the handshake is continued after ssl_verify_cert_chain();
14867 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14868 can lead to 'unexplainable' connection aborts later.
14870 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14872 * Major EVP API cipher revision.
14873 Add hooks for extra EVP features. This allows various cipher
14874 parameters to be set in the EVP interface. Support added for variable
14875 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14876 setting of RC2 and RC5 parameters.
14878 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14881 Remove lots of duplicated code from the EVP library. For example *every*
14882 cipher init() function handles the 'iv' in the same way according to the
14883 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14884 for CFB and OFB modes they zero ctx->num.
14886 New functionality allows removal of S/MIME code RC2 hack.
14888 Most of the routines have the same form and so can be declared in terms
14891 By shifting this to the top level EVP_CipherInit() it can be removed from
14892 all individual ciphers. If the cipher wants to handle IVs or keys
14893 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14896 Change lots of functions like EVP_EncryptUpdate() to now return a
14897 value: although software versions of the algorithms cannot fail
14898 any installed hardware versions can.
14902 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14903 this option is set, tolerate broken clients that send the negotiated
14904 protocol version number instead of the requested protocol version
14909 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14910 i.e. non-zero for export ciphersuites, zero otherwise.
14911 Previous versions had this flag inverted, inconsistent with
14912 rsa_tmp_cb (..._TMP_RSA_CB).
14914 *Bodo Moeller; problem reported by Amit Chopra*
14916 * Add missing DSA library text string. Work around for some IIS
14917 key files with invalid SEQUENCE encoding.
14921 * Add a document (doc/standards.txt) that list all kinds of standards
14922 and so on that are implemented in OpenSSL.
14926 * Enhance c_rehash script. Old version would mishandle certificates
14927 with the same subject name hash and wouldn't handle CRLs at all.
14928 Added -fingerprint option to crl utility, to support new c_rehash
14933 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14937 * Fix for SSL server purpose checking. Server checking was
14938 rejecting certificates which had extended key usage present
14939 but no ssl client purpose.
14941 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14943 * Make PKCS#12 code work with no password. The PKCS#12 spec
14944 is a little unclear about how a blank password is handled.
14945 Since the password in encoded as a BMPString with terminating
14946 double NULL a zero length password would end up as just the
14947 double NULL. However no password at all is different and is
14948 handled differently in the PKCS#12 key generation code. NS
14949 treats a blank password as zero length. MSIE treats it as no
14950 password on export: but it will try both on import. We now do
14951 the same: PKCS12_parse() tries zero length and no password if
14952 the password is set to "" or NULL (NULL is now a valid password:
14953 it wasn't before) as does the pkcs12 application.
14957 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14958 perror when PEM_read_bio_X509_REQ fails, the error message must
14959 be obtained from the error queue.
14963 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14964 it in ERR_remove_state if appropriate, and change ERR_get_state
14965 accordingly to avoid race conditions (this is necessary because
14966 thread_hash is no longer constant once set).
14970 * Bugfix for linux-elf makefile.one.
14974 * RSA_get_default_method() will now cause a default
14975 RSA_METHOD to be chosen if one doesn't exist already.
14976 Previously this was only set during a call to RSA_new()
14977 or RSA_new_method(NULL) meaning it was possible for
14978 RSA_get_default_method() to return NULL.
14982 * Added native name translation to the existing DSO code
14983 that will convert (if the flag to do so is set) filenames
14984 that are sufficiently small and have no path information
14985 into a canonical native form. Eg. "blah" converted to
14986 "libblah.so" or "blah.dll" etc.
14990 * New function ERR_error_string_n(e, buf, len) which is like
14991 ERR_error_string(e, buf), but writes at most 'len' bytes
14992 including the 0 terminator. For ERR_error_string_n, 'buf'
14995 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14997 * CONF library reworked to become more general. A new CONF
14998 configuration file reader "class" is implemented as well as a
14999 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15000 old `CONF_*` functions are still there, but are reimplemented to
15001 work in terms of the new functions. Also, a set of functions
15002 to handle the internal storage of the configuration data is
15003 provided to make it easier to write new configuration file
15004 reader "classes" (I can definitely see something reading a
15005 configuration file in XML format, for example), called `_CONF_*`,
15006 or "the configuration storage API"...
15008 The new configuration file reading functions are:
15010 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15011 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15013 NCONF_default, NCONF_WIN32
15015 NCONF_dump_fp, NCONF_dump_bio
15017 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15018 NCONF_new creates a new CONF object. This works in the same way
15019 as other interfaces in OpenSSL, like the BIO interface.
15020 `NCONF_dump_*` dump the internal storage of the configuration file,
15021 which is useful for debugging. All other functions take the same
15022 arguments as the old `CONF_*` functions with the exception of the
15023 first that must be a `CONF *` instead of a `LHASH *`.
15025 To make it easier to use the new classes with the old `CONF_*` functions,
15026 the function CONF_set_default_method is provided.
15030 * Add '-tls1' option to 'openssl ciphers', which was already
15031 mentioned in the documentation but had not been implemented.
15032 (This option is not yet really useful because even the additional
15033 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15037 * Initial DSO code added into libcrypto for letting OpenSSL (and
15038 OpenSSL-based applications) load shared libraries and bind to
15039 them in a portable way.
15041 *Geoff Thorpe, with contributions from Richard Levitte*
15043 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15045 * Make sure _lrotl and _lrotr are only used with MSVC.
15047 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15048 (the default implementation of RAND_status).
15050 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15051 to '-clrext' (= clear extensions), as intended and documented.
15052 *Bodo Moeller; inconsistency pointed out by Michael Attili
15053 <attili@amaxo.com>*
15055 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15056 was larger than the MD block size.
15058 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15060 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15061 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15062 using the passed key: if the passed key was a private key the result
15063 of X509_print(), for example, would be to print out all the private key
15068 * des_quad_cksum() byte order bug fix.
15069 *Ulf Möller, using the problem description in krb4-0.9.7, where
15070 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15072 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15075 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15077 * For easily testing in shell scripts whether some command
15078 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15079 returns with exit code 0 iff no command of the given name is available.
15080 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15081 the output goes to stdout and nothing is printed to stderr.
15082 Additional arguments are always ignored.
15084 Since for each cipher there is a command of the same name,
15085 the 'no-cipher' compilation switches can be tested this way.
15087 ('openssl no-XXX' is not able to detect pseudo-commands such
15088 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15092 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15096 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15097 is set; it will be thrown away anyway because each handshake creates
15099 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15100 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15101 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15102 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15106 * New s_client option -ign_eof: EOF at stdin is ignored, and
15107 'Q' and 'R' lose their special meanings (quit/renegotiate).
15108 This is part of what -quiet does; unlike -quiet, -ign_eof
15109 does not suppress any output.
15113 * Add compatibility options to the purpose and trust code. The
15114 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15115 accepts a certificate or CA, this was the previous behaviour,
15116 with all the associated security issues.
15118 X509_TRUST_COMPAT is the old trust behaviour: only and
15119 automatically trust self signed roots in certificate store. A
15120 new trust setting X509_TRUST_DEFAULT is used to specify that
15121 a purpose has no associated trust setting and it should instead
15122 use the value in the default purpose.
15126 * Fix the PKCS#8 DSA private key code so it decodes keys again
15127 and fix a memory leak.
15131 * In util/mkerr.pl (which implements 'make errors'), preserve
15132 reason strings from the previous version of the .c file, as
15133 the default to have only downcase letters (and digits) in
15134 automatically generated reasons codes is not always appropriate.
15138 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15139 using strerror. Previously, ERR_reason_error_string() returned
15140 library names as reason strings for SYSerr; but SYSerr is a special
15141 case where small numbers are errno values, not library numbers.
15145 * Add '-dsaparam' option to 'openssl dhparam' application. This
15146 converts DSA parameters into DH parameters. (When creating parameters,
15147 DSA_generate_parameters is used.)
15151 * Include 'length' (recommended exponent length) in C code generated
15152 by 'openssl dhparam -C'.
15156 * The second argument to set_label in perlasm was already being used
15157 so couldn't be used as a "file scope" flag. Moved to third argument
15162 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15163 instead of RAND_bytes for encryption IVs and salts.
15167 * Include RAND_status() into RAND_METHOD instead of implementing
15168 it only for md_rand.c Otherwise replacing the PRNG by calling
15169 RAND_set_rand_method would be impossible.
15173 * Don't let DSA_generate_key() enter an infinite loop if the random
15174 number generation fails.
15178 * New 'rand' application for creating pseudo-random output.
15182 * Added configuration support for Linux/IA64
15184 *Rolf Haberrecker <rolf@suse.de>*
15186 * Assembler module support for Mingw32.
15190 * Shared library support for HPUX (in shlib/).
15192 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15194 * Shared library support for Solaris gcc.
15196 *Lutz Behnke <behnke@trustcenter.de>*
15198 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15200 * PKCS7_encrypt() was adding text MIME headers twice because they
15201 were added manually and by SMIME_crlf_copy().
15205 * In bntest.c don't call BN_rand with zero bits argument.
15207 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15209 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15210 case was implemented. This caused BN_div_recp() to fail occasionally.
15214 * Add an optional second argument to the set_label() in the perl
15215 assembly language builder. If this argument exists and is set
15216 to 1 it signals that the assembler should use a symbol whose
15217 scope is the entire file, not just the current function. This
15218 is needed with MASM which uses the format label:: for this scope.
15220 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15222 * Change the ASN1 types so they are typedefs by default. Before
15223 almost all types were #define'd to ASN1_STRING which was causing
15224 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15229 * Change names of new functions to the new get1/get0 naming
15230 convention: After 'get1', the caller owns a reference count
15231 and has to call `..._free`; 'get0' returns a pointer to some
15232 data structure without incrementing reference counters.
15233 (Some of the existing 'get' functions increment a reference
15234 counter, some don't.)
15235 Similarly, 'set1' and 'add1' functions increase reference
15236 counters or duplicate objects.
15240 * Allow for the possibility of temp RSA key generation failure:
15241 the code used to assume it always worked and crashed on failure.
15245 * Fix potential buffer overrun problem in BIO_printf().
15246 *Ulf Möller, using public domain code by Patrick Powell; problem
15247 pointed out by David Sacerdote <das33@cornell.edu>*
15249 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15250 RAND_egd() and RAND_status(). In the command line application,
15251 the EGD socket can be specified like a seed file using RANDFILE
15256 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15257 Some CAs (e.g. Verisign) distribute certificates in this form.
15261 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15262 list to exclude them. This means that no special compilation option
15263 is needed to use anonymous DH: it just needs to be included in the
15268 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15269 EVP_MD_type. The old functionality is available in a new macro called
15270 EVP_MD_md(). Change code that uses it and update docs.
15274 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15275 where the `void *` argument is replaced by a function pointer argument.
