4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * Added a library context that applications as well as other
27 libraries can use to form a separate context within which libcrypto
28 operations are performed.
30 There are two ways this can be used:
32 - Directly, by passing a library context to functions that take
33 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
35 - Indirectly, by creating a new library context and then assigning
36 it as the new default, with `OPENSSL_CTX_set0_default`.
38 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
39 apart from the functions directly related to `OPENSSL_CTX`, accept
40 NULL to indicate that the default library context should be used.
42 Library code that changes the default library context using
43 `OPENSSL_CTX_set0_default` should take care to restore it with a
44 second call before returning to the caller.
48 * Handshake now fails if Extended Master Secret extension is dropped
53 * Dropped interactive mode from the 'openssl' program. From now on,
54 the `openssl` command without arguments is equivalent to `openssl
59 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
60 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
61 While the old function names have been retained for backward compatibility
62 they should not be used in new developments
63 because their return values are confusing: Unlike other `_cmp()` functions
64 they do not return 0 in case their arguments are equal.
68 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
69 EC_GROUP_get_field_type().
73 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
74 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
75 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
76 Applications should rely on the library automatically assigning a suitable
77 EC_METHOD internally upon EC_GROUP construction.
81 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
82 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
83 assigned internally without application intervention.
84 Users of EC_GROUP_new() should switch to a different suitable constructor.
88 * Add CAdES-BES signature verification support, mostly derived
89 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
91 *Filipe Raimundo da Silva*
93 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
97 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
98 functions are not widely used and now OpenSSL automatically perform this
99 conversion when needed.
103 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
104 EC_KEY_precompute_mult(). These functions are not widely used and
105 applications should instead switch to named curves which OpenSSL has
106 hardcoded lookup tables for.
110 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
111 should instead use the L<EC_POINT_mul(3)> function.
115 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
116 that are not applicable to the new provider model. Applications should
117 instead use EVP_default_properties_is_fips_enabled() and
118 EVP_default_properties_enable_fips().
122 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
123 is set, an unexpected EOF is ignored, it pretends a close notify was received
124 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
128 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
129 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
130 used and applications should instead use the
131 L<EC_POINT_set_affine_coordinates(3)> and
132 L<EC_POINT_get_affine_coordinates(3)> functions.
136 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
137 arrays to be more easily constructed via a series of utility functions.
138 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
139 the various push functions and finally convert to a passable OSSL_PARAM
140 array using OSSL_PARAM_BLD_to_param().
144 * The security strength of SHA1 and MD5 based signatures in TLS has been
145 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
146 working at the default security level of 1 and instead requires security
147 level 0. The security level can be changed either using the cipher string
148 with @SECLEVEL, or calling SSL_CTX_set_security_level().
152 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
153 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
154 internal keys, if they correspond to one of those built in types.
158 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
159 contain a provider side internal key.
163 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
164 They are old functions that we don't use, and that you could disable with
165 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
169 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
170 have been converted to Markdown with the goal to produce documents
171 which not only look pretty when viewed online in the browser, but
172 remain well readable inside a plain text editor.
174 To achieve this goal, a 'minimalistic' Markdown style has been applied
175 which avoids formatting elements that interfere too much with the
176 reading flow in the text file. For example, it
178 * avoids [ATX headings][] and uses [setext headings][] instead
179 (which works for `<h1>` and `<h2>` headings only).
180 * avoids [inline links][] and uses [reference links][] instead.
181 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
183 [ATX headings]: https://github.github.com/gfm/#atx-headings
184 [setext headings]: https://github.github.com/gfm/#setext-headings
185 [inline links]: https://github.github.com/gfm/#inline-link
186 [reference links]: https://github.github.com/gfm/#reference-link
187 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
188 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
190 *Matthias St. Pierre*
192 * The test suite is changed to preserve results of each test recipe.
193 A new directory test-runs/ with subdirectories named like the
194 test recipes are created in the build tree for this purpose.
198 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
199 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
200 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
202 *David von Oheimb, Martin Peylo*
204 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
205 The legacy OCSP-focused and only partly documented API is retained.
206 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
210 * BIO_do_connect and BIO_do_handshake have been extended:
211 If domain name resolution yields multiple IP addresses all of them are tried
212 after connect() failures.
216 * All of the low level RSA functions have been deprecated including:
218 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
219 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
220 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
221 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
222 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
223 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
224 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
225 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
226 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
227 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
228 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
229 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
230 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
231 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
232 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
233 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
234 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
235 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
236 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
237 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
238 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
239 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
240 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
241 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
242 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
243 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
244 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
245 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
247 Use of these low level functions has been informally discouraged for a long
248 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
249 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
250 L<EVP_PKEY_decrypt(3)>.
254 * X509 certificates signed using SHA1 are no longer allowed at security
256 In TLS/SSL the default security level is 1. It can be set either
257 using the cipher string with @SECLEVEL, or calling
258 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
259 a call to SSL_CTX_use_certificate() will fail if the security level is not
261 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
262 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
267 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
268 modified to use PKEY APIs. These commands are now in maintenance mode
269 and no new features will be added to them.
273 * The command line utility rsautl has been deprecated.
274 Instead use the pkeyutl program.
278 * The command line utilities genrsa and rsa have been modified to use PKEY
279 APIs These commands are now in maintenance mode and no new features will
284 * All of the low level DH functions have been deprecated including:
286 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
287 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
288 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
289 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
290 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
291 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
292 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
293 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
294 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
295 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
296 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
297 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
298 DH_meth_get_generate_params and DH_meth_set_generate_params.
300 Use of these low level functions has been informally discouraged for a long
301 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
302 and L<EVP_PKEY_derive(3)>.
306 * All of the low level DSA functions have been deprecated including:
308 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
309 DSA_get_default_method, DSA_set_method, DSA_get_method,
310 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
311 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
312 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
313 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
314 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
315 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
316 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
317 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
318 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
319 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
320 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
323 Use of these low level functions has been informally discouraged for a long
324 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
325 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
329 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
330 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
331 This means that applications don't have to look at the curve NID and
332 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
333 However, they still can, that EVP_PKEY_set_alias_type() call acts as
334 a no-op when the EVP_PKEY is already of the given type.
336 Parameter and key generation is also reworked to make it possible
337 to generate EVP_PKEY_SM2 parameters and keys without having to go
338 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
339 However, code that does the latter will still work as before.
343 * Deprecated low level ECDH and ECDSA functions. These include:
345 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
346 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
349 Use of these low level functions has been informally discouraged for a long
350 time. Instead applications should use the EVP_PKEY_derive(3),
351 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
355 * Deprecated the EC_KEY_METHOD functions. These include:
357 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
358 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
359 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
360 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
361 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
362 EC_KEY_METHOD_get_verify.
364 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
368 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
369 and EVP_PKEY_decrypt() instead.
370 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
371 and EVP_PKEY_encrypt() instead.
375 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
376 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
377 a new formulation to include all the things it can be used for,
378 as well as words of caution.
382 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
383 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
387 * All of the low level HMAC functions have been deprecated including:
389 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
390 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
393 Use of these low level functions has been informally discouraged for a long
394 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
395 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
396 and L<EVP_MAC_final(3)>.
400 * Over two thousand fixes were made to the documentation, including:
401 - Common options (such as -rand/-writerand, TLS version control, etc)
402 were refactored and point to newly-enhanced descriptions in openssl.pod.
403 - Added style conformance for all options (with help from Richard Levitte),
404 documented all reported missing options, added a CI build to check
405 that all options are documented and that no unimplemented options
407 - Documented some internals, such as all use of environment variables.
408 - Addressed all internal broken L<> references.
412 * All of the low level CMAC functions have been deprecated including:
414 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
415 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
417 Use of these low level functions has been informally discouraged for a long
418 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
419 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
420 and L<EVP_MAC_final(3)>.
424 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
425 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
428 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
429 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
430 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
431 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
432 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
433 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
434 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
435 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
436 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
437 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
439 Use of these low level functions has been informally discouraged
440 for a long time. Applications should use the EVP_DigestInit_ex(3),
441 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
445 * Corrected the documentation of the return values from the `EVP_DigestSign*`
446 set of functions. The documentation mentioned negative values for some
447 errors, but this was never the case, so the mention of negative values
450 Code that followed the documentation and thereby check with something
451 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
455 * All of the low level cipher functions have been deprecated including:
457 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
458 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
459 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
460 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
461 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
462 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
463 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
464 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
465 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
466 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
467 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
468 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
469 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
470 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
471 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
472 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
473 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
474 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
475 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
476 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
477 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
478 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
479 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
480 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
481 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
482 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
483 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
484 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
485 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
487 Use of these low level functions has been informally discouraged for
488 a long time. Applications should use the high level EVP APIs, e.g.
489 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
490 equivalently named decrypt functions instead.
492 *Matt Caswell and Paul Dale*
494 * Removed include/openssl/opensslconf.h.in and replaced it with
495 include/openssl/configuration.h.in, which differs in not including
496 <openssl/macros.h>. A short header include/openssl/opensslconf.h
497 was added to include both.
499 This allows internal hacks where one might need to modify the set
500 of configured macros, for example this if deprecated symbols are
501 still supposed to be available internally:
503 #include <openssl/configuration.h>
505 #undef OPENSSL_NO_DEPRECATED
506 #define OPENSSL_SUPPRESS_DEPRECATED
508 #include <openssl/macros.h>
510 This should not be used by applications that use the exported
511 symbols, as that will lead to linking errors.
515 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
516 used in exponentiation with 512-bit moduli. No EC algorithms are
517 affected. Analysis suggests that attacks against 2-prime RSA1024,
518 3-prime RSA1536, and DSA1024 as a result of this defect would be very
519 difficult to perform and are not believed likely. Attacks against DH512
520 are considered just feasible. However, for an attack the target would
521 have to re-use the DH512 private key, which is not recommended anyway.
522 Also applications directly using the low level API BN_mod_exp may be
523 affected if they use BN_FLG_CONSTTIME.
528 * Most memory-debug features have been deprecated, and the functionality
529 replaced with no-ops.
533 * Added documentation for the STACK API. OpenSSL only defines the STACK
534 functions where they are used.
538 * Introduced a new method type and API, OSSL_SERIALIZER, to
539 represent generic serializers. An implementation is expected to
540 be able to serialize an object associated with a given name (such
541 as an algorithm name for an asymmetric key) into forms given by
542 implementation properties.
544 Serializers are primarily used from inside libcrypto, through
545 calls to functions like EVP_PKEY_print_private(),
546 PEM_write_bio_PrivateKey() and similar.
548 Serializers are specified in such a way that they can be made to
549 directly handle the provider side portion of an object, if this
550 provider side part comes from the same provider as the serializer
551 itself, but can also be made to handle objects in parametrized
552 form (as an OSSL_PARAM array of data). This allows a provider to
553 offer generic serializers as a service for any other provider.
557 * Added a .pragma directive to the syntax of configuration files, to
558 allow varying behavior in a supported and predictable manner.
559 Currently added pragma:
563 This allows dollar signs to be a keyword character unless it's
564 followed by a opening brace or parenthesis. This is useful for
565 platforms where dollar signs are commonly used in names, such as
566 volume names and system directory names on VMS.
570 * Added functionality to create an EVP_PKEY from user data. This
571 is effectively the same as creating a RSA, DH or DSA object and
572 then assigning them to an EVP_PKEY, but directly using algorithm
573 agnostic EVP functions. A benefit is that this should be future
574 proof for public key algorithms to come.
578 * Change the interpretation of the '--api' configuration option to
579 mean that this is a desired API compatibility level with no
580 further meaning. The previous interpretation, that this would
581 also mean to remove all deprecated symbols up to and including
582 the given version, no requires that 'no-deprecated' is also used
583 in the configuration.
585 When building applications, the desired API compatibility level
586 can be set with the OPENSSL_API_COMPAT macro like before. For
587 API compatibility version below 3.0, the old style numerical
588 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
589 For version 3.0 and on, the value is expected to be the decimal
590 value calculated from the major and minor version like this:
592 MAJOR * 10000 + MINOR * 100
596 -DOPENSSL_API_COMPAT=30000 For 3.0
597 -DOPENSSL_API_COMPAT=30200 For 3.2
599 To hide declarations that are deprecated up to and including the
600 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
601 given when building the application as well.
605 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
606 access to certificate and CRL stores via URIs and OSSL_STORE
609 This adds the following functions:
611 - X509_LOOKUP_store()
612 - X509_STORE_load_file()
613 - X509_STORE_load_path()
614 - X509_STORE_load_store()
615 - SSL_add_store_cert_subjects_to_stack()
616 - SSL_CTX_set_default_verify_store()
617 - SSL_CTX_load_verify_file()
618 - SSL_CTX_load_verify_dir()
619 - SSL_CTX_load_verify_store()
623 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
624 The presence of this system service is determined at run-time.
628 * Added functionality to create an EVP_PKEY context based on data
629 for methods from providers. This takes an algorithm name and a
630 property query string and simply stores them, with the intent
631 that any operation that uses this context will use those strings
632 to fetch the needed methods implicitly, thereby making the port
633 of application written for pre-3.0 OpenSSL easier.
637 * The undocumented function NCONF_WIN32() has been deprecated; for
638 conversion details see the HISTORY section of doc/man5/config.pod
642 * Introduced the new functions EVP_DigestSignInit_ex() and
643 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
644 EVP_DigestVerifyUpdate() have been converted to functions. See the man
645 pages for further details.
649 * Over two thousand fixes were made to the documentation, including:
650 adding missing command flags, better style conformance, documentation
653 *Rich Salz, Richard Levitte*
655 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
656 X25519, X448, Ed25519 and Ed448.
660 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
665 * Deprecated the public definition of ERR_STATE as well as the function
666 ERR_get_state(). This is done in preparation of making ERR_STATE an
671 * Added ERR functionality to give callers access to the stored function
672 names that have replaced the older function code based functions.
674 New functions are ERR_get_error_func(), ERR_peek_error_func(),
675 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
676 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
677 and ERR_peek_last_error_all().
679 These functions have become deprecated: ERR_get_error_line_data(),
680 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
681 ERR_func_error_string().
685 * Extended testing to be verbose for failing tests only. The make variables
686 VERBOSE_FAILURE or VF can be used to enable this:
688 $ make VF=1 test # Unix
689 $ mms /macro=(VF=1) test ! OpenVMS
690 $ nmake VF=1 test # Windows
694 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
695 used even when parsing explicit parameters, when loading a serialized key
696 or calling `EC_GROUP_new_from_ecpkparameters()`/
697 `EC_GROUP_new_from_ecparameters()`.
698 This prevents bypass of security hardening and performance gains,
699 especially for curves with specialized EC_METHODs.
700 By default, if a key encoded with explicit parameters is loaded and later
701 serialized, the output is still encoded with explicit parameters, even if
702 internally a "named" EC_GROUP is used for computation.
706 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
707 this change, EC_GROUP_set_generator would accept order and/or cofactor as
708 NULL. After this change, only the cofactor parameter can be NULL. It also
709 does some minimal sanity checks on the passed order.
714 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
715 An attack is simple, if the first CMS_recipientInfo is valid but the
716 second CMS_recipientInfo is chosen ciphertext. If the second
717 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
718 encryption key will be replaced by garbage, and the message cannot be
719 decoded, but if the RSA decryption fails, the correct encryption key is
720 used and the recipient will not notice the attack.
721 As a work around for this potential attack the length of the decrypted
722 key must be equal to the cipher default key length, in case the
723 certifiate is not given and all recipientInfo are tried out.
724 The old behaviour can be re-enabled in the CMS code by setting the
725 CMS_DEBUG_DECRYPT flag.
729 * Early start up entropy quality from the DEVRANDOM seed source has been
730 improved for older Linux systems. The RAND subsystem will wait for
731 /dev/random to be producing output before seeding from /dev/urandom.
732 The seeded state is stored for future library initialisations using
733 a system global shared memory segment. The shared memory identifier
734 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
735 the desired value. The default identifier is 114.
739 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
740 when primes for RSA keys are computed.
741 Since we previously always generated primes == 2 (mod 3) for RSA keys,
742 the 2-prime and 3-prime RSA modules were easy to distinguish, since
743 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
744 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
745 This avoids possible fingerprinting of newly generated RSA modules.
749 * Correct the extended master secret constant on EBCDIC systems. Without this
750 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
751 negotiate EMS will fail. Unfortunately this also means that TLS connections
752 between EBCDIC systems with this fix, and EBCDIC systems without this
753 fix will fail if they negotiate EMS.
757 * Changed the library initialisation so that the config file is now loaded
758 by default. This was already the case for libssl. It now occurs for both
759 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
760 OPENSSL_init_crypto() to suppress automatic loading of a config file.
764 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
765 where the former acts as a replacement for ERR_put_error(), and the
766 latter replaces the combination ERR_put_error()+ERR_add_error_data().
767 ERR_raise_data() adds more flexibility by taking a format string and
768 an arbitrary number of arguments following it, to be processed with
773 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
774 to check if a named provider is loaded and available. When called, it
775 will also activate all fallback providers if such are still present.
779 * Enforce a minimum DH modulus size of 512 bits.
783 * Changed DH parameters to generate the order q subgroup instead of 2q.
784 Previously generated DH parameters are still accepted by DH_check
785 but DH_generate_key works around that by clearing bit 0 of the
786 private key for those. This avoids leaking bit 0 of the private key.
790 * Significantly reduce secure memory usage by the randomness pools.
794 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
799 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
800 algorithms. An implementation of a key exchange algorithm can be obtained
801 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
802 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
803 the older EVP_PKEY_derive_init() function. See the man pages for the new
804 functions for further details.
808 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
812 * Removed the function names from error messages and deprecated the
815 * Removed NextStep support and the macro OPENSSL_UNISTD
819 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
820 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
821 Also removed "export var as function" capability; we do not export
822 variables, only functions.
826 * RC5_32_set_key has been changed to return an int type, with 0 indicating
827 an error and 1 indicating success. In previous versions of OpenSSL this
828 was a void type. If a key was set longer than the maximum possible this
833 * Support SM2 signing and verification schemes with X509 certificate.
837 * Use SHA256 as the default digest for TS query in the ts app.
841 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
842 This checks that the salt length is at least 128 bits, the derived key
843 length is at least 112 bits, and that the iteration count is at least 1000.
844 For backwards compatibility these checks are disabled by default in the
845 default provider, but are enabled by default in the fips provider.
846 To enable or disable these checks use the control
847 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
851 * Default cipher lists/suites are now available via a function, the
852 #defines are deprecated.
856 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
857 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
858 for Windows Store apps easier. Also, the "no-uplink" option has been added.
862 * Join the directories crypto/x509 and crypto/x509v3
866 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
867 This changes the size when using the genpkey app when no size is given. It
868 fixes an omission in earlier changes that changed all RSA, DSA and DH
869 generation apps to use 2048 bits by default.
873 * Added command 'openssl kdf' that uses the EVP_KDF API.
877 * Added command 'openssl mac' that uses the EVP_MAC API.
881 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
882 as default directories. Also added the command 'openssl info'
883 for scripting purposes.
887 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
888 deprecated. These undocumented functions were never integrated into the EVP
889 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
890 Bi-directional IGE mode. These modes were never formally standardised and
891 usage of these functions is believed to be very small. In particular
892 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
893 is ever used. The security implications are believed to be minimal, but
894 this issue was never fixed for backwards compatibility reasons. New code
895 should not use these modes.
899 * Add prediction resistance to the DRBG reseeding process.
903 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
904 mandated by IEEE Std 1619-2018.
908 * Added newline escaping functionality to a filename when using openssl dgst.
909 This output format is to replicate the output format found in the `*sum`
910 checksum programs. This aims to preserve backward compatibility.
912 *Matt Eaton, Richard Levitte, and Paul Dale*
914 * Removed the heartbeat message in DTLS feature, as it has very
915 little usage and doesn't seem to fulfill a valuable purpose.
916 The configuration option is now deprecated.
920 * Changed the output of 'openssl {digestname} < file' to display the
921 digest name in its output.
925 * Added a new generic trace API which provides support for enabling
926 instrumentation through trace output. This feature is mainly intended
927 as an aid for developers and is disabled by default. To utilize it,
928 OpenSSL needs to be configured with the `enable-trace` option.
930 If the tracing API is enabled, the application can activate trace output
931 by registering BIOs as trace channels for a number of tracing and debugging
934 The 'openssl' application has been expanded to enable any of the types
935 available via environment variables defined by the user, and serves as
936 one possible example on how to use this functionality.
938 *Richard Levitte & Matthias St. Pierre*
940 * Added build tests for C++. These are generated files that only do one
941 thing, to include one public OpenSSL head file each. This tests that
942 the public header files can be usefully included in a C++ application.
944 This test isn't enabled by default. It can be enabled with the option
945 'enable-buildtest-c++'.
949 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
953 * Add KMAC to EVP_MAC.
957 * Added property based algorithm implementation selection framework to
962 * Added SCA hardening for modular field inversion in EC_GROUP through
963 a new dedicated field_inv() pointer in EC_METHOD.
964 This also addresses a leakage affecting conversions from projective
965 to affine coordinates.
967 *Billy Bob Brumley, Nicola Tuveri*
969 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
970 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
971 those algorithms that were already supported through the EVP_PKEY API
972 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
973 and scrypt are now wrappers that call EVP_KDF.
977 * Build devcrypto engine as a dynamic engine.
981 * Add keyed BLAKE2 to EVP_MAC.
985 * Fix a bug in the computation of the endpoint-pair shared secret used
986 by DTLS over SCTP. This breaks interoperability with older versions
987 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
988 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
989 interoperability with such broken implementations. However, enabling
990 this switch breaks interoperability with correct implementations.
992 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
993 re-used X509_PUBKEY object if the second PUBKEY is malformed.
997 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1001 * Change the license to the Apache License v2.0.
1005 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1007 - Major releases (indicated by incrementing the MAJOR release number)
1008 may introduce incompatible API/ABI changes.
1009 - Minor releases (indicated by incrementing the MINOR release number)
1010 may introduce new features but retain API/ABI compatibility.
1011 - Patch releases (indicated by incrementing the PATCH number)
1012 are intended for bug fixes and other improvements of existing
1013 features only (like improving performance or adding documentation)
1014 and retain API/ABI compatibility.
1018 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1022 * Remove the 'dist' target and add a tarball building script. The
1023 'dist' target has fallen out of use, and it shouldn't be
1024 necessary to configure just to create a source distribution.
1028 * Recreate the OS390-Unix config target. It no longer relies on a
1029 special script like it did for OpenSSL pre-1.1.0.
1033 * Instead of having the source directories listed in Configure, add
1034 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1039 * Add GMAC to EVP_MAC.
1043 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1047 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1048 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1049 to facilitate the continued use of MACs through raw private keys in
1050 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
1054 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1055 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1059 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1060 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1061 are retained for backwards compatibility.