15276 Previously `void *` was abused to point to functions, which works on
15277 many platforms, but is not correct. As these functions are usually
15278 called by macros defined in OpenSSL header files, most source code
15279 should work without changes.
15283 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15284 sections with information on -D... compiler switches used for
15285 compiling the library so that applications can see them. To enable
15286 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15287 must be defined. E.g.,
15288 #define OPENSSL_ALGORITHM_DEFINES
15289 #include <openssl/opensslconf.h>
15290 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15292 *Richard Levitte, Ulf and Bodo Möller*
15294 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15299 * Change the 'other' type in certificate aux info to a STACK_OF
15300 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15301 the required ASN1 format: arbitrary types determined by an OID.
15305 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15306 argument to 'req'. This is not because the function is newer or
15307 better than others it just uses the work 'NEW' in the certificate
15308 request header lines. Some software needs this.
15312 * Reorganise password command line arguments: now passwords can be
15313 obtained from various sources. Delete the PEM_cb function and make
15314 it the default behaviour: i.e. if the callback is NULL and the
15315 usrdata argument is not NULL interpret it as a null terminated pass
15316 phrase. If usrdata and the callback are NULL then the pass phrase
15317 is prompted for as usual.
15321 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15322 the support is automatically enabled. The resulting binaries will
15323 autodetect the card and use it if present.
15325 *Ben Laurie and Compaq Inc.*
15327 * Work around for Netscape hang bug. This sends certificate request
15328 and server done in one record. Since this is perfectly legal in the
15329 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15330 the bugs/SSLv3 entry for more info.
15334 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15338 * Add -rand argument to smime and pkcs12 applications and read/write
15343 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15347 * Add command line password options to the remaining applications.
15351 * Bug fix for BN_div_recp() for numerators with an even number of
15356 * More tests in bntest.c, and changed test_bn output.
15360 * ./config recognizes MacOS X now.
15364 * Bug fix for BN_div() when the first words of num and divisor are
15365 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15369 * Add support for various broken PKCS#8 formats, and command line
15370 options to produce them.
15374 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15375 get temporary BIGNUMs from a BN_CTX.
15379 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15384 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15385 include a #define from the old name to the new. The original intent
15386 was that statically linked binaries could for example just call
15387 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15388 link with digests. This never worked because SSLeay_add_all_digests()
15389 and SSLeay_add_all_ciphers() were in the same source file so calling
15390 one would link with the other. They are now in separate source files.
15394 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15398 * Use a less unusual form of the Miller-Rabin primality test (it used
15399 a binary algorithm for exponentiation integrated into the Miller-Rabin
15400 loop, our standard modexp algorithms are faster).
15404 * Support for the EBCDIC character set completed.
15406 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15408 * Source code cleanups: use const where appropriate, eliminate casts,
15409 use `void *` instead of `char *` in lhash.
15413 * Bugfix: ssl3_send_server_key_exchange was not restartable
15414 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15415 this the server could overwrite ephemeral keys that the client
15420 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15421 using 50 iterations of the Rabin-Miller test.
15423 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15424 iterations of the Rabin-Miller test as required by the appendix
15425 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15426 As BN_is_prime_fasttest includes trial division, DSA parameter
15427 generation becomes much faster.
15429 This implies a change for the callback functions in DSA_is_prime
15430 and DSA_generate_parameters: The callback function is called once
15431 for each positive witness in the Rabin-Miller test, not just
15432 occasionally in the inner loop; and the parameters to the
15433 callback function now provide an iteration count for the outer
15434 loop rather than for the current invocation of the inner loop.
15435 DSA_generate_parameters additionally can call the callback
15436 function with an 'iteration count' of -1, meaning that a
15437 candidate has passed the trial division test (when q is generated
15438 from an application-provided seed, trial division is skipped).
15442 * New function BN_is_prime_fasttest that optionally does trial
15443 division before starting the Rabin-Miller test and has
15444 an additional BN_CTX * argument (whereas BN_is_prime always
15445 has to allocate at least one BN_CTX).
15446 'callback(1, -1, cb_arg)' is called when a number has passed the
15447 trial division stage.
15451 * Fix for bug in CRL encoding. The validity dates weren't being handled
15456 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15460 * New function BN_pseudo_rand().
15464 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15465 bignum version of BN_from_montgomery() with the working code from
15466 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15471 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15472 made it impossible to use the same SSL_SESSION data structure in
15473 SSL2 clients in multiple threads.
15477 * The return value of RAND_load_file() no longer counts bytes obtained
15478 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15479 to seed the PRNG (previously an explicit byte count was required).
15481 *Ulf Möller, Bodo Möller*
15483 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15484 used `char *` instead of `void *` and had casts all over the place.
15488 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15492 * Retain source code compatibility for BN_prime_checks macro:
15493 BN_is_prime(..., BN_prime_checks, ...) now uses
15494 BN_prime_checks_for_size to determine the appropriate number of
15495 Rabin-Miller iterations.
15499 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15500 DH_CHECK_P_NOT_SAFE_PRIME.
15501 (Check if this is true? OpenPGP calls them "strong".)
15505 * Merge the functionality of "dh" and "gendh" programs into a new program
15506 "dhparam". The old programs are retained for now but will handle DH keys
15507 (instead of parameters) in future.
15511 * Make the ciphers, s_server and s_client programs check the return values
15512 when a new cipher list is set.
15516 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15517 ciphers. Before when the 56bit ciphers were enabled the sorting was
15520 The syntax for the cipher sorting has been extended to support sorting by
15521 cipher-strength (using the strength_bits hard coded in the tables).
15522 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15524 Fix a bug in the cipher-command parser: when supplying a cipher command
15525 string with an "undefined" symbol (neither command nor alphanumeric
15526 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15527 an error is flagged.
15529 Due to the strength-sorting extension, the code of the
15530 ssl_create_cipher_list() function was completely rearranged. I hope that
15531 the readability was also increased :-)
15533 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15535 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15536 for the first serial number and places 2 in the serial number file. This
15537 avoids problems when the root CA is created with serial number zero and
15538 the first user certificate has the same issuer name and serial number
15543 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15544 the new code. Add documentation for this stuff.
15548 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15549 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15550 structures and behave in an analogous way to the X509v3 functions:
15551 they shouldn't be called directly but wrapper functions should be used
15554 So we also now have some wrapper functions that call the X509at functions
15555 when passed certificate requests. (TO DO: similar things can be done with
15556 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15557 things. Some of these need some d2i or i2d and print functionality
15558 because they handle more complex structures.)
15562 * Add missing #ifndefs that caused missing symbols when building libssl
15563 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15564 NO_RSA in `ssl/s2*.c`.
15566 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15568 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15569 has a return value which indicates the quality of the random data
15570 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15571 error queue. New function RAND_pseudo_bytes() generates output that is
15572 guaranteed to be unique but not unpredictable. RAND_add is like
15573 RAND_seed, but takes an extra argument for an entropy estimate
15574 (RAND_seed always assumes full entropy).
15578 * Do more iterations of Rabin-Miller probable prime test (specifically,
15579 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15580 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15581 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15582 false-positive rate of at most 2^-80 for random input.
15586 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15590 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15591 in the 0.9.5 release), this returns the chain
15592 from an X509_CTX structure with a dup of the stack and all
15593 the X509 reference counts upped: so the stack will exist
15594 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15597 Also make SSL_SESSION_print() print out the verify return
15602 * Add manpage for the pkcs12 command. Also change the default
15603 behaviour so MAC iteration counts are used unless the new
15604 -nomaciter option is used. This improves file security and
15605 only older versions of MSIE (4.0 for example) need it.
15609 * Honor the no-xxx Configure options when creating .DEF files.
15613 * Add PKCS#10 attributes to field table: challengePassword,
15614 unstructuredName and unstructuredAddress. These are taken from
15615 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15616 international characters are used.
15618 More changes to X509_ATTRIBUTE code: allow the setting of types
15619 based on strings. Remove the 'loc' parameter when adding
15620 attributes because these will be a SET OF encoding which is sorted
15625 * Initial changes to the 'req' utility to allow request generation
15626 automation. This will allow an application to just generate a template
15627 file containing all the field values and have req construct the
15630 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15631 used all over the place including certificate requests and PKCS#7
15632 structures. They are currently handled manually where necessary with
15633 some primitive wrappers for PKCS#7. The new functions behave in a
15634 manner analogous to the X509 extension functions: they allow
15635 attributes to be looked up by NID and added.
15637 Later something similar to the X509V3 code would be desirable to
15638 automatically handle the encoding, decoding and printing of the
15639 more complex types. The string types like challengePassword can
15640 be handled by the string table functions.
15642 Also modified the multi byte string table handling. Now there is
15643 a 'global mask' which masks out certain types. The table itself
15644 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15645 is useful when for example there is only one permissible type
15646 (as in countryName) and using the mask might result in no valid
15651 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15652 SSL_get_peer_finished to allow applications to obtain the latest
15653 Finished messages sent to the peer or expected from the peer,
15654 respectively. (SSL_get_peer_finished is usually the Finished message
15655 actually received from the peer, otherwise the protocol will be aborted.)
15657 As the Finished message are message digests of the complete handshake
15658 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15659 be used for external authentication procedures when the authentication
15660 provided by SSL/TLS is not desired or is not enough.
15664 * Enhanced support for Alpha Linux is added. Now ./config checks if
15665 the host supports BWX extension and if Compaq C is present on the
15666 $PATH. Just exploiting of the BWX extension results in 20-30%
15667 performance kick for some algorithms, e.g. DES and RC4 to mention
15668 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15673 * Add support for MS "fast SGC". This is arguably a violation of the
15674 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15675 weak crypto and after checking the certificate is SGC a second one
15676 with strong crypto. MS SGC stops the first handshake after receiving
15677 the server certificate message and sends a second client hello. Since
15678 a server will typically do all the time consuming operations before
15679 expecting any further messages from the client (server key exchange
15680 is the most expensive) there is little difference between the two.
15682 To get OpenSSL to support MS SGC we have to permit a second client
15683 hello message after we have sent server done. In addition we have to
15684 reset the MAC if we do get this second client hello.
15688 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15689 if a DER encoded private key is RSA or DSA traditional format. Changed
15690 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15691 format DER encoded private key. Newer code should use PKCS#8 format which
15692 has the key type encoded in the ASN1 structure. Added DER private key
15693 support to pkcs8 application.
15697 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15698 ciphersuites has been selected (as required by the SSL 3/TLS 1
15699 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15700 is set, we interpret this as a request to violate the specification
15701 (the worst that can happen is a handshake failure, and 'correct'
15702 behaviour would result in a handshake failure anyway).
15706 * In SSL_CTX_add_session, take into account that there might be multiple
15707 SSL_SESSION structures with the same session ID (e.g. when two threads
15708 concurrently obtain them from an external cache).