1065 * AES-XTS mode now enforces that its two keys are different to mitigate
1066 the attacked described in "Efficient Instantiations of Tweakable
1067 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1068 Details of this attack can be obtained from:
1069 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1073 * Rename the object files, i.e. give them other names than in previous
1074 versions. Their names now include the name of the final product, as
1075 well as its type mnemonic (bin, lib, shlib).
1079 * Added new option for 'openssl list', '-objects', which will display the
1080 list of built in objects, i.e. OIDs with names.
1084 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1085 improves application performance by removing data copies and providing
1086 applications with zero-copy system calls such as sendfile and splice.
1093 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1095 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1097 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1098 while reading in libssl then we would report an error back to the
1099 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1100 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1101 therefore give a hint as to what went wrong.
1105 * Check that ed25519 and ed448 are allowed by the security level. Previously
1106 signature algorithms not using an MD were not being checked that they were
1107 allowed by the security level.
1111 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1112 was not quite right. The behaviour was not consistent between resumption
1113 and normal handshakes, and also not quite consistent with historical
1114 behaviour. The behaviour in various scenarios has been clarified and
1115 it has been updated to make it match historical behaviour as closely as
1120 * *[VMS only]* The header files that the VMS compilers include automatically,
1121 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1122 that the C++ compiler doesn't understand. This is a shortcoming in the
1123 compiler, but can be worked around with `__cplusplus` guards.
1125 C++ applications that use OpenSSL libraries must be compiled using the
1126 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1127 functions. Otherwise, only functions with symbols of less than 31
1128 characters can be used, as the linker will not be able to successfully
1129 resolve symbols with longer names.
1133 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1134 The presence of this system service is determined at run-time.
1138 * Added newline escaping functionality to a filename when using openssl dgst.
1139 This output format is to replicate the output format found in the `*sum`
1140 checksum programs. This aims to preserve backward compatibility.
1142 *Matt Eaton, Richard Levitte, and Paul Dale*
1144 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1149 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1151 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1152 number generator (RNG). This was intended to include protection in the
1153 event of a fork() system call in order to ensure that the parent and child
1154 processes did not share the same RNG state. However this protection was not
1155 being used in the default case.
1157 A partial mitigation for this issue is that the output from a high
1158 precision timer is mixed into the RNG state so the likelihood of a parent
1159 and child process sharing state is significantly reduced.
1161 If an application already calls OPENSSL_init_crypto() explicitly using
1162 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1165 *Matthias St. Pierre*
1167 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1168 used even when parsing explicit parameters, when loading a serialized key
1169 or calling `EC_GROUP_new_from_ecpkparameters()`/
1170 `EC_GROUP_new_from_ecparameters()`.
1171 This prevents bypass of security hardening and performance gains,
1172 especially for curves with specialized EC_METHODs.
1173 By default, if a key encoded with explicit parameters is loaded and later
1174 serialized, the output is still encoded with explicit parameters, even if
1175 internally a "named" EC_GROUP is used for computation.
1179 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1180 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1181 NULL. After this change, only the cofactor parameter can be NULL. It also
1182 does some minimal sanity checks on the passed order.
1187 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1188 An attack is simple, if the first CMS_recipientInfo is valid but the
1189 second CMS_recipientInfo is chosen ciphertext. If the second
1190 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1191 encryption key will be replaced by garbage, and the message cannot be
1192 decoded, but if the RSA decryption fails, the correct encryption key is
1193 used and the recipient will not notice the attack.
1194 As a work around for this potential attack the length of the decrypted
1195 key must be equal to the cipher default key length, in case the
1196 certifiate is not given and all recipientInfo are tried out.
1197 The old behaviour can be re-enabled in the CMS code by setting the
1198 CMS_DEBUG_DECRYPT flag.
1203 * Early start up entropy quality from the DEVRANDOM seed source has been
1204 improved for older Linux systems. The RAND subsystem will wait for
1205 /dev/random to be producing output before seeding from /dev/urandom.
1206 The seeded state is stored for future library initialisations using
1207 a system global shared memory segment. The shared memory identifier
1208 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1209 the desired value. The default identifier is 114.
1213 * Correct the extended master secret constant on EBCDIC systems. Without this
1214 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1215 negotiate EMS will fail. Unfortunately this also means that TLS connections
1216 between EBCDIC systems with this fix, and EBCDIC systems without this
1217 fix will fail if they negotiate EMS.
1221 * Use Windows installation paths in the mingw builds
1223 Mingw isn't a POSIX environment per se, which means that Windows
1224 paths should be used for installation.
1229 * Changed DH_check to accept parameters with order q and 2q subgroups.
1230 With order 2q subgroups the bit 0 of the private key is not secret
1231 but DH_generate_key works around that by clearing bit 0 of the
1232 private key for those. This avoids leaking bit 0 of the private key.
1236 * Significantly reduce secure memory usage by the randomness pools.
1240 * Revert the DEVRANDOM_WAIT feature for Linux systems
1242 The DEVRANDOM_WAIT feature added a select() call to wait for the
1243 /dev/random device to become readable before reading from the
1244 /dev/urandom device.
1246 It turned out that this change had negative side effects on
1247 performance which were not acceptable. After some discussion it
1248 was decided to revert this feature and leave it up to the OS
1249 resp. the platform maintainer to ensure a proper initialization
1250 during early boot time.
1252 *Matthias St. Pierre*
1254 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1256 * Add build tests for C++. These are generated files that only do one
1257 thing, to include one public OpenSSL head file each. This tests that
1258 the public header files can be usefully included in a C++ application.
1260 This test isn't enabled by default. It can be enabled with the option
1261 'enable-buildtest-c++'.
1265 * Enable SHA3 pre-hashing for ECDSA and DSA.
1269 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1270 This changes the size when using the genpkey app when no size is given. It
1271 fixes an omission in earlier changes that changed all RSA, DSA and DH
1272 generation apps to use 2048 bits by default.
1276 * Reorganize the manual pages to consistently have RETURN VALUES,
1277 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1278 util/fix-doc-nits accordingly.
1280 *Paul Yang, Joshua Lock*
1282 * Add the missing accessor EVP_PKEY_get0_engine()
1286 * Have apps like 's_client' and 's_server' output the signature scheme
1287 along with other cipher suite parameters when debugging.
1291 * Make OPENSSL_config() error agnostic again.
1295 * Do the error handling in RSA decryption constant time.
1299 * Prevent over long nonces in ChaCha20-Poly1305.
1301 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1302 for every encryption operation. RFC 7539 specifies that the nonce value
1303 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1304 and front pads the nonce with 0 bytes if it is less than 12
1305 bytes. However it also incorrectly allows a nonce to be set of up to 16
1306 bytes. In this case only the last 12 bytes are significant and any
1307 additional leading bytes are ignored.
1309 It is a requirement of using this cipher that nonce values are
1310 unique. Messages encrypted using a reused nonce value are susceptible to
1311 serious confidentiality and integrity attacks. If an application changes
1312 the default nonce length to be longer than 12 bytes and then makes a
1313 change to the leading bytes of the nonce expecting the new value to be a
1314 new unique nonce then such an application could inadvertently encrypt
1315 messages with a reused nonce.
1317 Additionally the ignored bytes in a long nonce are not covered by the
1318 integrity guarantee of this cipher. Any application that relies on the
1319 integrity of these ignored leading bytes of a long nonce may be further
1320 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1321 is safe because no such use sets such a long nonce value. However user
1322 applications that use this cipher directly and set a non-default nonce
1323 length to be longer than 12 bytes may be vulnerable.
1325 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1331 * Add DEVRANDOM_WAIT feature for Linux systems
1333 On older Linux systems where the getrandom() system call is not available,
1334 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1335 Contrary to getrandom(), the /dev/urandom device will not block during
1336 early boot when the kernel CSPRNG has not been seeded yet.
1338 To mitigate this known weakness, use select() to wait for /dev/random to
1339 become readable before reading from /dev/urandom.
1341 * Ensure that SM2 only uses SM3 as digest algorithm
1345 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1347 * Change the info callback signals for the start and end of a post-handshake
1348 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1349 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1350 confused by this and assume that a TLSv1.2 renegotiation has started. This
1351 can break KeyUpdate handling. Instead we no longer signal the start and end
1352 of a post handshake message exchange (although the messages themselves are
1353 still signalled). This could break some applications that were expecting
1354 the old signals. However without this KeyUpdate is not usable for many
1359 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1361 * Timing vulnerability in DSA signature generation
1363 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1364 timing side channel attack. An attacker could use variations in the signing
1365 algorithm to recover the private key.
1367 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1372 * Timing vulnerability in ECDSA signature generation
1374 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1375 timing side channel attack. An attacker could use variations in the signing
1376 algorithm to recover the private key.
1378 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1383 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1384 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1385 of two gigabytes and the error handling improved.
1387 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1388 categorized as a normal bug, not a security issue, because the DRBG reseeds
1389 automatically and is fully functional even without additional randomness
1390 provided by the application.
1392 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1394 * Add a new ClientHello callback. Provides a callback interface that gives
1395 the application the ability to adjust the nascent SSL object at the
1396 earliest stage of ClientHello processing, immediately after extensions have
1397 been collected but before they have been processed. In particular, this
1398 callback can adjust the supported TLS versions in response to the contents
1403 * Add SM2 base algorithm support.
1407 * s390x assembly pack: add (improved) hardware-support for the following
1408 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1409 aes-cfb/cfb8, aes-ecb.
1413 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1414 parameter is no longer accepted, as it leads to a corrupt table. NULL
1415 pem_str is reserved for alias entries only.
1419 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1420 step for prime curves. The new implementation is based on formulae from
1421 differential addition-and-doubling in homogeneous projective coordinates
1422 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1423 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1424 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1425 to work in projective coordinates.
1427 *Billy Bob Brumley, Nicola Tuveri*
1429 * Change generating and checking of primes so that the error rate of not
1430 being prime depends on the intended use based on the size of the input.
1431 For larger primes this will result in more rounds of Miller-Rabin.
1432 The maximal error rate for primes with more than 1080 bits is lowered
1435 *Kurt Roeckx, Annie Yousar*
1437 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1441 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1442 moving between systems, and to avoid confusion when a Windows build is
1443 done with mingw vs with MSVC. For POSIX installs, there's still a
1444 symlink or copy named 'tsget' to avoid that confusion as well.
1448 * Revert blinding in ECDSA sign and instead make problematic addition
1449 length-invariant. Switch even to fixed-length Montgomery multiplication.
1453 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1454 step for binary curves. The new implementation is based on formulae from
1455 differential addition-and-doubling in mixed Lopez-Dahab projective
1456 coordinates, modified to independently blind the operands.
1458 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1460 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1461 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1462 EC_METHODs to implement their own specialized "ladder step", to take
1463 advantage of more favorable coordinate systems or more efficient
1464 differential addition-and-doubling algorithms.
1466 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1468 * Modified the random device based seed sources to keep the relevant
1469 file descriptors open rather than reopening them on each access.
1470 This allows such sources to operate in a chroot() jail without
1471 the associated device nodes being available. This behaviour can be
1472 controlled using RAND_keep_random_devices_open().
1476 * Numerous side-channel attack mitigations have been applied. This may have
1477 performance impacts for some algorithms for the benefit of improved
1478 security. Specific changes are noted in this change log by their respective
1483 * AIX shared library support overhaul. Switch to AIX "natural" way of
1484 handling shared libraries, which means collecting shared objects of
1485 different versions and bitnesses in one common archive. This allows to
1486 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1487 doesn't affect the way 3rd party applications are linked, only how
1488 multi-version installation is managed.
1492 * Make ec_group_do_inverse_ord() more robust and available to other
1493 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1494 mitigations are applied to the fallback BN_mod_inverse().