15709 The internal cache can handle only one SSL_SESSION with a given ID,
15710 so if there's a conflict, we now throw out the old one to achieve
15715 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15716 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15717 some routines that use cipher OIDs: some ciphers do not have OIDs
15718 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15723 * Simplify the trust setting structure and code. Now we just have
15724 two sequences of OIDs for trusted and rejected settings. These will
15725 typically have values the same as the extended key usage extension
15726 and any application specific purposes.
15728 The trust checking code now has a default behaviour: it will just
15729 check for an object with the same NID as the passed id. Functions can
15730 be provided to override either the default behaviour or the behaviour
15731 for a given id. SSL client, server and email already have functions
15732 in place for compatibility: they check the NID and also return "trusted"
15733 if the certificate is self signed.
15737 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15738 traditional format into an EVP_PKEY structure.
15742 * Add a password callback function PEM_cb() which either prompts for
15743 a password if usr_data is NULL or otherwise assumes it is a null
15744 terminated password. Allow passwords to be passed on command line
15745 environment or config files in a few more utilities.
15749 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15750 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15751 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15752 Update documentation.
15756 * Support for ASN1 "NULL" type. This could be handled before by using
15757 ASN1_TYPE but there wasn't any function that would try to read a NULL
15758 and produce an error if it couldn't. For compatibility we also have
15759 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15760 don't allocate anything because they don't need to.
15764 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15767 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15769 * Rebuild of the memory allocation routines used by OpenSSL code and
15770 possibly others as well. The purpose is to make an interface that
15771 provide hooks so anyone can build a separate set of allocation and
15772 deallocation routines to be used by OpenSSL, for example memory
15773 pool implementations, or something else, which was previously hard
15774 since Malloc(), Realloc() and Free() were defined as macros having
15775 the values malloc, realloc and free, respectively (except for Win32
15776 compilations). The same is provided for memory debugging code.
15777 OpenSSL already comes with functionality to find memory leaks, but
15778 this gives people a chance to debug other memory problems.
15780 With these changes, a new set of functions and macros have appeared:
15782 CRYPTO_set_mem_debug_functions() [F]
15783 CRYPTO_get_mem_debug_functions() [F]
15784 CRYPTO_dbg_set_options() [F]
15785 CRYPTO_dbg_get_options() [F]
15786 CRYPTO_malloc_debug_init() [M]
15788 The memory debug functions are NULL by default, unless the library
15789 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15790 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15791 gives the standard debugging functions that come with OpenSSL) or
15792 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15793 provided by the library user) must be used. When the standard
15794 debugging functions are used, CRYPTO_dbg_set_options can be used to
15795 request additional information:
15796 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15797 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15799 Also, things like CRYPTO_set_mem_functions will always give the
15800 expected result (the new set of functions is used for allocation
15801 and deallocation) at all times, regardless of platform and compiler
15804 To finish it up, some functions that were never use in any other
15805 way than through macros have a new API and new semantic:
15807 CRYPTO_dbg_malloc()
15808 CRYPTO_dbg_realloc()
15811 All macros of value have retained their old syntax.
15813 *Richard Levitte and Bodo Moeller*
15815 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15816 ordering of SMIMECapabilities wasn't in "strength order" and there
15817 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15822 * Some ASN1 types with illegal zero length encoding (INTEGER,
15823 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15825 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15827 * Merge in my S/MIME library for OpenSSL. This provides a simple
15828 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15829 functionality to handle multipart/signed properly) and a utility
15830 called 'smime' to call all this stuff. This is based on code I
15831 originally wrote for Celo who have kindly allowed it to be
15832 included in OpenSSL.
15836 * Add variants des_set_key_checked and des_set_key_unchecked of
15837 des_set_key (aka des_key_sched). Global variable des_check_key
15838 decides which of these is called by des_set_key; this way
15839 des_check_key behaves as it always did, but applications and
15840 the library itself, which was buggy for des_check_key == 1,
15841 have a cleaner way to pick the version they need.
15845 * New function PKCS12_newpass() which changes the password of a
15850 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15851 dynamic mix. In both cases the ids can be used as an index into the
15852 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15853 functions so they accept a list of the field values and the
15854 application doesn't need to directly manipulate the X509_TRUST
15859 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15864 * Modify the way the V3 extension code looks up extensions. This now
15865 works in a similar way to the object code: we have some "standard"
15866 extensions in a static table which is searched with OBJ_bsearch()
15867 and the application can add dynamic ones if needed. The file
15868 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15869 updated whenever a new extension is added to the core code and kept
15870 in ext_nid order. There is a simple program 'tabtest.c' which checks
15871 this. New extensions are not added too often so this file can readily
15872 be maintained manually.
15874 There are two big advantages in doing things this way. The extensions
15875 can be looked up immediately and no longer need to be "added" using
15876 X509V3_add_standard_extensions(): this function now does nothing.
15877 Side note: I get *lots* of email saying the extension code doesn't
15878 work because people forget to call this function.
15879 Also no dynamic allocation is done unless new extensions are added:
15880 so if we don't add custom extensions there is no need to call
15881 X509V3_EXT_cleanup().
15885 * Modify enc utility's salting as follows: make salting the default. Add a
15886 magic header, so unsalted files fail gracefully instead of just decrypting
15887 to garbage. This is because not salting is a big security hole, so people
15888 should be discouraged from doing it.
15892 * Fixes and enhancements to the 'x509' utility. It allowed a message
15893 digest to be passed on the command line but it only used this
15894 parameter when signing a certificate. Modified so all relevant
15895 operations are affected by the digest parameter including the
15896 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15897 DSA key was used because it didn't fix the digest.
15901 * Initial certificate chain verify code. Currently tests the untrusted
15902 certificates for consistency with the verify purpose (which is set
15903 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15905 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15906 this is because it will reject chains with invalid extensions whereas
15907 every previous version of OpenSSL and SSLeay made no checks at all.
15909 Trust code: checks the root CA for the relevant trust settings. Trust
15910 settings have an initial value consistent with the verify purpose: e.g.
15911 if the verify purpose is for SSL client use it expects the CA to be
15912 trusted for SSL client use. However the default value can be changed to
15913 permit custom trust settings: one example of this would be to only trust
15914 certificates from a specific "secure" set of CAs.
15916 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15917 which should be used for version portability: especially since the
15918 verify structure is likely to change more often now.
15920 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15921 to set them. If not set then assume SSL clients will verify SSL servers
15924 Two new options to the verify program: -untrusted allows a set of
15925 untrusted certificates to be passed in and -purpose which sets the
15926 intended purpose of the certificate. If a purpose is set then the
15927 new chain verify code is used to check extension consistency.
15931 * Support for the authority information access extension.
15935 * Modify RSA and DSA PEM read routines to transparently handle
15936 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15937 public keys in a format compatible with certificate
15938 SubjectPublicKeyInfo structures. Unfortunately there were already
15939 functions called *_PublicKey_* which used various odd formats so
15940 these are retained for compatibility: however the DSA variants were
15941 never in a public release so they have been deleted. Changed dsa/rsa
15942 utilities to handle the new format: note no releases ever handled public
15943 keys so we should be OK.
15945 The primary motivation for this change is to avoid the same fiasco
15946 that dogs private keys: there are several incompatible private key
15947 formats some of which are standard and some OpenSSL specific and
15948 require various evil hacks to allow partial transparent handling and
15949 even then it doesn't work with DER formats. Given the option anything
15950 other than PKCS#8 should be dumped: but the other formats have to
15951 stay in the name of compatibility.
15953 With public keys and the benefit of hindsight one standard format
15954 is used which works with EVP_PKEY, RSA or DSA structures: though
15955 it clearly returns an error if you try to read the wrong kind of key.
15957 Added a -pubkey option to the 'x509' utility to output the public key.
15958 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15959 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15960 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15961 that do the same as the `EVP_PKEY_assign_*()` except they up the
15962 reference count of the added key (they don't "swallow" the
15967 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15968 CRLs would fail if the file contained no certificates or no CRLs:
15969 added a new function to read in both types and return the number
15970 read: this means that if none are read it will be an error. The
15971 DER versions of the certificate and CRL reader would always fail
15972 because it isn't possible to mix certificates and CRLs in DER format
15973 without choking one or the other routine. Changed this to just read
15974 a certificate: this is the best we can do. Also modified the code
15975 in apps/verify.c to take notice of return codes: it was previously
15976 attempting to read in certificates from NULL pointers and ignoring
15977 any errors: this is one reason why the cert and CRL reader seemed
15978 to work. It doesn't check return codes from the default certificate
15979 routines: these may well fail if the certificates aren't installed.
15983 * Code to support otherName option in GeneralName.
15987 * First update to verify code. Change the verify utility
15988 so it warns if it is passed a self signed certificate:
15989 for consistency with the normal behaviour. X509_verify
15990 has been modified to it will now verify a self signed
15991 certificate if *exactly* the same certificate appears
15992 in the store: it was previously impossible to trust a
15993 single self signed certificate. This means that:
15994 openssl verify ss.pem
15995 now gives a warning about a self signed certificate but
15996 openssl verify -CAfile ss.pem ss.pem
16001 * For servers, store verify_result in SSL_SESSION data structure
16002 (and add it to external session representation).
16003 This is needed when client certificate verifications fails,
16004 but an application-provided verification callback (set by
16005 SSL_CTX_set_cert_verify_callback) allows accepting the session
16006 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16007 but returns 1): When the session is reused, we have to set
16008 ssl->verify_result to the appropriate error code to avoid
16011 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16013 * Fix a bug in the new PKCS#7 code: it didn't consider the
16014 case in PKCS7_dataInit() where the signed PKCS7 structure
16015 didn't contain any existing data because it was being created.
16017 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16019 * Add a salt to the key derivation routines in enc.c. This
16020 forms the first 8 bytes of the encrypted file. Also add a
16021 -S option to allow a salt to be input on the command line.
16025 * New function X509_cmp(). Oddly enough there wasn't a function
16026 to compare two certificates. We do this by working out the SHA1
16027 hash and comparing that. X509_cmp() will be needed by the trust
16032 * SSL_get1_session() is like SSL_get_session(), but increments
16033 the reference count in the SSL_SESSION returned.
16035 *Geoff Thorpe <geoff@eu.c2.net>*
16037 * Fix for 'req': it was adding a null to request attributes.
16038 Also change the X509_LOOKUP and X509_INFO code to handle
16039 certificate auxiliary information.
16043 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16048 * Add the possibility to add extra information to the memory leak
16049 detecting output, to form tracebacks, showing from where each
16050 allocation was originated: CRYPTO_push_info("constant string") adds
16051 the string plus current file name and line number to a per-thread
16052 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16053 is like calling CYRPTO_pop_info() until the stack is empty.
16054 Also updated memory leak detection code to be multi-thread-safe.
16058 * Add options -text and -noout to pkcs7 utility and delete the
16059 encryption options which never did anything. Update docs.