1495 When using this function rather than BN_mod_inverse() directly, new
1496 EC cryptosystem implementations are then safer-by-default.
1500 * Add coordinate blinding for EC_POINT and implement projective
1501 coordinate blinding for generic prime curves as a countermeasure to
1502 chosen point SCA attacks.
1504 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1506 * Add blinding to ECDSA and DSA signatures to protect against side channel
1507 attacks discovered by Keegan Ryan (NCC Group).
1511 * Enforce checking in the pkeyutl command line app to ensure that the input
1512 length does not exceed the maximum supported digest length when performing
1513 a sign, verify or verifyrecover operation.
1517 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1518 I/O in combination with something like select() or poll() will hang. This
1519 can be turned off again using SSL_CTX_clear_mode().
1520 Many applications do not properly handle non-application data records, and
1521 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1522 around the problems in those applications, but can also break some.
1523 It's recommended to read the manpages about SSL_read(), SSL_write(),
1524 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1525 SSL_CTX_set_read_ahead() again.
1529 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1530 now allow empty (zero character) pass phrases.
1534 * Apply blinding to binary field modular inversion and remove patent
1535 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1539 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1540 binary and prime elliptic curves.
1544 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1545 constant time fixed point multiplication.
1549 * Revise elliptic curve scalar multiplication with timing attack
1550 defenses: ec_wNAF_mul redirects to a constant time implementation
1551 when computing fixed point and variable point multiplication (which
1552 in OpenSSL are mostly used with secret scalars in keygen, sign,
1553 ECDH derive operations).
1554 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1557 * Updated CONTRIBUTING
1561 * Updated DRBG / RAND to request nonce and additional low entropy
1562 randomness from the system.
1564 *Matthias St. Pierre*
1566 * Updated 'openssl rehash' to use OpenSSL consistent default.
1570 * Moved the load of the ssl_conf module to libcrypto, which helps
1571 loading engines that libssl uses before libssl is initialised.
1575 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1579 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1581 *Ingo Schwarze, Rich Salz*
1583 * Added output of accepting IP address and port for 'openssl s_server'
1587 * Added a new API for TLSv1.3 ciphersuites:
1588 SSL_CTX_set_ciphersuites()
1589 SSL_set_ciphersuites()
1593 * Memory allocation failures consistently add an error to the error
1598 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1599 in libcrypto when run as setuid/setgid.
1603 * Load any config file by default when libssl is used.
1607 * Added new public header file <openssl/rand_drbg.h> and documentation
1608 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1610 *Matthias St. Pierre*
1612 * QNX support removed (cannot find contributors to get their approval
1613 for the license change).
1617 * TLSv1.3 replay protection for early data has been implemented. See the
1618 SSL_read_early_data() man page for further details.
1622 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1623 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1624 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1625 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1626 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1627 configuration has been separated out. See the ciphers man page or the
1628 SSL_CTX_set_ciphersuites() man page for more information.
1632 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1633 in responder mode now supports the new "-multi" option, which
1634 spawns the specified number of child processes to handle OCSP
1635 requests. The "-timeout" option now also limits the OCSP
1636 responder's patience to wait to receive the full client request
1637 on a newly accepted connection. Child processes are respawned
1638 as needed, and the CA index file is automatically reloaded
1639 when changed. This makes it possible to run the "ocsp" responder
1640 as a long-running service, making the OpenSSL CA somewhat more
1641 feature-complete. In this mode, most diagnostic messages logged
1642 after entering the event loop are logged via syslog(3) rather than
1647 * Added support for X448 and Ed448. Heavily based on original work by
1652 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1653 objects loaded. This adds the functions OSSL_STORE_expect() and
1654 OSSL_STORE_find() as well as needed tools to construct searches and
1655 get the search data out of them.
1659 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1660 version of OpenSSL should review their configuration settings to ensure
1661 that they are still appropriate for TLSv1.3. For further information see:
1662 <https://wiki.openssl.org/index.php/TLS1.3>
1666 * Grand redesign of the OpenSSL random generator
1668 The default RAND method now utilizes an AES-CTR DRBG according to
1669 NIST standard SP 800-90Ar1. The new random generator is essentially
1670 a port of the default random generator from the OpenSSL FIPS 2.0
1671 object module. It is a hybrid deterministic random bit generator
1672 using an AES-CTR bit stream and which seeds and reseeds itself
1673 automatically using trusted system entropy sources.
1675 Some of its new features are:
1676 - Support for multiple DRBG instances with seed chaining.
1677 - The default RAND method makes use of a DRBG.
1678 - There is a public and private DRBG instance.
1679 - The DRBG instances are fork-safe.
1680 - Keep all global DRBG instances on the secure heap if it is enabled.
1681 - The public and private DRBG instance are per thread for lock free
1684 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1686 * Changed Configure so it only says what it does and doesn't dump
1687 so much data. Instead, ./configdata.pm should be used as a script
1688 to display all sorts of configuration data.
1692 * Added processing of "make variables" to Configure.
1696 * Added SHA512/224 and SHA512/256 algorithm support.
1700 * The last traces of Netware support, first removed in 1.1.0, have
1705 * Get rid of Makefile.shared, and in the process, make the processing
1706 of certain files (rc.obj, or the .def/.map/.opt files produced from
1707 the ordinal files) more visible and hopefully easier to trace and
1708 debug (or make silent).
1712 * Make it possible to have environment variable assignments as
1713 arguments to config / Configure.
1717 * Add multi-prime RSA (RFC 8017) support.
1721 * Add SM3 implemented according to GB/T 32905-2016
1722 * Jack Lloyd <jack.lloyd@ribose.com>,
1723 Ronald Tse <ronald.tse@ribose.com>,
1724 Erick Borsboom <erick.borsboom@ribose.com> *
1726 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1727 as documented in RFC6066.
1728 Based on a patch from Tomasz Moń
1730 *Filipe Raimundo da Silva*
1732 * Add SM4 implemented according to GB/T 32907-2016.
1733 * Jack Lloyd <jack.lloyd@ribose.com>,
1734 Ronald Tse <ronald.tse@ribose.com>,
1735 Erick Borsboom <erick.borsboom@ribose.com> *
1737 * Reimplement -newreq-nodes and ERR_error_string_n; the
1738 original author does not agree with the license change.
1742 * Add ARIA AEAD TLS support.
1746 * Some macro definitions to support VS6 have been removed. Visual
1747 Studio 6 has not worked since 1.1.0
1751 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1752 without clearing the errors.
1756 * Add "atfork" functions. If building on a system that without
1757 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1758 requirements. The RAND facility now uses/requires this.
1766 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1767 not possible to disable entirely. However, it's still possible to
1768 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1771 To disable, configure with 'no-ui-console'. 'no-ui' is still
1772 possible to use as an alias. Check at compile time with the
1773 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1774 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1778 * Add a STORE module, which implements a uniform and URI based reader of
1779 stores that can contain keys, certificates, CRLs and numerous other
1780 objects. The main API is loosely based on a few stdio functions,
1781 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1782 OSSL_STORE_error and OSSL_STORE_close.
1783 The implementation uses backends called "loaders" to implement arbitrary
1784 URI schemes. There is one built in "loader" for the 'file' scheme.
1788 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1789 then adjusted to work on FreeBSD 8.4 as well.
1790 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1791 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1795 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1796 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1797 error code calls like this:
1799 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1801 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1802 that can be encoded in C. For the foreseeable future, this will only
1805 *Richard Levitte and Tim Hudson*
1807 * Removed BSD cryptodev engine.
1811 * Add a build target 'build_all_generated', to build all generated files
1812 and only that. This can be used to prepare everything that requires
1813 things like perl for a system that lacks perl and then move everything
1814 to that system and do the rest of the build there.
1818 * In the UI interface, make it possible to duplicate the user data. This
1819 can be used by engines that need to retain the data for a longer time
1820 than just the call where this user data is passed.
1824 * Ignore the '-named_curve auto' value for compatibility of applications
1827 *Tomas Mraz <tmraz@fedoraproject.org>*
1829 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1830 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1831 alerts across multiple records (some of which could be empty). In practice
1832 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1833 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1834 support this at all. Supporting it adds significant complexity to the
1835 record layer, and its removal is unlikely to cause interoperability
1840 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1841 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1842 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1847 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1848 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1850 *Richard Levitte, Andy Polyakov*
1852 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1857 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1858 platform rather than 'mingw'.
1862 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1863 success if they are asked to add an object which already exists
1864 in the store. This change cascades to other functions which load
1865 certificates and CRLs.
1869 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1870 facilitate stack unwinding even from assembly subroutines.
1874 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1875 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1879 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1880 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1881 which is the minimum version we support.
1885 * Certificate time validation (X509_cmp_time) enforces stricter
1886 compliance with RFC 5280. Fractional seconds and timezone offsets
1887 are no longer allowed.
1891 * Add support for ARIA
1895 * s_client will now send the Server Name Indication (SNI) extension by
1896 default unless the new "-noservername" option is used. The server name is
1897 based on the host provided to the "-connect" option unless overridden by
1898 using "-servername".
1902 * Add support for SipHash
1906 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1907 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1908 prevent issues where no progress is being made and the peer continually
1909 sends unrecognised record types, using up resources processing them.
1913 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1914 using the algorithm defined in
1915 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1919 * Heartbeat support has been removed; the ABI is changed for now.
1921 *Richard Levitte, Rich Salz*
1923 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1927 * The RSA "null" method, which was partially supported to avoid patent
1928 issues, has been replaced to always returns NULL.
1935 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1937 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1938 used even when parsing explicit parameters, when loading a serialized key
1939 or calling `EC_GROUP_new_from_ecpkparameters()`/
1940 `EC_GROUP_new_from_ecparameters()`.
1941 This prevents bypass of security hardening and performance gains,
1942 especially for curves with specialized EC_METHODs.
1943 By default, if a key encoded with explicit parameters is loaded and later
1944 serialized, the output is still encoded with explicit parameters, even if
1945 internally a "named" EC_GROUP is used for computation.
1949 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1950 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1951 NULL. After this change, only the cofactor parameter can be NULL. It also
1952 does some minimal sanity checks on the passed order.
1957 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1958 An attack is simple, if the first CMS_recipientInfo is valid but the
1959 second CMS_recipientInfo is chosen ciphertext. If the second
1960 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1961 encryption key will be replaced by garbage, and the message cannot be
1962 decoded, but if the RSA decryption fails, the correct encryption key is
1963 used and the recipient will not notice the attack.
1964 As a work around for this potential attack the length of the decrypted
1965 key must be equal to the cipher default key length, in case the
1966 certifiate is not given and all recipientInfo are tried out.
1967 The old behaviour can be re-enabled in the CMS code by setting the
1968 CMS_DEBUG_DECRYPT flag.
1973 * Use Windows installation paths in the mingw builds
1975 Mingw isn't a POSIX environment per se, which means that Windows
1976 paths should be used for installation.
1981 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
1983 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1984 This changes the size when using the genpkey app when no size is given. It
1985 fixes an omission in earlier changes that changed all RSA, DSA and DH
1986 generation apps to use 2048 bits by default.
1990 * Prevent over long nonces in ChaCha20-Poly1305.
1992 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1993 for every encryption operation. RFC 7539 specifies that the nonce value
1994 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1995 and front pads the nonce with 0 bytes if it is less than 12
1996 bytes. However it also incorrectly allows a nonce to be set of up to 16
1997 bytes. In this case only the last 12 bytes are significant and any
1998 additional leading bytes are ignored.