16063 * Add options to some of the utilities to allow the pass phrase
16064 to be included on either the command line (not recommended on
16065 OSes like Unix) or read from the environment. Update the
16066 manpages and fix a few bugs.
16070 * Add a few manpages for some of the openssl commands.
16074 * Fix the -revoke option in ca. It was freeing up memory twice,
16075 leaking and not finding already revoked certificates.
16079 * Extensive changes to support certificate auxiliary information.
16080 This involves the use of X509_CERT_AUX structure and X509_AUX
16081 functions. An X509_AUX function such as PEM_read_X509_AUX()
16082 can still read in a certificate file in the usual way but it
16083 will also read in any additional "auxiliary information". By
16084 doing things this way a fair degree of compatibility can be
16085 retained: existing certificates can have this information added
16086 using the new 'x509' options.
16088 Current auxiliary information includes an "alias" and some trust
16089 settings. The trust settings will ultimately be used in enhanced
16090 certificate chain verification routines: currently a certificate
16091 can only be trusted if it is self signed and then it is trusted
16096 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16097 The problem was that one of the replacement routines had not been working
16098 since SSLeay releases. For now the offending routine has been replaced
16099 with non-optimised assembler. Even so, this now gives around 95%
16100 performance improvement for 1024 bit RSA signs.
16104 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16105 handling. Most clients have the effective key size in bits equal to
16106 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16107 A few however don't do this and instead use the size of the decrypted key
16108 to determine the RC2 key length and the AlgorithmIdentifier to determine
16109 the effective key length. In this case the effective key length can still
16110 be 40 bits but the key length can be 168 bits for example. This is fixed
16111 by manually forcing an RC2 key into the EVP_PKEY structure because the
16112 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16113 the key length and effective key length are equal.
16117 * Add a bunch of functions that should simplify the creation of
16118 X509_NAME structures. Now you should be able to do:
16119 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16120 and have it automatically work out the correct field type and fill in
16121 the structures. The more adventurous can try:
16122 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16123 and it will (hopefully) work out the correct multibyte encoding.
16127 * Change the 'req' utility to use the new field handling and multibyte
16128 copy routines. Before the DN field creation was handled in an ad hoc
16129 way in req, ca, and x509 which was rather broken and didn't support
16130 BMPStrings or UTF8Strings. Since some software doesn't implement
16131 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16132 using the dirstring_type option. See the new comment in the default
16133 openssl.cnf for more info.
16137 * Make crypto/rand/md_rand.c more robust:
16138 - Assure unique random numbers after fork().
16139 - Make sure that concurrent threads access the global counter and
16140 md serializably so that we never lose entropy in them
16141 or use exactly the same state in multiple threads.
16142 Access to the large state is not always serializable because
16143 the additional locking could be a performance killer, and
16144 md should be large enough anyway.
16148 * New file apps/app_rand.c with commonly needed functionality
16149 for handling the random seed file.
16151 Use the random seed file in some applications that previously did not:
16153 dsaparam -genkey (which also ignored its '-rand' option),
16156 x509 (when signing).
16157 Except on systems with /dev/urandom, it is crucial to have a random
16158 seed file at least for key creation, DSA signing, and for DH exchanges;
16159 for RSA signatures we could do without one.
16161 gendh and gendsa (unlike genrsa) used to read only the first byte
16162 of each file listed in the '-rand' option. The function as previously
16163 found in genrsa is now in app_rand.c and is used by all programs
16164 that support '-rand'.
16168 * In RAND_write_file, use mode 0600 for creating files;
16169 don't just chmod when it may be too late.
16173 * Report an error from X509_STORE_load_locations
16174 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16178 * New function ASN1_mbstring_copy() this copies a string in either
16179 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16180 into an ASN1_STRING type. A mask of permissible types is passed
16181 and it chooses the "minimal" type to use or an error if not type
16186 * Add function equivalents to the various macros in asn1.h. The old
16187 macros are retained with an `M_` prefix. Code inside the library can
16188 use the `M_` macros. External code (including the openssl utility)
16189 should *NOT* in order to be "shared library friendly".
16193 * Add various functions that can check a certificate's extensions
16194 to see if it usable for various purposes such as SSL client,
16195 server or S/MIME and CAs of these types. This is currently
16196 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16197 verification. Also added a -purpose flag to x509 utility to
16198 print out all the purposes.
16202 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16207 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16208 for, obtain and decode and extension and obtain its critical flag.
16209 This allows all the necessary extension code to be handled in a
16210 single function call.
16214 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16215 platforms. See crypto/rc4/rc4_enc.c for further details.
16219 * New -noout option to asn1parse. This causes no output to be produced
16220 its main use is when combined with -strparse and -out to extract data
16221 from a file (which may not be in ASN.1 format).
16225 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16226 when producing the local key id.
16228 *Richard Levitte <levitte@stacken.kth.se>*
16230 * New option -dhparam in s_server. This allows a DH parameter file to be
16231 stated explicitly. If it is not stated then it tries the first server
16232 certificate file. The previous behaviour hard coded the filename
16237 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16238 a public key to be input or output. For example:
16239 openssl rsa -in key.pem -pubout -out pubkey.pem
16240 Also added necessary DSA public key functions to handle this.
16244 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16245 in the message. This was handled by allowing
16246 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16248 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16250 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16251 to the end of the strings whereas this didn't. This would cause problems
16252 if strings read with d2i_ASN1_bytes() were later modified.
16254 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16256 * Fix for base64 decode bug. When a base64 bio reads only one line of
16257 data and it contains EOF it will end up returning an error. This is
16258 caused by input 46 bytes long. The cause is due to the way base64
16259 BIOs find the start of base64 encoded data. They do this by trying a
16260 trial decode on each line until they find one that works. When they
16261 do a flag is set and it starts again knowing it can pass all the
16262 data directly through the decoder. Unfortunately it doesn't reset
16263 the context it uses. This means that if EOF is reached an attempt
16264 is made to pass two EOFs through the context and this causes the
16265 resulting error. This can also cause other problems as well. As is
16266 usual with these problems it takes *ages* to find and the fix is
16267 trivial: move one line.
16269 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16271 * Ugly workaround to get s_client and s_server working under Windows. The
16272 old code wouldn't work because it needed to select() on sockets and the
16273 tty (for keypresses and to see if data could be written). Win32 only
16274 supports select() on sockets so we select() with a 1s timeout on the
16275 sockets and then see if any characters are waiting to be read, if none
16276 are present then we retry, we also assume we can always write data to
16277 the tty. This isn't nice because the code then blocks until we've
16278 received a complete line of data and it is effectively polling the
16279 keyboard at 1s intervals: however it's quite a bit better than not
16280 working at all :-) A dedicated Windows application might handle this
16281 with an event loop for example.
16285 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16286 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16287 will be called when RSA_sign() and RSA_verify() are used. This is useful
16288 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16289 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16290 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16291 This necessitated the support of an extra signature type NID_md5_sha1
16292 for SSL signatures and modifications to the SSL library to use it instead
16293 of calling RSA_public_decrypt() and RSA_private_encrypt().
16297 * Add new -verify -CAfile and -CApath options to the crl program, these
16298 will lookup a CRL issuers certificate and verify the signature in a
16299 similar way to the verify program. Tidy up the crl program so it
16300 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16301 less strict. It will now permit CRL extensions even if it is not
16302 a V2 CRL: this will allow it to tolerate some broken CRLs.
16306 * Initialize all non-automatic variables each time one of the openssl
16307 sub-programs is started (this is necessary as they may be started
16308 multiple times from the "OpenSSL>" prompt).
16310 *Lennart Bang, Bodo Moeller*
16312 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16313 removing all other RSA functionality (this is what NO_RSA does). This
16314 is so (for example) those in the US can disable those operations covered
16315 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16320 * Non-copying interface to BIO pairs.
16321 (still largely untested)
16325 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16326 ASCII string. This was handled independently in various places before.
16330 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16331 UTF8 strings a character at a time.
16335 * Use client_version from client hello to select the protocol
16336 (s23_srvr.c) and for RSA client key exchange verification
16337 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16341 * Add various utility functions to handle SPKACs, these were previously
16342 handled by poking round in the structure internals. Added new function
16343 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16344 print, verify and generate SPKACs. Based on an original idea from
16345 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16349 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16353 * Allow the config file extension section to be overwritten on the
16354 command line. Based on an original idea from Massimiliano Pala
16355 <madwolf@comune.modena.it>. The new option is called -extensions
16356 and can be applied to ca, req and x509. Also -reqexts to override
16357 the request extensions in req and -crlexts to override the crl extensions
16362 * Add new feature to the SPKAC handling in ca. Now you can include
16363 the same field multiple times by preceding it by "XXXX." for example:
16366 this is the same syntax as used in the req config file.
16370 * Allow certificate extensions to be added to certificate requests. These
16371 are specified in a 'req_extensions' option of the req section of the
16372 config file. They can be printed out with the -text option to req but
16373 are otherwise ignored at present.
16377 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16378 data read consists of only the final block it would not decrypted because
16379 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16380 A misplaced 'break' also meant the decrypted final block might not be
16381 copied until the next read.
16385 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16386 a few extra parameters to the DH structure: these will be useful if
16387 for example we want the value of 'q' or implement X9.42 DH.
16391 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16392 provides hooks that allow the default DSA functions or functions on a
16393 "per key" basis to be replaced. This allows hardware acceleration and
16394 hardware key storage to be handled without major modification to the
16395 library. Also added low level modexp hooks and CRYPTO_EX structure and
16396 associated functions.
16400 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16401 as "read only": it can't be written to and the buffer it points to will
16402 not be freed. Reading from a read only BIO is much more efficient than
16403 a normal memory BIO. This was added because there are several times when
16404 an area of memory needs to be read from a BIO. The previous method was
16405 to create a memory BIO and write the data to it, this results in two
16406 copies of the data and an O(n^2) reading algorithm. There is a new
16407 function BIO_new_mem_buf() which creates a read only memory BIO from
16408 an area of memory. Also modified the PKCS#7 routines to use read only
16413 * Bugfix: ssl23_get_client_hello did not work properly when called in
16414 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16415 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16416 but a retry condition occurred while trying to read the rest.
16420 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16421 NID_pkcs7_encrypted by default: this was wrong since this should almost
16422 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16423 the encrypted data type: this is a more sensible place to put it and it
16424 allows the PKCS#12 code to be tidied up that duplicated this
16429 * Changed obj_dat.pl script so it takes its input and output files on
16430 the command line. This should avoid shell escape redirection problems
16435 * Initial support for certificate extension requests, these are included
16436 in things like Xenroll certificate requests. Included functions to allow
16437 extensions to be obtained and added.
16441 * -crlf option to s_client and s_server for sending newlines as
16442 CRLF (as required by many protocols).
16446 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16448 * Install libRSAglue.a when OpenSSL is built with RSAref.