2000 It is a requirement of using this cipher that nonce values are
2001 unique. Messages encrypted using a reused nonce value are susceptible to
2002 serious confidentiality and integrity attacks. If an application changes
2003 the default nonce length to be longer than 12 bytes and then makes a
2004 change to the leading bytes of the nonce expecting the new value to be a
2005 new unique nonce then such an application could inadvertently encrypt
2006 messages with a reused nonce.
2008 Additionally the ignored bytes in a long nonce are not covered by the
2009 integrity guarantee of this cipher. Any application that relies on the
2010 integrity of these ignored leading bytes of a long nonce may be further
2011 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2012 is safe because no such use sets such a long nonce value. However user
2013 applications that use this cipher directly and set a non-default nonce
2014 length to be longer than 12 bytes may be vulnerable.
2016 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2022 * Added SCA hardening for modular field inversion in EC_GROUP through
2023 a new dedicated field_inv() pointer in EC_METHOD.
2024 This also addresses a leakage affecting conversions from projective
2025 to affine coordinates.
2027 *Billy Bob Brumley, Nicola Tuveri*
2029 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2030 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2034 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2038 * Remove the 'dist' target and add a tarball building script. The
2039 'dist' target has fallen out of use, and it shouldn't be
2040 necessary to configure just to create a source distribution.
2044 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2046 * Timing vulnerability in DSA signature generation
2048 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2049 timing side channel attack. An attacker could use variations in the signing
2050 algorithm to recover the private key.
2052 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2057 * Timing vulnerability in ECDSA signature generation
2059 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2060 timing side channel attack. An attacker could use variations in the signing
2061 algorithm to recover the private key.
2063 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2068 * Add coordinate blinding for EC_POINT and implement projective
2069 coordinate blinding for generic prime curves as a countermeasure to
2070 chosen point SCA attacks.
2072 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2074 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2076 * Client DoS due to large DH parameter
2078 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2079 malicious server can send a very large prime value to the client. This will
2080 cause the client to spend an unreasonably long period of time generating a
2081 key for this prime resulting in a hang until the client has finished. This
2082 could be exploited in a Denial Of Service attack.
2084 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2089 * Cache timing vulnerability in RSA Key Generation
2091 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2092 a cache timing side channel attack. An attacker with sufficient access to
2093 mount cache timing attacks during the RSA key generation process could
2094 recover the private key.
2096 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2097 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2102 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2103 parameter is no longer accepted, as it leads to a corrupt table. NULL
2104 pem_str is reserved for alias entries only.
2108 * Revert blinding in ECDSA sign and instead make problematic addition
2109 length-invariant. Switch even to fixed-length Montgomery multiplication.
2113 * Change generating and checking of primes so that the error rate of not
2114 being prime depends on the intended use based on the size of the input.
2115 For larger primes this will result in more rounds of Miller-Rabin.
2116 The maximal error rate for primes with more than 1080 bits is lowered
2119 *Kurt Roeckx, Annie Yousar*
2121 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2125 * Add blinding to ECDSA and DSA signatures to protect against side channel
2126 attacks discovered by Keegan Ryan (NCC Group).
2130 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2131 now allow empty (zero character) pass phrases.
2135 * Certificate time validation (X509_cmp_time) enforces stricter
2136 compliance with RFC 5280. Fractional seconds and timezone offsets
2137 are no longer allowed.
2141 * Fixed a text canonicalisation bug in CMS
2143 Where a CMS detached signature is used with text content the text goes
2144 through a canonicalisation process first prior to signing or verifying a
2145 signature. This process strips trailing space at the end of lines, converts
2146 line terminators to CRLF and removes additional trailing line terminators
2147 at the end of a file. A bug in the canonicalisation process meant that
2148 some characters, such as form-feed, were incorrectly treated as whitespace
2149 and removed. This is contrary to the specification (RFC5485). This fix
2150 could mean that detached text data signed with an earlier version of
2151 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2152 signed with a fixed OpenSSL may fail to verify with an earlier version of
2153 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2154 and use the "-binary" flag (for the "cms" command line application) or set
2155 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2159 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2161 * Constructed ASN.1 types with a recursive definition could exceed the stack
2163 Constructed ASN.1 types with a recursive definition (such as can be found
2164 in PKCS7) could eventually exceed the stack given malicious input with
2165 excessive recursion. This could result in a Denial Of Service attack. There
2166 are no such structures used within SSL/TLS that come from untrusted sources
2167 so this is considered safe.
2169 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2175 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2177 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2178 effectively reduced to only comparing the least significant bit of each
2179 byte. This allows an attacker to forge messages that would be considered as
2180 authenticated in an amount of tries lower than that guaranteed by the
2181 security claims of the scheme. The module can only be compiled by the
2182 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2184 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2190 * Add a build target 'build_all_generated', to build all generated files
2191 and only that. This can be used to prepare everything that requires
2192 things like perl for a system that lacks perl and then move everything
2193 to that system and do the rest of the build there.
2197 * Backport SSL_OP_NO_RENGOTIATION
2199 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2200 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2201 changes this is no longer possible in 1.1.0. Therefore the new
2202 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2203 1.1.0 to provide equivalent functionality.
2205 Note that if an application built against 1.1.0h headers (or above) is run
2206 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2207 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2211 * Removed the OS390-Unix config target. It relied on a script that doesn't
2216 * rsaz_1024_mul_avx2 overflow bug on x86_64
2218 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2219 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2220 Analysis suggests that attacks against RSA and DSA as a result of this
2221 defect would be very difficult to perform and are not believed likely.
2222 Attacks against DH1024 are considered just feasible, because most of the
2223 work necessary to deduce information about a private key may be performed
2224 offline. The amount of resources required for such an attack would be
2225 significant. However, for an attack on TLS to be meaningful, the server
2226 would have to share the DH1024 private key among multiple clients, which is
2227 no longer an option since CVE-2016-0701.
2229 This only affects processors that support the AVX2 but not ADX extensions
2230 like Intel Haswell (4th generation).
2232 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2233 was originally found via the OSS-Fuzz project.
2238 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2240 * bn_sqrx8x_internal carry bug on x86_64
2242 There is a carry propagating bug in the x86_64 Montgomery squaring
2243 procedure. No EC algorithms are affected. Analysis suggests that attacks
2244 against RSA and DSA as a result of this defect would be very difficult to
2245 perform and are not believed likely. Attacks against DH are considered just
2246 feasible (although very difficult) because most of the work necessary to
2247 deduce information about a private key may be performed offline. The amount
2248 of resources required for such an attack would be very significant and
2249 likely only accessible to a limited number of attackers. An attacker would
2250 additionally need online access to an unpatched system using the target
2251 private key in a scenario with persistent DH parameters and a private
2252 key that is shared between multiple clients.
2254 This only affects processors that support the BMI1, BMI2 and ADX extensions
2255 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2257 This issue was reported to OpenSSL by the OSS-Fuzz project.
2262 * Malformed X.509 IPAddressFamily could cause OOB read
2264 If an X.509 certificate has a malformed IPAddressFamily extension,
2265 OpenSSL could do a one-byte buffer overread. The most likely result
2266 would be an erroneous display of the certificate in text format.
2268 This issue was reported to OpenSSL by the OSS-Fuzz project.
2273 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2275 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2276 platform rather than 'mingw'.
2280 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2281 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2282 which is the minimum version we support.
2286 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2288 * Encrypt-Then-Mac renegotiation crash
2290 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2291 negotiated where it was not in the original handshake (or vice-versa) then
2292 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2293 and servers are affected.
2295 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2300 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2302 * Truncated packet could crash via OOB read
2304 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2305 cipher is being used, then a truncated packet can cause that host to
2306 perform an out-of-bounds read, usually resulting in a crash.
2308 This issue was reported to OpenSSL by Robert Święcki of Google.
2313 * Bad (EC)DHE parameters cause a client crash
2315 If a malicious server supplies bad parameters for a DHE or ECDHE key
2316 exchange then this can result in the client attempting to dereference a
2317 NULL pointer leading to a client crash. This could be exploited in a Denial
2320 This issue was reported to OpenSSL by Guido Vranken.
2325 * BN_mod_exp may produce incorrect results on x86_64
2327 There is a carry propagating bug in the x86_64 Montgomery squaring
2328 procedure. No EC algorithms are affected. Analysis suggests that attacks
2329 against RSA and DSA as a result of this defect would be very difficult to
2330 perform and are not believed likely. Attacks against DH are considered just
2331 feasible (although very difficult) because most of the work necessary to
2332 deduce information about a private key may be performed offline. The amount
2333 of resources required for such an attack would be very significant and
2334 likely only accessible to a limited number of attackers. An attacker would
2335 additionally need online access to an unpatched system using the target
2336 private key in a scenario with persistent DH parameters and a private
2337 key that is shared between multiple clients. For example this can occur by
2338 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2339 similar to CVE-2015-3193 but must be treated as a separate problem.
2341 This issue was reported to OpenSSL by the OSS-Fuzz project.
2346 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2348 * ChaCha20/Poly1305 heap-buffer-overflow
2350 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2351 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2352 crash. This issue is not considered to be exploitable beyond a DoS.
2354 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2359 * CMS Null dereference
2361 Applications parsing invalid CMS structures can crash with a NULL pointer
2362 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2363 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2364 structure callback if an attempt is made to free certain invalid encodings.
2365 Only CHOICE structures using a callback which do not handle NULL value are
2368 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2373 * Montgomery multiplication may produce incorrect results
2375 There is a carry propagating bug in the Broadwell-specific Montgomery
2376 multiplication procedure that handles input lengths divisible by, but
2377 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2378 and DH private keys are impossible. This is because the subroutine in
2379 question is not used in operations with the private key itself and an input
2380 of the attacker's direct choice. Otherwise the bug can manifest itself as
2381 transient authentication and key negotiation failures or reproducible
2382 erroneous outcome of public-key operations with specially crafted input.
2383 Among EC algorithms only Brainpool P-512 curves are affected and one
2384 presumably can attack ECDH key negotiation. Impact was not analyzed in
2385 detail, because pre-requisites for attack are considered unlikely. Namely
2386 multiple clients have to choose the curve in question and the server has to
2387 share the private key among them, neither of which is default behaviour.
2388 Even then only clients that chose the curve will be affected.
2390 This issue was publicly reported as transient failures and was not
2391 initially recognized as a security issue. Thanks to Richard Morgan for
2392 providing reproducible case.
2397 * Removed automatic addition of RPATH in shared libraries and executables,
2398 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2402 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2404 * Fix Use After Free for large message sizes
2406 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2407 message larger than approx 16k is received then the underlying buffer to
2408 store the incoming message is reallocated and moved. Unfortunately a
2409 dangling pointer to the old location is left which results in an attempt to
2410 write to the previously freed location. This is likely to result in a
2411 crash, however it could potentially lead to execution of arbitrary code.
2413 This issue only affects OpenSSL 1.1.0a.
2415 This issue was reported to OpenSSL by Robert Święcki.
2420 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2422 * OCSP Status Request extension unbounded memory growth
2424 A malicious client can send an excessively large OCSP Status Request
2425 extension. If that client continually requests renegotiation, sending a
2426 large OCSP Status Request extension each time, then there will be unbounded
2427 memory growth on the server. This will eventually lead to a Denial Of
2428 Service attack through memory exhaustion. Servers with a default
2429 configuration are vulnerable even if they do not support OCSP. Builds using
2430 the "no-ocsp" build time option are not affected.
2432 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2437 * SSL_peek() hang on empty record
2439 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2440 sends an empty record. This could be exploited by a malicious peer in a
2441 Denial Of Service attack.
2443 This issue was reported to OpenSSL by Alex Gaynor.
2448 * Excessive allocation of memory in tls_get_message_header() and
2449 dtls1_preprocess_fragment()
2451 A (D)TLS message includes 3 bytes for its length in the header for the
2452 message. This would allow for messages up to 16Mb in length. Messages of
2453 this length are excessive and OpenSSL includes a check to ensure that a
2454 peer is sending reasonably sized messages in order to avoid too much memory
2455 being consumed to service a connection. A flaw in the logic of version
2456 1.1.0 means that memory for the message is allocated too early, prior to
2457 the excessive message length check. Due to way memory is allocated in
2458 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2459 to service a connection. This could lead to a Denial of Service through
2460 memory exhaustion. However, the excessive message length check still takes
2461 place, and this would cause the connection to immediately fail. Assuming
2462 that the application calls SSL_free() on the failed connection in a timely
2463 manner then the 21Mb of allocated memory will then be immediately freed
2464 again. Therefore the excessive memory allocation will be transitory in
2465 nature. This then means that there is only a security impact if:
2467 1) The application does not call SSL_free() in a timely manner in the event
2468 that the connection fails
2470 2) The application is working in a constrained environment where there is
2471 very little free memory
2473 3) The attacker initiates multiple connection attempts such that there are
2474 multiple connections in a state where memory has been allocated for the
2475 connection; SSL_free() has not yet been called; and there is insufficient
2476 memory to service the multiple requests.
2478 Except in the instance of (1) above any Denial Of Service is likely to be
2479 transitory because as soon as the connection fails the memory is
2480 subsequently freed again in the SSL_free() call. However there is an
2481 increased risk during this period of application crashes due to the lack of
2482 memory - which would then mean a more serious Denial of Service.
2484 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2485 (CVE-2016-6307 and CVE-2016-6308)
2489 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2490 had to be removed. Primary reason is that vendor assembler can't
2491 assemble our modules with -KPIC flag. As result it, assembly
2492 support, was not even available as option. But its lack means
2493 lack of side-channel resistant code, which is incompatible with
2494 security by todays standards. Fortunately gcc is readily available
2495 prepackaged option, which we firmly point at...
2499 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2501 * Windows command-line tool supports UTF-8 opt-in option for arguments
2502 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2503 (to any value) allows Windows user to access PKCS#12 file generated
2504 with Windows CryptoAPI and protected with non-ASCII password, as well
2505 as files generated under UTF-8 locale on Linux also protected with
2510 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2511 have been disabled by default and removed from DEFAULT, just like RC4.
2512 See the RC4 item below to re-enable both.
2516 * The method for finding the storage location for the Windows RAND seed file
2517 has changed. First we check %RANDFILE%. If that is not set then we check
2518 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2519 all else fails we fall back to C:\.
2523 * The EVP_EncryptUpdate() function has had its return type changed from void
2524 to int. A return of 0 indicates and error while a return of 1 indicates
2529 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2530 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2531 off the constant time implementation for RSA, DSA and DH have been made
2532 no-ops and deprecated.
2536 * Windows RAND implementation was simplified to only get entropy by
2537 calling CryptGenRandom(). Various other RAND-related tickets
2540 *Joseph Wylie Yandle, Rich Salz*
2542 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2543 and `OPENSSL_LH_`, respectively. The old names are available
2544 with API compatibility. They new names are now completely documented.
2548 * Unify TYPE_up_ref(obj) methods signature.
2549 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2550 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2551 int (instead of void) like all others TYPE_up_ref() methods.
2552 So now these methods also check the return value of CRYPTO_atomic_add(),
2553 and the validity of object reference counter.
2555 *fdasilvayy@gmail.com*
2557 * With Windows Visual Studio builds, the .pdb files are installed
2558 alongside the installed libraries and executables. For a static
2559 library installation, ossl_static.pdb is the associate compiler
2560 generated .pdb file to be used when linking programs.
2564 * Remove openssl.spec. Packaging files belong with the packagers.
2568 * Automatic Darwin/OSX configuration has had a refresh, it will now
2569 recognise x86_64 architectures automatically. You can still decide
2570 to build for a different bitness with the environment variable
2571 KERNEL_BITS (can be 32 or 64), for example:
2573 KERNEL_BITS=32 ./config
2577 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2578 256 bit AES and HMAC with SHA256.
2582 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2586 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2590 * To enable users to have their own config files and build file templates,
2591 Configure looks in the directory indicated by the environment variable
2592 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2593 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2594 name and is used as is.
2598 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2599 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2600 X509_CERT_FILE_CTX was removed.
2604 * "shared" builds are now the default. To create only static libraries use
2605 the "no-shared" Configure option.
2609 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2610 All of these option have not worked for some while and are fundamental
2615 * Make various cleanup routines no-ops and mark them as deprecated. Most
2616 global cleanup functions are no longer required because they are handled
2617 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2618 Explicitly de-initing can cause problems (e.g. where a library that uses
2619 OpenSSL de-inits, but an application is still using it). The affected
2620 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2621 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2622 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2623 COMP_zlib_cleanup().
2627 * --strict-warnings no longer enables runtime debugging options
2628 such as REF_DEBUG. Instead, debug options are automatically
2629 enabled with '--debug' builds.
2631 *Andy Polyakov, Emilia Käsper*
2633 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2634 have been moved out of the public header files. New functions for managing
2635 these have been added.
2639 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2640 objects have been moved out of the public header files. New
2641 functions for managing these have been added.
2645 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2646 have been moved out of the public header files. New functions for managing
2647 these have been added.
2651 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2652 moved out of the public header files. New functions for managing these
2657 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2661 * Removed the mk1mf build scripts.
2665 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2666 it is always safe to #include a header now.
2670 * Removed the aged BC-32 config and all its supporting scripts
2674 * Removed support for Ultrix, Netware, and OS/2.
2678 * Add support for HKDF.
2680 *Alessandro Ghedini*
2682 * Add support for blake2b and blake2s
2686 * Added support for "pipelining". Ciphers that have the
2687 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2688 encryptions/decryptions simultaneously. There are currently no built-in
2689 ciphers with this property but the expectation is that engines will be able
2690 to offer it to significantly improve throughput. Support has been extended
2691 into libssl so that multiple records for a single connection can be
2692 processed in one go (for >=TLS 1.1).
2696 * Added the AFALG engine. This is an async capable engine which is able to
2697 offload work to the Linux kernel. In this initial version it only supports
2698 AES128-CBC. The kernel must be version 4.1.0 or greater.
2702 * OpenSSL now uses a new threading API. It is no longer necessary to
2703 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2704 are two supported threading models: pthreads and windows threads. It is
2705 also possible to configure OpenSSL at compile time for "no-threads". The
2706 old threading API should no longer be used. The functions have been
2707 replaced with "no-op" compatibility macros.
2709 *Alessandro Ghedini, Matt Caswell*
2711 * Modify behavior of ALPN to invoke callback after SNI/servername
2712 callback, such that updates to the SSL_CTX affect ALPN.
2716 * Add SSL_CIPHER queries for authentication and key-exchange.
2720 * Changes to the DEFAULT cipherlist:
2721 - Prefer (EC)DHE handshakes over plain RSA.
2722 - Prefer AEAD ciphers over legacy ciphers.
2723 - Prefer ECDSA over RSA when both certificates are available.
2724 - Prefer TLSv1.2 ciphers/PRF.
2725 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2730 * Change the ECC default curve list to be this, in order: x25519,
2731 secp256r1, secp521r1, secp384r1.
2735 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2736 disabled by default. They can be re-enabled using the
2737 enable-weak-ssl-ciphers option to Configure.
2741 * If the server has ALPN configured, but supports no protocols that the
2742 client advertises, send a fatal "no_application_protocol" alert.
2743 This behaviour is SHALL in RFC 7301, though it isn't universally
2744 implemented by other servers.
2748 * Add X25519 support.
2749 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2750 for public and private key encoding using the format documented in
2751 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2752 key generation and key derivation.
2754 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2759 * Deprecate SRP_VBASE_get_by_user.
2760 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2761 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2762 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2763 seed, even if the seed is configured.
2765 Users should use SRP_VBASE_get1_by_user instead. Note that in
2766 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2767 also that even though configuring the SRP seed attempts to hide
2768 invalid usernames by continuing the handshake with fake
2769 credentials, this behaviour is not constant time and no strong
2770 guarantees are made that the handshake is indistinguishable from
2771 that of a valid user.
2775 * Configuration change; it's now possible to build dynamic engines
2776 without having to build shared libraries and vice versa. This
2777 only applies to the engines in engines/, those in crypto/engine/
2778 will always be built into libcrypto (i.e. "static").
2780 Building dynamic engines is enabled by default; to disable, use
2781 the configuration option "disable-dynamic-engine".
2783 The only requirements for building dynamic engines are the
2784 presence of the DSO module and building with position independent
2785 code, so they will also automatically be disabled if configuring
2786 with "disable-dso" or "disable-pic".
2788 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2789 are also taken away from openssl/opensslconf.h, as they are
2794 * Configuration change; if there is a known flag to compile
2795 position independent code, it will always be applied on the
2796 libcrypto and libssl object files, and never on the application
2797 object files. This means other libraries that use routines from
2798 libcrypto / libssl can be made into shared libraries regardless
2799 of how OpenSSL was configured.
2801 If this isn't desirable, the configuration options "disable-pic"
2802 or "no-pic" can be used to disable the use of PIC. This will
2803 also disable building shared libraries and dynamic engines.
2807 * Removed JPAKE code. It was experimental and has no wide use.
2811 * The INSTALL_PREFIX Makefile variable has been renamed to
2812 DESTDIR. That makes for less confusion on what this variable
2813 is for. Also, the configuration option --install_prefix is
2818 * Heartbeat for TLS has been removed and is disabled by default
2819 for DTLS; configure with enable-heartbeats. Code that uses the
2820 old #define's might need to be updated.
2822 *Emilia Käsper, Rich Salz*
2824 * Rename REF_CHECK to REF_DEBUG.
2828 * New "unified" build system
2830 The "unified" build system is aimed to be a common system for all
2831 platforms we support. With it comes new support for VMS.
2833 This system builds supports building in a different directory tree
2834 than the source tree. It produces one Makefile (for unix family
2835 or lookalikes), or one descrip.mms (for VMS).
2837 The source of information to make the Makefile / descrip.mms is
2838 small files called 'build.info', holding the necessary
2839 information for each directory with source to compile, and a
2840 template in Configurations, like unix-Makefile.tmpl or
2843 With this change, the library names were also renamed on Windows
2844 and on VMS. They now have names that are closer to the standard
2845 on Unix, and include the major version number, and in certain
2846 cases, the architecture they are built for. See "Notes on shared
2847 libraries" in INSTALL.
2849 We rely heavily on the perl module Text::Template.
2853 * Added support for auto-initialisation and de-initialisation of the library.
2854 OpenSSL no longer requires explicit init or deinit routines to be called,
2855 except in certain circumstances. See the OPENSSL_init_crypto() and
2856 OPENSSL_init_ssl() man pages for further information.
2860 * The arguments to the DTLSv1_listen function have changed. Specifically the
2861 "peer" argument is now expected to be a BIO_ADDR object.
2863 * Rewrite of BIO networking library. The BIO library lacked consistent
2864 support of IPv6, and adding it required some more extensive
2865 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2866 which hold all types of addresses and chains of address information.
2867 It also introduces a new API, with functions like BIO_socket,
2868 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2869 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2870 have been adapted accordingly.