16450 *Ralf S. Engelschall*
16452 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16454 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16456 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16461 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16462 DH parameters/keys (q is lost during that conversion, but the resulting
16463 DH parameters contain its length).
16465 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16466 much faster than DH_generate_parameters (which creates parameters
16467 where `p = 2*q + 1`), and also the smaller q makes DH computations
16468 much more efficient (160-bit exponentiation instead of 1024-bit
16469 exponentiation); so this provides a convenient way to support DHE
16470 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16471 utter importance to use
16472 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16474 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16475 when such DH parameters are used, because otherwise small subgroup
16476 attacks may become possible!
16480 * Avoid memory leak in i2d_DHparams.
16484 * Allow the -k option to be used more than once in the enc program:
16485 this allows the same encrypted message to be read by multiple recipients.
16489 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16490 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16491 it will always use the numerical form of the OID, even if it has a short
16496 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16497 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16498 otherwise bn_mod_exp was called. In the case of hardware keys for example
16499 no private key components need be present and it might store extra data
16500 in the RSA structure, which cannot be accessed from bn_mod_exp.
16501 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16502 private key operations.
16506 * Added support for SPARC Linux.
16510 * pem_password_cb function type incompatibly changed from
16511 typedef int pem_password_cb(char *buf, int size, int rwflag);
16513 ....(char *buf, int size, int rwflag, void *userdata);
16514 so that applications can pass data to their callbacks:
16515 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16516 additional void * argument, which is just handed through whenever
16517 the password callback is called.
16519 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16521 New function SSL_CTX_set_default_passwd_cb_userdata.
16523 Compatibility note: As many C implementations push function arguments
16524 onto the stack in reverse order, the new library version is likely to
16525 interoperate with programs that have been compiled with the old
16526 pem_password_cb definition (PEM_whatever takes some data that
16527 happens to be on the stack as its last argument, and the callback
16528 just ignores this garbage); but there is no guarantee whatsoever that
16531 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16532 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16533 problems not only on Windows, but also on some Unix platforms.
16534 To avoid problematic command lines, these definitions are now in an
16535 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16536 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16540 * MIPS III/IV assembler module is reimplemented.
16544 * More DES library cleanups: remove references to srand/rand and
16545 delete an unused file.
16549 * Add support for the free Netwide assembler (NASM) under Win32,
16550 since not many people have MASM (ml) and it can be hard to obtain.
16551 This is currently experimental but it seems to work OK and pass all
16552 the tests. Check out INSTALL.W32 for info.
16556 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16557 without temporary keys kept an extra copy of the server key,
16558 and connections with temporary keys did not free everything in case
16563 * New function RSA_check_key and new openssl rsa option -check
16564 for verifying the consistency of RSA keys.
16566 *Ulf Moeller, Bodo Moeller*
16568 * Various changes to make Win32 compile work:
16569 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16570 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16571 comparison" warnings.
16572 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16576 * Add a debugging option to PKCS#5 v2 key generation function: when
16577 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16578 derived keys are printed to stderr.
16582 * Copy the flags in ASN1_STRING_dup().
16584 *Roman E. Pavlov <pre@mo.msk.ru>*
16586 * The x509 application mishandled signing requests containing DSA
16587 keys when the signing key was also DSA and the parameters didn't match.
16589 It was supposed to omit the parameters when they matched the signing key:
16590 the verifying software was then supposed to automatically use the CA's
16591 parameters if they were absent from the end user certificate.
16593 Omitting parameters is no longer recommended. The test was also
16594 the wrong way round! This was probably due to unusual behaviour in
16595 EVP_cmp_parameters() which returns 1 if the parameters match.
16596 This meant that parameters were omitted when they *didn't* match and
16597 the certificate was useless. Certificates signed with 'ca' didn't have
16600 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16602 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16603 The interface is as follows:
16604 Applications can use
16605 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16606 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16607 "off" is now the default.
16608 The library internally uses
16609 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16610 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16611 to disable memory-checking temporarily.
16613 Some inconsistent states that previously were possible (and were
16614 even the default) are now avoided.
16616 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16617 with each memory chunk allocated; this is occasionally more helpful
16618 than just having a counter.
16620 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16622 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16627 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16628 which largely parallels "options", but is for changing API behaviour,
16629 whereas "options" are about protocol behaviour.
16630 Initial "mode" flags are:
16632 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16633 a single record has been written.
16634 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16635 retries use the same buffer location.
16636 (But all of the contents must be
16641 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16644 * Fix problems with no-hmac etc.
16646 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16648 * New functions RSA_get_default_method(), RSA_set_method() and
16649 RSA_get_method(). These allows replacement of RSA_METHODs without having
16650 to mess around with the internals of an RSA structure.
16654 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16655 Also really enable memory leak checks in openssl.c and in some
16658 *Chad C. Mulligan, Bodo Moeller*
16660 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16661 up the length of negative integers. This has now been simplified to just
16662 store the length when it is first determined and use it later, rather
16663 than trying to keep track of where data is copied and updating it to
16665 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16667 * Add a new function PKCS7_signatureVerify. This allows the verification
16668 of a PKCS#7 signature but with the signing certificate passed to the
16669 function itself. This contrasts with PKCS7_dataVerify which assumes the
16670 certificate is present in the PKCS#7 structure. This isn't always the
16671 case: certificates can be omitted from a PKCS#7 structure and be
16672 distributed by "out of band" means (such as a certificate database).
16676 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16677 function prototypes in pem.h, also change util/mkdef.pl to add the
16678 necessary function names.
16682 * mk1mf.pl (used by Windows builds) did not properly read the
16683 options set by Configure in the top level Makefile, and Configure
16684 was not even able to write more than one option correctly.
16685 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16689 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16690 file to be loaded from a BIO or FILE pointer. The BIO version will
16691 for example allow memory BIOs to contain config info.
16695 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16696 Whoever hopes to achieve shared-library compatibility across versions
16697 must use this, not the compile-time macro.
16698 (Exercise 0.9.4: Which is the minimum library version required by
16700 Note: All this applies only to multi-threaded programs, others don't
16705 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16706 through a BIO pair triggered the default case, i.e.
16707 SSLerr(...,SSL_R_UNKNOWN_STATE).
16711 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16712 can use the SSL library even if none of the specific BIOs is
16717 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16718 for the encoded length.
16720 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16722 * Add initial documentation of the X509V3 functions.
16726 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16727 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16728 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16729 secure PKCS#8 private key format with a high iteration count.
16733 * Fix determination of Perl interpreter: A perl or perl5
16734 *directory* in $PATH was also accepted as the interpreter.
16736 *Ralf S. Engelschall*
16738 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16739 wrong with it but it was very old and did things like calling
16740 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16741 unusual formatting.
16745 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16746 to use the new extension code.
16750 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16751 with macros. This should make it easier to change their form, add extra
16752 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16757 * Add to configuration table a new entry that can specify an alternative
16758 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16759 according to Mark Crispin <MRC@Panda.COM>.
16763 * DES CBC did not update the IV. Weird.
16767 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16768 Changing the behaviour of the former might break existing programs --
16769 where IV updating is needed, des_ncbc_encrypt can be used.
16772 * When bntest is run from "make test" it drives bc to check its
16773 calculations, as well as internally checking them. If an internal check
16774 fails, it needs to cause bc to give a non-zero result or make test carries
16775 on without noticing the failure. Fixed.
16779 * DES library cleanups.
16783 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16784 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16785 ciphers. NOTE: although the key derivation function has been verified
16786 against some published test vectors it has not been extensively tested
16787 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16792 * Instead of "mkdir -p", which is not fully portable, use new
16793 Perl script "util/mkdir-p.pl".
16797 * Rewrite the way password based encryption (PBE) is handled. It used to
16798 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16799 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16800 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16801 the 'parameter' field of the AlgorithmIdentifier is passed to the
16802 underlying key generation function so it must do its own ASN1 parsing.
16803 This has also changed the EVP_PBE_CipherInit() function which now has a
16804 'parameter' argument instead of literal salt and iteration count values
16805 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16809 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16810 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16811 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16812 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16813 value was just used as a "magic string" and not used directly its
16814 value doesn't matter.
16818 * Introduce some semblance of const correctness to BN. Shame C doesn't
16823 * "linux-sparc64" configuration (ultrapenguin).
16825 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16826 "linux-sparc" configuration.
16828 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16830 * config now generates no-xxx options for missing ciphers.
16834 * Support the EBCDIC character set (work in progress).
16835 File ebcdic.c not yet included because it has a different license.
16837 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16839 * Support BS2000/OSD-POSIX.
16841 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16843 * Make callbacks for key generation use `void *` instead of `char *`.
16847 * Make S/MIME samples compile (not yet tested).
16851 * Additional typesafe stacks.
16855 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16859 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16861 * New configuration variant "sco5-gcc".
16863 * Updated some demos.
16865 *Sean O Riordain, Wade Scholine*
16867 * Add missing BIO_free at exit of pkcs12 application.
16871 * Fix memory leak in conf.c.
16875 * Updates for Win32 to assembler version of MD5.
16879 * Set #! path to perl in apps/der_chop to where we found it
16880 instead of using a fixed path.
16884 * SHA library changes for irix64-mips4-cc.
16888 * Improvements for VMS support.
16892 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16894 * Bignum library bug fix. IRIX 6 passes "make test" now!
16895 This also avoids the problems with SC4.2 and unpatched SC5.
16897 *Andy Polyakov <appro@fy.chalmers.se>*
16899 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16900 These are required because of the typesafe stack would otherwise break
16901 existing code. If old code used a structure member which used to be STACK
16902 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16903 sk_num or sk_value it would produce an error because the num, data members
16904 are not present in STACK_OF. Now it just produces a warning. sk_set
16905 replaces the old method of assigning a value to sk_value
16906 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16907 that does this will no longer work (and should use sk_set instead) but
16908 this could be regarded as a "questionable" behaviour anyway.
16912 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16913 correctly handle encrypted S/MIME data.
16917 * Change type of various DES function arguments from des_cblock
16918 (which means, in function argument declarations, pointer to char)
16919 to des_cblock * (meaning pointer to array with 8 char elements),
16920 which allows the compiler to do more typechecking; it was like
16921 that back in SSLeay, but with lots of ugly casts.
16923 Introduce new type const_des_cblock.
16927 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16928 problems: find RecipientInfo structure that matches recipient certificate
16929 and initialise the ASN1 structures properly based on passed cipher.
16933 * Belatedly make the BN tests actually check the results.
16937 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16938 to and from BNs: it was completely broken. New compilation option
16939 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16940 key elements as negative integers.
16944 * Reorganize and speed up MD5.
16946 *Andy Polyakov <appro@fy.chalmers.se>*
16950 *Richard Levitte <richard@levitte.org>*
16952 * New option -out to asn1parse to allow the parsed structure to be
16953 output to a file. This is most useful when combined with the -strparse
16954 option to examine the output of things like OCTET STRINGS.