2874 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2879 * CRIME protection: disable compression by default, even if OpenSSL is
2880 compiled with zlib enabled. Applications can still enable compression
2881 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2882 using the SSL_CONF library to configure compression.
2886 * The signature of the session callback configured with
2887 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2888 was explicitly marked as `const unsigned char*` instead of
2893 * Always DPURIFY. Remove the use of uninitialized memory in the
2894 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2898 * Removed many obsolete configuration items, including
2899 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2900 MD2_CHAR, MD2_INT, MD2_LONG
2902 IDEA_SHORT, IDEA_LONG
2903 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2905 *Rich Salz, with advice from Andy Polyakov*
2907 * Many BN internals have been moved to an internal header file.
2909 *Rich Salz with help from Andy Polyakov*
2911 * Configuration and writing out the results from it has changed.
2912 Files such as Makefile include/openssl/opensslconf.h and are now
2913 produced through general templates, such as Makefile.in and
2914 crypto/opensslconf.h.in and some help from the perl module
2917 Also, the center of configuration information is no longer
2918 Makefile. Instead, Configure produces a perl module in
2919 configdata.pm which holds most of the config data (in the hash
2920 table %config), the target data that comes from the target
2921 configuration in one of the `Configurations/*.conf~ files (in
2926 * To clarify their intended purposes, the Configure options
2927 --prefix and --openssldir change their semantics, and become more
2928 straightforward and less interdependent.
2930 --prefix shall be used exclusively to give the location INSTALLTOP
2931 where programs, scripts, libraries, include files and manuals are
2932 going to be installed. The default is now /usr/local.
2934 --openssldir shall be used exclusively to give the default
2935 location OPENSSLDIR where certificates, private keys, CRLs are
2936 managed. This is also where the default openssl.cnf gets
2938 If the directory given with this option is a relative path, the
2939 values of both the --prefix value and the --openssldir value will
2940 be combined to become OPENSSLDIR.
2941 The default for --openssldir is INSTALLTOP/ssl.
2943 Anyone who uses --openssldir to specify where OpenSSL is to be
2944 installed MUST change to use --prefix instead.
2948 * The GOST engine was out of date and therefore it has been removed. An up
2949 to date GOST engine is now being maintained in an external repository.
2950 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2951 support for GOST ciphersuites (these are only activated if a GOST engine
2956 * EGD is no longer supported by default; use enable-egd when
2959 *Ben Kaduk and Rich Salz*
2961 * The distribution now has Makefile.in files, which are used to
2962 create Makefile's when Configure is run. *Configure must be run
2963 before trying to build now.*
2967 * The return value for SSL_CIPHER_description() for error conditions
2972 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2974 Obtaining and performing DNSSEC validation of TLSA records is
2975 the application's responsibility. The application provides
2976 the TLSA records of its choice to OpenSSL, and these are then
2977 used to authenticate the peer.
2979 The TLSA records need not even come from DNS. They can, for
2980 example, be used to implement local end-entity certificate or
2981 trust-anchor "pinning", where the "pin" data takes the form
2982 of TLSA records, which can augment or replace verification
2983 based on the usual WebPKI public certification authorities.
2987 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2988 continues to support deprecated interfaces in default builds.
2989 However, applications are strongly advised to compile their
2990 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2991 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2992 or the 1.1.0 releases.
2994 In environments in which all applications have been ported to
2995 not use any deprecated interfaces OpenSSL's Configure script
2996 should be used with the --api=1.1.0 option to entirely remove
2997 support for the deprecated features from the library and
2998 unconditionally disable them in the installed headers.
2999 Essentially the same effect can be achieved with the "no-deprecated"
3000 argument to Configure, except that this will always restrict
3001 the build to just the latest API, rather than a fixed API
3004 As applications are ported to future revisions of the API,
3005 they should update their compile-time OPENSSL_API_COMPAT define
3006 accordingly, but in most cases should be able to continue to
3007 compile with later releases.
3009 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3010 0x10000000L and 0x00908000L, respectively. However those
3011 versions did not support the OPENSSL_API_COMPAT feature, and
3012 so applications are not typically tested for explicit support
3013 of just the undeprecated features of either release.
3017 * Add support for setting the minimum and maximum supported protocol.
3018 It can bet set via the SSL_set_min_proto_version() and
3019 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3020 MaxProtocol. It's recommended to use the new APIs to disable
3021 protocols instead of disabling individual protocols using
3022 SSL_set_options() or SSL_CONF's Protocol. This change also
3023 removes support for disabling TLS 1.2 in the OpenSSL TLS
3024 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3028 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3032 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3033 and integrates ECDSA and ECDH functionality into EC. Implementations can
3034 now redirect key generation and no longer need to convert to or from
3037 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3038 include the ec.h header file instead.
3042 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3043 ciphers who are no longer supported and drops support the ephemeral RSA key
3044 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3048 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3049 opaque. For HMAC_CTX, the following constructors and destructors
3052 HMAC_CTX *HMAC_CTX_new(void);
3053 void HMAC_CTX_free(HMAC_CTX *ctx);
3055 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3056 destroy such methods has been added. See EVP_MD_meth_new(3) and
3057 EVP_CIPHER_meth_new(3) for documentation.
3060 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
3061 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
3062 EVP_MD_CTX_reset() should be called instead to reinitialise
3063 an already created structure.
3064 2) For consistency with the majority of our object creators and
3065 destructors, EVP_MD_CTX_(create|destroy) were renamed to
3066 EVP_MD_CTX_(new|free). The old names are retained as macros
3067 for deprecated builds.
3071 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3072 cryptographic operations to be performed asynchronously as long as an
3073 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3074 further details. Libssl has also had this capability integrated with the
3075 introduction of the new mode SSL_MODE_ASYNC and associated error
3076 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3077 pages. This work was developed in partnership with Intel Corp.
3081 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3082 always enabled now. If you want to disable the support you should
3083 exclude it using the list of supported ciphers. This also means that the
3084 "-no_ecdhe" option has been removed from s_server.
3088 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3089 SSL_{CTX_}set1_curves() which can set a list.
3093 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3094 curve you want to support using SSL_{CTX_}set1_curves().
3098 * State machine rewrite. The state machine code has been significantly
3099 refactored in order to remove much duplication of code and solve issues
3100 with the old code (see ssl/statem/README for further details). This change
3101 does have some associated API changes. Notably the SSL_state() function
3102 has been removed and replaced by SSL_get_state which now returns an
3103 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
3104 altogether. The previous handshake states defined in ssl.h and ssl3.h have
3109 * All instances of the string "ssleay" in the public API were replaced
3110 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3111 Some error codes related to internal RSA_eay API's were renamed.
3115 * The demo files in crypto/threads were moved to demo/threads.
3119 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3122 *Matt Caswell, Rich Salz*
3124 * New ASN.1 embed macro.
3126 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3127 structure is not allocated: it is part of the parent. That is instead of
3135 This reduces memory fragmentation and make it impossible to accidentally
3136 set a mandatory field to NULL.
3138 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3139 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3140 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3145 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3149 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3150 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3151 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3152 DES and RC4 ciphersuites.
3156 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3157 This changes the decoding behaviour for some invalid messages,
3158 though the change is mostly in the more lenient direction, and
3159 legacy behaviour is preserved as much as possible.
3163 * Fix no-stdio build.
3164 * David Woodhouse <David.Woodhouse@intel.com> and also
3165 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3167 * New testing framework
3168 The testing framework has been largely rewritten and is now using
3169 perl and the perl modules Test::Harness and an extended variant of
3170 Test::More called OpenSSL::Test to do its work. All test scripts in
3171 test/ have been rewritten into test recipes, and all direct calls to
3172 executables in test/Makefile have become individual recipes using the
3173 simplified testing OpenSSL::Test::Simple.
3175 For documentation on our testing modules, do:
3177 perldoc test/testlib/OpenSSL/Test/Simple.pm
3178 perldoc test/testlib/OpenSSL/Test.pm
3182 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3183 are used; the latter aborts on memory leaks (usually checked on exit).
3184 Some undocumented "set malloc, etc., hooks" functions were removed
3185 and others were changed. All are now documented.
3189 * In DSA_generate_parameters_ex, if the provided seed is too short,
3192 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3194 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3195 from RFC4279, RFC4785, RFC5487, RFC5489.
3197 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3198 original RSA_PSK patch.
3202 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3203 era flag was never set throughout the codebase (only read). Also removed
3204 SSL3_FLAGS_POP_BUFFER which was only used if
3205 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3209 * Changed the default name options in the "ca", "crl", "req" and "x509"
3210 to be "oneline" instead of "compat".
3214 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3215 not aware of clients that still exhibit this bug, and the workaround
3216 hasn't been working properly for a while.
3220 * The return type of BIO_number_read() and BIO_number_written() as well as
3221 the corresponding num_read and num_write members in the BIO structure has
3222 changed from unsigned long to uint64_t. On platforms where an unsigned
3223 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3228 * Given the pervasive nature of TLS extensions it is inadvisable to run
3229 OpenSSL without support for them. It also means that maintaining
3230 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3231 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3235 * Removed support for the two export grade static DH ciphersuites
3236 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3237 were newly added (along with a number of other static DH ciphersuites) to
3238 1.0.2. However the two export ones have *never* worked since they were
3239 introduced. It seems strange in any case to be adding new export
3240 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3244 * Version negotiation has been rewritten. In particular SSLv23_method(),
3245 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3246 and turned into macros which simply call the new preferred function names
3247 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3248 should use the new names instead. Also as part of this change the ssl23.h
3249 header file has been removed.
3253 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3254 code and the associated standard is no longer considered fit-for-purpose.
3258 * RT2547 was closed. When generating a private key, try to make the
3259 output file readable only by the owner. This behavior change might
3260 be noticeable when interacting with other software.
3262 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3267 * Added HTTP GET support to the ocsp command.
3271 * Changed default digest for the dgst and enc commands from MD5 to
3276 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3280 * Added support for TLS extended master secret from
3281 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3282 initial patch which was a great help during development.
3286 * All libssl internal structures have been removed from the public header
3287 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3288 now redundant). Users should not attempt to access internal structures
3289 directly. Instead they should use the provided API functions.
3293 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3294 Access to deprecated functions can be re-enabled by running config with
3295 "enable-deprecated". In addition applications wishing to use deprecated
3296 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3297 will, by default, disable some transitive includes that previously existed
3298 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3302 * Added support for OCB mode. OpenSSL has been granted a patent license
3303 compatible with the OpenSSL license for use of OCB. Details are available
3304 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3305 for OCB can be removed by calling config with no-ocb.
3309 * SSLv2 support has been removed. It still supports receiving a SSLv2
3310 compatible client hello.
3314 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3315 done while fixing the error code for the key-too-small case.
3317 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3319 * CA.sh has been removed; use CA.pl instead.
3323 * Removed old DES API.
3327 * Remove various unsupported platforms:
3333 Sinix/ReliantUNIX RM400
3338 16-bit platforms such as WIN16
3342 * Clean up OPENSSL_NO_xxx #define's
3343 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3344 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3345 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3346 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3347 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3348 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3349 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3350 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3351 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3352 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3356 * Cleaned up dead code
3357 Remove all but one '#ifdef undef' which is to be looked at.
3361 * Clean up calling of xxx_free routines.
3362 Just like free(), fix most of the xxx_free routines to accept
3363 NULL. Remove the non-null checks from callers. Save much code.
3367 * Add secure heap for storage of private keys (when possible).
3368 Add BIO_s_secmem(), CBIGNUM, etc.
3369 Contributed by Akamai Technologies under our Corporate CLA.
3373 * Experimental support for a new, fast, unbiased prime candidate generator,
3374 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3376 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3378 * New output format NSS in the sess_id command line tool. This allows
3379 exporting the session id and the master key in NSS keylog format.