16958 * Make SSL library a little more fool-proof by not requiring any longer
16959 that `SSL_set_{accept,connect}_state` be called before
16960 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16961 in many applications because usually everything *appeared* to work as
16962 intended anyway -- now it really works as intended).
16966 * Move openssl.cnf out of lib/.
16970 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16971 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16972 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16974 *Ralf S. Engelschall*
16976 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16977 handle PKCS#7 enveloped data properly.
16979 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16981 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16982 copying pointers. The cert_st handling is changed by this in
16983 various ways (and thus what used to be known as ctx->default_cert
16984 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16985 any longer when s->cert does not give us what we need).
16986 ssl_cert_instantiate becomes obsolete by this change.
16987 As soon as we've got the new code right (possibly it already is?),
16988 we have solved a couple of bugs of the earlier code where s->cert
16989 was used as if it could not have been shared with other SSL structures.
16991 Note that using the SSL API in certain dirty ways now will result
16992 in different behaviour than observed with earlier library versions:
16993 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16994 does not influence s as it used to.
16996 In order to clean up things more thoroughly, inside SSL_SESSION
16997 we don't use CERT any longer, but a new structure SESS_CERT
16998 that holds per-session data (if available); currently, this is
16999 the peer's certificate chain and, for clients, the server's certificate
17000 and temporary key. CERT holds only those values that can have
17001 meaningful defaults in an SSL_CTX.
17005 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17006 from the internal representation. Various PKCS#7 fixes: remove some
17007 evil casts and set the enc_dig_alg field properly based on the signing
17012 * Allow PKCS#12 password to be set from the command line or the
17013 environment. Let 'ca' get its config file name from the environment
17014 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17019 * Allow certificate policies extension to use an IA5STRING for the
17020 organization field. This is contrary to the PKIX definition but
17021 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17026 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17027 without disallowing inline assembler and the like for non-pedantic builds.
17031 * Support Borland C++ builder.
17033 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17039 * SHA-1 cleanups and performance enhancements.
17041 *Andy Polyakov <appro@fy.chalmers.se>*
17043 * Sparc v8plus assembler for the bignum library.
17045 *Andy Polyakov <appro@fy.chalmers.se>*
17047 * Accept any -xxx and +xxx compiler options in Configure.
17051 * Update HPUX configuration.
17055 * Add missing `sk_<type>_unshift()` function to safestack.h
17057 *Ralf S. Engelschall*
17059 * New function SSL_CTX_use_certificate_chain_file that sets the
17060 "extra_cert"s in addition to the certificate. (This makes sense
17061 only for "PEM" format files, as chains as a whole are not
17066 * Support verify_depth from the SSL API.
17067 x509_vfy.c had what can be considered an off-by-one-error:
17068 Its depth (which was not part of the external interface)
17069 was actually counting the number of certificates in a chain;
17070 now it really counts the depth.
17074 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17075 instead of X509err, which often resulted in confusing error
17076 messages since the error codes are not globally unique
17077 (e.g. an alleged error in ssl3_accept when a certificate
17078 didn't match the private key).
17080 * New function SSL_CTX_set_session_id_context that allows to set a default
17081 value (so that you don't need SSL_set_session_id_context for each
17082 connection using the SSL_CTX).
17086 * OAEP decoding bug fix.
17090 * Support INSTALL_PREFIX for package builders, as proposed by
17095 * New Configure options "threads" and "no-threads". For systems
17096 where the proper compiler options are known (currently Solaris
17097 and Linux), "threads" is the default.
17101 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17105 * Install various scripts to $(OPENSSLDIR)/misc, not to
17106 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17107 such as /usr/local/bin.
17111 * "make linux-shared" to build shared libraries.
17113 *Niels Poppe <niels@netbox.org>*
17115 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17119 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17120 extension adding in x509 utility.
17124 * Remove NOPROTO sections and error code comments.
17128 * Partial rewrite of the DEF file generator to now parse the ANSI
17133 * New Configure options --prefix=DIR and --openssldir=DIR.
17137 * Complete rewrite of the error code script(s). It is all now handled
17138 by one script at the top level which handles error code gathering,
17139 header rewriting and C source file generation. It should be much better
17140 than the old method: it now uses a modified version of Ulf's parser to
17141 read the ANSI prototypes in all header files (thus the old K&R definitions
17142 aren't needed for error creation any more) and do a better job of
17143 translating function codes into names. The old 'ASN1 error code embedded
17144 in a comment' is no longer necessary and it doesn't use .err files which
17145 have now been deleted. Also the error code call doesn't have to appear all
17146 on one line (which resulted in some large lines...).
17150 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17154 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17155 0 (which usually indicates a closed connection), but continue reading.
17159 * Fix some race conditions.
17163 * Add support for CRL distribution points extension. Add Certificate
17164 Policies and CRL distribution points documentation.
17168 * Move the autogenerated header file parts to crypto/opensslconf.h.
17172 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17173 8 of keying material. Merlin has also confirmed interop with this fix
17174 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17176 *Merlin Hughes <merlin@baltimore.ie>*
17178 * Fix lots of warnings.
17180 *Richard Levitte <levitte@stacken.kth.se>*
17182 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17183 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17185 *Richard Levitte <levitte@stacken.kth.se>*
17187 * Fix problems with sizeof(long) == 8.
17189 *Andy Polyakov <appro@fy.chalmers.se>*
17191 * Change functions to ANSI C.
17195 * Fix typos in error codes.
17197 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17199 * Remove defunct assembler files from Configure.
17203 * SPARC v8 assembler BIGNUM implementation.
17205 *Andy Polyakov <appro@fy.chalmers.se>*
17207 * Support for Certificate Policies extension: both print and set.
17208 Various additions to support the r2i method this uses.
17212 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17213 return a const string when you are expecting an allocated buffer.
17217 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17218 types DirectoryString and DisplayText.
17222 * Add code to allow r2i extensions to access the configuration database,
17223 add an LHASH database driver and add several ctx helper functions.
17227 * Fix an evil bug in bn_expand2() which caused various BN functions to
17228 fail when they extended the size of a BIGNUM.
17232 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17233 support typesafe stack.
17237 * Fix typo in SSL_[gs]et_options().
17239 *Nils Frostberg <nils@medcom.se>*
17241 * Delete various functions and files that belonged to the (now obsolete)
17242 old X509V3 handling code.
17246 * New Configure option "rsaref".
17250 * Don't auto-generate pem.h.
17254 * Introduce type-safe ASN.1 SETs.
17258 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17260 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17262 * Introduce type-safe STACKs. This will almost certainly break lots of code
17263 that links with OpenSSL (well at least cause lots of warnings), but fear
17264 not: the conversion is trivial, and it eliminates loads of evil casts. A
17265 few STACKed things have been converted already. Feel free to convert more.
17266 In the fullness of time, I'll do away with the STACK type altogether.
17270 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17271 specified in `<certfile>` by updating the entry in the index.txt file.
17272 This way one no longer has to edit the index.txt file manually for
17273 revoking a certificate. The -revoke option does the gory details now.
17275 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17277 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17278 `-text` option at all and this way the `-noout -text` combination was
17279 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17281 *Ralf S. Engelschall*
17283 * Make sure a corresponding plain text error message exists for the
17284 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17285 verify callback function determined that a certificate was revoked.
17287 *Ralf S. Engelschall*
17289 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17290 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17291 all available ciphers including rc5, which was forgotten until now.
17292 In order to let the testing shell script know which algorithms
17293 are available, a new (up to now undocumented) command
17294 `openssl list-cipher-commands` is used.
17298 * Bugfix: s_client occasionally would sleep in select() when
17299 it should have checked SSL_pending() first.
17303 * New functions DSA_do_sign and DSA_do_verify to provide access to
17304 the raw DSA values prior to ASN.1 encoding.
17308 * Tweaks to Configure
17310 *Niels Poppe <niels@netbox.org>*
17312 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17317 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17321 * New config option to avoid instructions that are illegal on the 80386.
17322 The default code is faster, but requires at least a 486.
17326 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17327 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17328 same as SSL2_VERSION anyway.
17332 * New "-showcerts" option for s_client.
17336 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17337 application. Various cleanups and fixes.
17341 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17342 modify error routines to work internally. Add error codes and PBE init
17343 to library startup routines.
17347 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17348 packing functions to asn1 and evp. Changed function names and error
17349 codes along the way.
17353 * PKCS12 integration: and so it begins... First of several patches to
17354 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17355 objects to objects.h
17359 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17360 and display support for Thawte strong extranet extension.
17364 * Add LinuxPPC support.
17366 *Jeff Dubrule <igor@pobox.org>*
17368 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17369 bn_div_words in alpha.s.
17371 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17373 * Make sure the RSA OAEP test is skipped under -DRSAref because
17374 OAEP isn't supported when OpenSSL is built with RSAref.
17376 *Ulf Moeller <ulf@fitug.de>*
17378 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17379 so they no longer are missing under -DNOPROTO.
17381 *Soren S. Jorvang <soren@t.dk>*
17383 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17385 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17386 doesn't work when the session is reused. Coming soon!
17390 * Fix a security hole, that allows sessions to be reused in the wrong
17391 context thus bypassing client cert protection! All software that uses
17392 client certs and session caches in multiple contexts NEEDS PATCHING to
17393 allow session reuse! A fuller solution is in the works.
17395 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17397 * Some more source tree cleanups (removed obsolete files
17398 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17399 permission on "config" script to be executable) and a fix for the INSTALL
17402 *Ulf Moeller <ulf@fitug.de>*
17404 * Remove some legacy and erroneous uses of malloc, free instead of
17407 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17409 * Make rsa_oaep_test return non-zero on error.
17411 *Ulf Moeller <ulf@fitug.de>*
17413 * Add support for native Solaris shared libraries. Configure
17414 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17415 if someone would make that last step automatic.
17417 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17419 * ctx_size was not built with the right compiler during "make links". Fixed.
17423 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17424 except NULL ciphers". This means the default cipher list will no longer
17425 enable NULL ciphers. They need to be specifically enabled e.g. with
17426 the string "DEFAULT:eNULL".
17430 * Fix to RSA private encryption routines: if p < q then it would
17431 occasionally produce an invalid result. This will only happen with
17432 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17436 * Be less restrictive and allow also `perl util/perlpath.pl
17437 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17438 because this way one can also use an interpreter named `perl5` (which is
17439 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17440 installed as `perl`).
17442 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17444 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17446 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17448 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17449 advapi32.lib to Win32 build and change the pem test comparison
17450 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17451 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17452 and crypto/des/ede_cbcm_enc.c.
17456 * DES quad checksum was broken on big-endian architectures. Fixed.
17460 * Comment out two functions in bio.h that aren't implemented. Fix up the
17461 Win32 test batch file so it (might) work again. The Win32 test batch file
17462 is horrible: I feel ill....
17466 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17467 in e_os.h. Audit of header files to check ANSI and non ANSI
17468 sections: 10 functions were absent from non ANSI section and not exported
17469 from Windows DLLs. Fixed up libeay.num for new functions.
17473 * Make `openssl version` output lines consistent.
17475 *Ralf S. Engelschall*
17477 * Fix Win32 symbol export lists for BIO functions: Added
17478 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17479 to ms/libeay{16,32}.def.
17481 *Ralf S. Engelschall*
17483 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17484 fine under Unix and passes some trivial tests I've now added. But the
17485 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17486 added to make sure no one expects that this stuff really works in the
17487 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17488 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17491 *Ralf S. Engelschall*
17493 * Fix the generation of two part addresses in perl.
17495 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17497 * Add config entry for Linux on MIPS.
17499 *John Tobey <jtobey@channel1.com>*
17501 * Make links whenever Configure is run, unless we are on Windoze.
17505 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17506 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17511 * Add a useful kludge to allow package maintainers to specify compiler and
17512 other platforms details on the command line without having to patch the
17513 Configure script every time: One now can use
17514 `perl Configure <id>:<details>`,
17515 i.e. platform ids are allowed to have details appended
17516 to them (separated by colons). This is treated as there would be a static
17517 pre-configured entry in Configure's %table under key `<id>` with value
17518 `<details>` and `perl Configure <id>` is called. So, when you want to
17519 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17520 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17521 now, which overrides the FreeBSD-elf entry on-the-fly.
17523 *Ralf S. Engelschall*
17525 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17529 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17530 on the `perl Configure ...` command line. This way one can compile
17531 OpenSSL libraries with Position Independent Code (PIC) which is needed
17532 for linking it into DSOs.
17534 *Ralf S. Engelschall*
17536 * Remarkably, export ciphers were totally broken and no-one had noticed!
17541 * Cleaned up the LICENSE document: The official contact for any license
17542 questions now is the OpenSSL core team under openssl-core@openssl.org.
17543 And add a paragraph about the dual-license situation to make sure people
17544 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17545 to the OpenSSL toolkit.
17547 *Ralf S. Engelschall*
17549 * General source tree makefile cleanups: Made `making xxx in yyy...`
17550 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17551 Additionally cleaned up the `make links` target: Remove unnecessary
17552 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17553 to speed processing and no longer clutter the display with confusing
17554 stuff. Instead only the actually done links are displayed.
17556 *Ralf S. Engelschall*
17558 * Permit null encryption ciphersuites, used for authentication only. It used
17559 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17560 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17565 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17566 signed attributes when verifying signatures (this would break them),
17567 the detached data encoding was wrong and public keys obtained using
17568 X509_get_pubkey() weren't freed.
17572 * Add text documentation for the BUFFER functions. Also added a work around
17573 to a Win95 console bug. This was triggered by the password read stuff: the
17574 last character typed gets carried over to the next fread(). If you were
17575 generating a new cert request using 'req' for example then the last
17576 character of the passphrase would be CR which would then enter the first
17581 * Added the new 'Includes OpenSSL Cryptography Software' button as
17582 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17583 button and can be used by applications based on OpenSSL to show the
17584 relationship to the OpenSSL project.
17586 *Ralf S. Engelschall*
17588 * Remove confusing variables in function signatures in files
17589 ssl/ssl_lib.c and ssl/ssl.h.
17591 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17593 * Don't install bss_file.c under PREFIX/include/
17595 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17597 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17598 functions that return function pointers and has support for NT specific
17599 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17600 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17601 unsigned to signed types: this was killing the Win32 compile.
17605 * Add new certificate file to stack functions,
17606 SSL_add_dir_cert_subjects_to_stack() and
17607 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17608 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17609 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17610 This means that Apache-SSL and similar packages don't have to mess around
17611 to add as many CAs as they want to the preferred list.
17615 * Experiment with doxygen documentation. Currently only partially applied to
17617 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17618 openssl.doxy as the configuration file.
17622 * Get rid of remaining C++-style comments which strict C compilers hate.
17624 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17626 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17627 compiled in by default: it has problems with large keys.
17631 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17632 DH private keys and/or callback functions which directly correspond to
17633 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17634 is needed for applications which have to configure certificates on a
17635 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17637 For the RSA certificate situation is makes no difference, but
17638 for the DSA certificate situation this fixes the "no shared cipher"
17639 problem where the OpenSSL cipher selection procedure failed because the
17640 temporary keys were not overtaken from the context and the API provided
17641 no way to reconfigure them.
17642 The new functions now let applications reconfigure the stuff and they
17643 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17644 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17645 non-public-API function ssl_cert_instantiate() is used as a helper
17646 function and also to reduce code redundancy inside ssl_rsa.c.
17648 *Ralf S. Engelschall*
17650 * Move s_server -dcert and -dkey options out of the undocumented feature
17651 area because they are useful for the DSA situation and should be
17652 recognized by the users.
17654 *Ralf S. Engelschall*
17656 * Fix the cipher decision scheme for export ciphers: the export bits are
17657 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17658 SSL_EXP_MASK. So, the original variable has to be used instead of the
17659 already masked variable.
17661 *Richard Levitte <levitte@stacken.kth.se>*
17663 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17665 *Richard Levitte <levitte@stacken.kth.se>*
17667 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17668 from `int` to `unsigned int` because it is a length and initialized by
17669 EVP_DigestFinal() which expects an `unsigned int *`.
17671 *Richard Levitte <levitte@stacken.kth.se>*
17673 * Don't hard-code path to Perl interpreter on shebang line of Configure
17674 script. Instead use the usual Shell->Perl transition trick.
17676 *Ralf S. Engelschall*
17678 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17679 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17680 -noout -modulus` as it's already the case for `openssl rsa -noout
17681 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17682 currently the public key is printed (a decision which was already done by
17683 `openssl dsa -modulus` in the past) which serves a similar purpose.
17684 Additionally the NO_RSA no longer completely removes the whole -modulus
17685 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17688 *Ralf S. Engelschall*
17690 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17691 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17693 *Arne Ansper <arne@ats.cyber.ee>*
17695 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17696 to be added. Now both 'req' and 'ca' can use new objects defined in the
17701 * Add cool BIO that does syslog (or event log on NT).
17703 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17705 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17706 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17707 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17708 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17712 * Add preliminary config info for new extension code.
17716 * Make RSA_NO_PADDING really use no padding.
17718 *Ulf Moeller <ulf@fitug.de>*
17720 * Generate errors when private/public key check is done.
17724 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17725 for some CRL extensions and new objects added.
17729 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17730 key usage extension and fuller support for authority key id.
17734 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17735 padding method for RSA, which is recommended for new applications in PKCS
17736 #1 v2.0 (RFC 2437, October 1998).
17737 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17738 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17739 against Bleichbacher's attack on RSA.
17740 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17743 * Updates to the new SSL compression code
17745 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17747 * Fix so that the version number in the master secret, when passed
17748 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17749 (because the server will not accept higher), that the version number
17750 is 0x03,0x01, not 0x03,0x00
17752 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17754 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17755 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17756 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17760 * Support for RAW extensions where an arbitrary extension can be
17761 created by including its DER encoding. See apps/openssl.cnf for
17766 * Make sure latest Perl versions don't interpret some generated C array
17767 code as Perl array code in the crypto/err/err_genc.pl script.
17769 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17771 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17772 not many people have the assembler. Various Win32 compilation fixes and
17773 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17774 build instructions.
17778 * Modify configure script 'Configure' to automatically create crypto/date.h
17779 file under Win32 and also build pem.h from pem.org. New script
17780 util/mkfiles.pl to create the MINFO file on environments that can't do a
17781 'make files': perl util/mkfiles.pl >MINFO should work.
17785 * Major rework of DES function declarations, in the pursuit of correctness
17786 and purity. As a result, many evil casts evaporated, and some weirdness,
17787 too. You may find this causes warnings in your code. Zapping your evil
17788 casts will probably fix them. Mostly.
17792 * Fix for a typo in asn1.h. Bug fix to object creation script
17793 obj_dat.pl. It considered a zero in an object definition to mean
17794 "end of object": none of the objects in objects.h have any zeros
17795 so it wasn't spotted.
17797 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17799 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17800 Masking (CBCM). In the absence of test vectors, the best I have been able
17801 to do is check that the decrypt undoes the encrypt, so far. Send me test
17802 vectors if you have them.
17806 * Correct calculation of key length for export ciphers (too much space was
17807 allocated for null ciphers). This has not been tested!
17811 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17812 message is now correct (it understands "crypto" and "ssl" on its
17813 command line). There is also now an "update" option. This will update
17814 the util/ssleay.num and util/libeay.num files with any new functions.
17816 perl util/mkdef.pl crypto ssl update
17817 it will update them.
17821 * Overhauled the Perl interface:
17822 - ported BN stuff to OpenSSL's different BN library
17823 - made the perl/ source tree CVS-aware
17824 - renamed the package from SSLeay to OpenSSL (the files still contain
17825 their history because I've copied them in the repository)
17826 - removed obsolete files (the test scripts will be replaced
17827 by better Test::Harness variants in the future)
17829 *Ralf S. Engelschall*
17831 * First cut for a very conservative source tree cleanup:
17832 1. merge various obsolete readme texts into doc/ssleay.txt
17833 where we collect the old documents and readme texts.
17834 2. remove the first part of files where I'm already sure that we no
17835 longer need them because of three reasons: either they are just temporary
17836 files which were left by Eric or they are preserved original files where
17837 I've verified that the diff is also available in the CVS via "cvs diff
17838 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17839 the crypto/md/ stuff).
17841 *Ralf S. Engelschall*
17843 * More extension code. Incomplete support for subject and issuer alt
17844 name, issuer and authority key id. Change the i2v function parameters
17845 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17846 what that's for :-) Fix to ASN1 macro which messed up
17847 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17851 * Preliminary support for ENUMERATED type. This is largely copied from the
17856 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17858 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17860 * Make sure `make rehash` target really finds the `openssl` program.
17862 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17864 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17865 like to hear about it if this slows down other processors.
17869 * Add CygWin32 platform information to Configure script.
17871 *Alan Batie <batie@aahz.jf.intel.com>*
17873 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17875 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17877 * New program nseq to manipulate netscape certificate sequences
17881 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17886 * Fixes to BN code. Previously the default was to define BN_RECURSION
17887 but the BN code had some problems that would cause failures when
17888 doing certificate verification and some other functions.
17890 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17892 * Add ASN1 and PEM code to support netscape certificate sequences.
17896 * Add ASN1 and PEM code to support netscape certificate sequences.
17900 * Add several PKIX and private extended key usage OIDs.
17904 * Modify the 'ca' program to handle the new extension code. Modify
17905 openssl.cnf for new extension format, add comments.
17909 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17910 and add a sample to openssl.cnf so req -x509 now adds appropriate
17915 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17916 error code, add initial support to X509_print() and x509 application.
17920 * Takes a deep breath and start adding X509 V3 extension support code. Add
17921 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17922 stuff is currently isolated and isn't even compiled yet.
17926 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17927 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17928 Removed the versions check from X509 routines when loading extensions:
17929 this allows certain broken certificates that don't set the version
17930 properly to be processed.
17934 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17935 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17936 can still be regenerated with "make depend".
17940 * Spelling mistake in C version of CAST-128.
17942 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17944 * Changes to the error generation code. The perl script err-code.pl
17945 now reads in the old error codes and retains the old numbers, only
17946 adding new ones if necessary. It also only changes the .err files if new
17947 codes are added. The makefiles have been modified to only insert errors
17948 when needed (to avoid needlessly modifying header files). This is done
17949 by only inserting errors if the .err file is newer than the auto generated
17950 C file. To rebuild all the error codes from scratch (the old behaviour)
17951 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17952 or delete all the .err files.
17956 * CAST-128 was incorrectly implemented for short keys. The C version has
17957 been fixed, but is untested. The assembler versions are also fixed, but
17958 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17959 to regenerate it if needed.
17960 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17961 Hagino <itojun@kame.net>*
17963 * File was opened incorrectly in randfile.c.
17965 *Ulf Möller <ulf@fitug.de>*
17967 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17968 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17969 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17970 al: it's just almost always a UTCTime. Note this patch adds new error
17971 codes so do a "make errors" if there are problems.
17975 * Correct Linux 1 recognition in config.
17977 *Ulf Möller <ulf@fitug.de>*
17979 * Remove pointless MD5 hash when using DSA keys in ca.
17981 *Anonymous <nobody@replay.com>*
17983 * Generate an error if given an empty string as a cert directory. Also
17984 generate an error if handed NULL (previously returned 0 to indicate an
17985 error, but didn't set one).
17987 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17989 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17993 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17994 parameters. This was causing a warning which killed off the Win32 compile.
17998 * Remove C++ style comments from crypto/bn/bn_local.h.
18000 *Neil Costigan <neil.costigan@celocom.com>*
18002 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18003 based on a text string, looking up short and long names and finally
18004 "dot" format. The "dot" format stuff didn't work. Added new function
18005 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18006 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18007 OID is not part of the table.
18011 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18012 X509_LOOKUP_by_alias().
18016 * Sort openssl functions by name.
18020 * Get the gendsa program working (hopefully) and add it to app list. Remove
18021 encryption from sample DSA keys (in case anyone is interested the password
18026 * Make *all* `*_free` functions accept a NULL pointer.
18028 *Frans Heymans <fheymans@isaserver.be>*
18030 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18033 *Anonymous <nobody@replay.com>*
18035 * s_server should send the CAfile as acceptable CAs, not its own cert.
18037 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18039 * Don't blow it for numeric -newkey arguments to apps/req.
18041 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18043 * Temp key "for export" tests were wrong in s3_srvr.c.
18045 *Anonymous <nobody@replay.com>*
18047 * Add prototype for temp key callback functions
18048 SSL_CTX_set_tmp_{rsa,dh}_callback().
18052 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18053 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18057 * X509_name_add_entry() freed the wrong thing after an error.
18059 *Arne Ansper <arne@ats.cyber.ee>*
18061 * rsa_eay.c would attempt to free a NULL context.
18063 *Arne Ansper <arne@ats.cyber.ee>*
18065 * BIO_s_socket() had a broken should_retry() on Windoze.
18067 *Arne Ansper <arne@ats.cyber.ee>*
18069 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18071 *Arne Ansper <arne@ats.cyber.ee>*
18073 * Make sure the already existing X509_STORE->depth variable is initialized
18074 in X509_STORE_new(), but document the fact that this variable is still
18075 unused in the certificate verification process.
18077 *Ralf S. Engelschall*
18079 * Fix the various library and apps files to free up pkeys obtained from
18080 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18084 * Fix reference counting in X509_PUBKEY_get(). This makes
18085 demos/maurice/example2.c work, amongst others, probably.
18087 *Steve Henson and Ben Laurie*
18089 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18090 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18091 are no longer created. This way we have a single and consistent command
18092 line interface `openssl <command>`, similar to `cvs <command>`.
18094 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18096 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18097 BIT STRING wrapper always have zero unused bits.
18101 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18105 * Make the top-level INSTALL documentation easier to understand.
18109 * Makefiles updated to exit if an error occurs in a sub-directory
18110 make (including if user presses ^C) [Paul Sutton]
18112 * Make Montgomery context stuff explicit in RSA data structure.
18116 * Fix build order of pem and err to allow for generated pem.h.
18120 * Fix renumbering bug in X509_NAME_delete_entry().
18124 * Enhanced the err-ins.pl script so it makes the error library number
18125 global and can add a library name. This is needed for external ASN1 and
18126 other error libraries.
18130 * Fixed sk_insert which never worked properly.
18134 * Fix ASN1 macros so they can handle indefinite length constructed
18135 EXPLICIT tags. Some non standard certificates use these: they can now
18140 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18141 into a single doc/ssleay.txt bundle. This way the information is still
18142 preserved but no longer messes up this directory. Now it's new room for
18143 the new set of documentation files.
18145 *Ralf S. Engelschall*
18147 * SETs were incorrectly DER encoded. This was a major pain, because they
18148 shared code with SEQUENCEs, which aren't coded the same. This means that
18149 almost everything to do with SETs or SEQUENCEs has either changed name or
18150 number of arguments.
18152 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18154 * Fix test data to work with the above.
18158 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18159 was already fixed by Eric for 0.9.1 it seems.
18161 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18163 * Autodetect FreeBSD3.
18167 * Fix various bugs in Configure. This affects the following platforms:
18171 unixware-2.0-pentium
18176 * Eliminate generated files from CVS. Reorder tests to regenerate files
18177 before they are needed.
18181 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18185 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18187 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18188 changed SSLeay to OpenSSL in version strings.
18190 *Ralf S. Engelschall*
18192 * Some fixups to the top-level documents.
18196 * Fixed the nasty bug where rsaref.h was not found under compile-time
18197 because the symlink to include/ was missing.
18199 *Ralf S. Engelschall*
18201 * Incorporated the popular no-RSA/DSA-only patches
18202 which allow to compile a RSA-free SSLeay.
18204 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18206 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18207 when "ssleay" is still not found.
18209 *Ralf S. Engelschall*
18211 * Added more platforms to Configure: Cray T3E, HPUX 11,
18213 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18215 * Updated the README file.
18217 *Ralf S. Engelschall*
18219 * Added various .cvsignore files in the CVS repository subdirs
18220 to make a "cvs update" really silent.
18222 *Ralf S. Engelschall*
18224 * Recompiled the error-definition header files and added
18225 missing symbols to the Win32 linker tables.
18227 *Ralf S. Engelschall*
18229 * Cleaned up the top-level documents;
18230 o new files: CHANGES and LICENSE
18231 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18232 o merged COPYRIGHT into LICENSE
18233 o removed obsolete TODO file
18234 o renamed MICROSOFT to INSTALL.W32
18236 *Ralf S. Engelschall*
18238 * Removed dummy files from the 0.9.1b source tree:
18239 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18240 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18241 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18242 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18243 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18245 *Ralf S. Engelschall*
18247 * Added various platform portability fixes.
18251 * The Genesis of the OpenSSL rpject:
18252 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18253 Young and Tim J. Hudson created while they were working for C2Net until
18256 *The OpenSSL Project*
18258 ### Changes between 0.9.0b and 0.9.1b [not released]
18260 * Updated a few CA certificates under certs/
18264 * Changed some BIGNUM api stuff.
18268 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18269 DGUX x86, Linux Alpha, etc.
18273 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18274 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18279 * Add -strparse option to asn1pars program which parses nested
18282 *Dr Stephen Henson <shenson@bigfoot.com>*
18284 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18288 * DSA fix for "ca" program.
18292 * Added "-genkey" option to "dsaparam" program.
18296 * Added RIPE MD160 (rmd160) message digest.
18300 * Added -a (all) option to "ssleay version" command.
18304 * Added PLATFORM define which is the id given to Configure.
18308 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18312 * Extended the ASN.1 parser routines.
18316 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18320 * Added a BN_CTX to the BN library.
18324 * Fixed the weak key values in DES library
18328 * Changed API in EVP library for cipher aliases.
18332 * Added support for RC2/64bit cipher.
18336 * Converted the lhash library to the crypto/mem.c functions.
18340 * Added more recognized ASN.1 object ids.
18344 * Added more RSA padding checks for SSL/TLS.
18348 * Added BIO proxy/filter functionality.
18352 * Added extra_certs to SSL_CTX which can be used
18353 send extra CA certificates to the client in the CA cert chain sending
18354 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18358 * Now Fortezza is denied in the authentication phase because
18359 this is key exchange mechanism is not supported by SSLeay at all.
18363 * Additional PKCS1 checks.
18367 * Support the string "TLSv1" for all TLS v1 ciphers.
18371 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18372 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18376 * Fixed a few memory leaks.
18380 * Fixed various code and comment typos.
18384 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18385 bytes sent in the client random.
18387 *Edward Bishop <ebishop@spyglass.com>*
18391 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18392 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18393 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18394 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18395 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18396 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18397 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18398 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18399 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18400 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18401 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18402 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18403 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18404 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18405 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18406 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18407 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18408 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18409 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18410 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18411 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18412 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18413 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18414 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18415 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18416 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18417 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18418 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18419 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18420 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18421 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18422 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18423 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18424 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18425 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18426 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18427 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18428 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18429 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18430 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18431 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18432 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18433 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18434 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18435 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18436 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18437 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18438 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18439 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18440 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18441 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18442 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18443 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18444 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18445 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18446 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18447 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18448 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18449 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18450 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18451 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18452 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18453 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18454 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18455 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18456 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18457 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18458 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18459 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18460 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18461 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18462 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18463 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18464 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18465 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18466 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18467 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18468 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18469 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18470 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18471 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18472 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18473 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18474 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18475 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18476 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18477 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18478 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18479 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18480 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18481 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18482 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18483 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18484 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18485 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18486 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18487 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18488 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18489 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18490 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18491 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18492 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18493 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18494 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18495 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18496 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18497 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18498 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18499 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18500 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18501 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18502 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18503 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18504 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18505 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18506 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18507 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18508 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18509 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18510 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18511 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18512 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18513 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18514 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18515 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18516 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18517 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18518 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18519 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18520 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18521 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18522 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18523 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18524 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18525 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18526 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18527 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18528 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18529 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18530 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18531 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18532 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18533 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18534 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18535 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18536 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18537 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18538 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18539 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18540 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18541 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18542 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18543 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18544 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18545 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18546 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18547 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18548 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18549 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18550 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655