3381 *Martin Kaiser <martin@kaiser.cx>*
3383 * Harmonize version and its documentation. -f flag is used to display
3386 *mancha <mancha1@zoho.com>*
3388 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3389 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3391 *mancha <mancha1@zoho.com>*
3393 * Fix some double frees. These are not thought to be exploitable.
3395 *mancha <mancha1@zoho.com>*
3397 * A missing bounds check in the handling of the TLS heartbeat extension
3398 can be used to reveal up to 64k of memory to a connected client or
3401 Thanks for Neel Mehta of Google Security for discovering this bug and to
3402 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3403 preparing the fix [CVE-2014-0160][]
3405 *Adam Langley, Bodo Moeller*
3407 * Fix for the attack described in the paper "Recovering OpenSSL
3408 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3409 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3410 <http://eprint.iacr.org/2014/140>
3412 Thanks to Yuval Yarom and Naomi Benger for discovering this
3413 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3415 *Yuval Yarom and Naomi Benger*
3417 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3418 this fixes a limitation in previous versions of OpenSSL.
3422 * Experimental encrypt-then-mac support.
3424 Experimental support for encrypt then mac from
3425 draft-gutmann-tls-encrypt-then-mac-02.txt
3427 To enable it set the appropriate extension number (0x42 for the test
3428 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3430 For non-compliant peers (i.e. just about everything) this should have no
3433 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3437 * Add EVP support for key wrapping algorithms, to avoid problems with
3438 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3439 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3440 algorithms and include tests cases.
3444 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3449 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3450 MGF1 digest and OAEP label.
3454 * Make openssl verify return errors.
3456 *Chris Palmer <palmer@google.com> and Ben Laurie*
3458 * New function ASN1_TIME_diff to calculate the difference between two
3459 ASN1_TIME structures or one structure and the current time.
3463 * Update fips_test_suite to support multiple command line options. New
3464 test to induce all self test errors in sequence and check expected
3469 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3470 sign or verify all in one operation.
3474 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3475 test programs and fips_test_suite. Includes functionality to parse
3476 the minimal script output of fipsalgest.pl directly.
3480 * Add authorisation parameter to FIPS_module_mode_set().
3484 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3488 * Use separate DRBG fields for internal and external flags. New function
3489 FIPS_drbg_health_check() to perform on demand health checking. Add
3490 generation tests to fips_test_suite with reduced health check interval to
3491 demonstrate periodic health checking. Add "nodh" option to
3492 fips_test_suite to skip very slow DH test.
3496 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3501 * More extensive health check for DRBG checking many more failure modes.
3502 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3503 combination: call this in fips_test_suite.
3507 * Add support for canonical generation of DSA parameter 'g'. See
3510 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3511 POST to handle HMAC cases.
3515 * Add functions FIPS_module_version() and FIPS_module_version_text()
3516 to return numerical and string versions of the FIPS module number.
3520 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3521 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3522 outside the validated module in the FIPS capable OpenSSL.
3526 * Minor change to DRBG entropy callback semantics. In some cases
3527 there is no multiple of the block length between min_len and
3528 max_len. Allow the callback to return more than max_len bytes
3529 of entropy but discard any extra: it is the callback's responsibility
3530 to ensure that the extra data discarded does not impact the
3531 requested amount of entropy.
3535 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3536 information in FIPS186-3, SP800-57 and SP800-131A.
3540 * CCM support via EVP. Interface is very similar to GCM case except we
3541 must supply all data in one chunk (i.e. no update, final) and the
3542 message length must be supplied if AAD is used. Add algorithm test
3547 * Initial version of POST overhaul. Add POST callback to allow the status
3548 of POST to be monitored and/or failures induced. Modify fips_test_suite
3549 to use callback. Always run all selftests even if one fails.
3553 * XTS support including algorithm test driver in the fips_gcmtest program.
3554 Note: this does increase the maximum key length from 32 to 64 bytes but
3555 there should be no binary compatibility issues as existing applications
3556 will never use XTS mode.
3560 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3561 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3562 performs algorithm blocking for unapproved PRNG types. Also do not
3563 set PRNG type in FIPS_mode_set(): leave this to the application.
3564 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3565 the standard OpenSSL PRNG: set additional data to a date time vector.
3569 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3570 This shouldn't present any incompatibility problems because applications
3571 shouldn't be using these directly and any that are will need to rethink
3572 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3576 * Extensive self tests and health checking required by SP800-90 DRBG.
3577 Remove strength parameter from FIPS_drbg_instantiate and always
3578 instantiate at maximum supported strength.
3582 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3586 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3590 * New function DH_compute_key_padded() to compute a DH key and pad with
3591 leading zeroes if needed: this complies with SP800-56A et al.
3595 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3596 anything, incomplete, subject to change and largely untested at present.
3600 * Modify fipscanisteronly build option to only build the necessary object
3601 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3605 * Add experimental option FIPSSYMS to give all symbols in
3606 fipscanister.o and FIPS or fips prefix. This will avoid
3607 conflicts with future versions of OpenSSL. Add perl script
3608 util/fipsas.pl to preprocess assembly language source files
3609 and rename any affected symbols.
3613 * Add selftest checks and algorithm block of non-fips algorithms in
3614 FIPS mode. Remove DES2 from selftests.
3618 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3619 return internal method without any ENGINE dependencies. Add new
3620 tiny fips sign and verify functions.
3624 * New build option no-ec2m to disable characteristic 2 code.
3628 * New build option "fipscanisteronly". This only builds fipscanister.o
3629 and (currently) associated fips utilities. Uses the file Makefile.fips
3630 instead of Makefile.org as the prototype.
3634 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3635 Update fips_gcmtest to use IV generator.
3639 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3640 setting output buffer to NULL. The `*Final` function must be
3641 called although it will not retrieve any additional data. The tag
3642 can be set or retrieved with a ctrl. The IV length is by default 12
3643 bytes (96 bits) but can be set to an alternative value. If the IV
3644 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3649 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3650 underlying do_cipher function handles all cipher semantics itself
3651 including padding and finalisation. This is useful if (for example)
3652 an ENGINE cipher handles block padding itself. The behaviour of
3653 do_cipher is subtly changed if this flag is set: the return value
3654 is the number of characters written to the output buffer (zero is
3655 no longer an error code) or a negative error code. Also if the
3656 input buffer is NULL and length 0 finalisation should be performed.
3660 * If a candidate issuer certificate is already part of the constructed
3661 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3665 * Improve forward-security support: add functions
3667 void SSL_CTX_set_not_resumable_session_callback(
3668 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3669 void SSL_set_not_resumable_session_callback(
3670 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3672 for use by SSL/TLS servers; the callback function will be called whenever a
3673 new session is created, and gets to decide whether the session may be
3674 cached to make it resumable (return 0) or not (return 1). (As by the
3675 SSL/TLS protocol specifications, the session_id sent by the server will be
3676 empty to indicate that the session is not resumable; also, the server will
3677 not generate RFC 4507 (RFC 5077) session tickets.)
3679 A simple reasonable callback implementation is to return is_forward_secure.
3680 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3681 by the SSL/TLS server library, indicating whether it can provide forward
3684 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3686 * New -verify_name option in command line utilities to set verification
3691 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3692 Add CMAC pkey methods.
3696 * Experimental renegotiation in s_server -www mode. If the client
3697 browses /reneg connection is renegotiated. If /renegcert it is
3698 renegotiated requesting a certificate.
3702 * Add an "external" session cache for debugging purposes to s_server. This
3703 should help trace issues which normally are only apparent in deployed
3704 multi-process servers.
3708 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3709 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3710 BIO_set_cipher() and some obscure PEM functions were changed so they
3711 can now return an error. The RAND changes required a change to the
3712 RAND_METHOD structure.
3716 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3717 a gcc attribute to warn if the result of a function is ignored. This
3718 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3719 whose return value is often ignored.
3723 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3724 These allow SCTs (signed certificate timestamps) to be requested and
3725 validated when establishing a connection.
3727 *Rob Percival <robpercival@google.com>*
3732 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3734 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3735 used even when parsing explicit parameters, when loading a serialized key
3736 or calling `EC_GROUP_new_from_ecpkparameters()`/
3737 `EC_GROUP_new_from_ecparameters()`.
3738 This prevents bypass of security hardening and performance gains,
3739 especially for curves with specialized EC_METHODs.
3740 By default, if a key encoded with explicit parameters is loaded and later
3741 serialized, the output is still encoded with explicit parameters, even if
3742 internally a "named" EC_GROUP is used for computation.
3746 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3747 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3748 NULL. After this change, only the cofactor parameter can be NULL. It also
3749 does some minimal sanity checks on the passed order.
3754 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3755 An attack is simple, if the first CMS_recipientInfo is valid but the
3756 second CMS_recipientInfo is chosen ciphertext. If the second
3757 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3758 encryption key will be replaced by garbage, and the message cannot be
3759 decoded, but if the RSA decryption fails, the correct encryption key is
3760 used and the recipient will not notice the attack.
3761 As a work around for this potential attack the length of the decrypted
3762 key must be equal to the cipher default key length, in case the
3763 certifiate is not given and all recipientInfo are tried out.
3764 The old behaviour can be re-enabled in the CMS code by setting the
3765 CMS_DEBUG_DECRYPT flag.
3770 * Document issue with installation paths in diverse Windows builds
3772 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3773 binaries and run-time config file.
3778 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3780 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3781 This changes the size when using the genpkey app when no size is given. It
3782 fixes an omission in earlier changes that changed all RSA, DSA and DH
3783 generation apps to use 2048 bits by default.
3787 * Add FIPS support for Android Arm 64-bit
3789 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3790 Module in Version 2.0.10. For some reason, the corresponding target
3791 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3792 built with FIPS support on Android Arm 64-bit. This omission has been
3795 *Matthias St. Pierre*
3797 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3799 * 0-byte record padding oracle
3801 If an application encounters a fatal protocol error and then calls
3802 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3803 then OpenSSL can respond differently to the calling application if a 0 byte
3804 record is received with invalid padding compared to if a 0 byte record is
3805 received with an invalid MAC. If the application then behaves differently
3806 based on that in a way that is detectable to the remote peer, then this
3807 amounts to a padding oracle that could be used to decrypt data.
3809 In order for this to be exploitable "non-stitched" ciphersuites must be in
3810 use. Stitched ciphersuites are optimised implementations of certain
3811 commonly used ciphersuites. Also the application must call SSL_shutdown()
3812 twice even if a protocol error has occurred (applications should not do
3813 this but some do anyway).
3815 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3816 Aviram, with additional investigation by Steven Collison and Andrew
3817 Hourselt. It was reported to OpenSSL on 10th December 2018.
3822 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3826 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3828 * Microarchitecture timing vulnerability in ECC scalar multiplication
3830 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3831 shown to be vulnerable to a microarchitecture timing side channel attack.
3832 An attacker with sufficient access to mount local timing attacks during
3833 ECDSA signature generation could recover the private key.
3835 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3836 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3842 * Timing vulnerability in DSA signature generation
3844 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3845 timing side channel attack. An attacker could use variations in the signing
3846 algorithm to recover the private key.
3848 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3853 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3854 Module, accidentally introduced while backporting security fixes from the
3855 development branch and hindering the use of ECC in FIPS mode.
3859 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3861 * Client DoS due to large DH parameter
3863 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3864 malicious server can send a very large prime value to the client. This will
3865 cause the client to spend an unreasonably long period of time generating a
3866 key for this prime resulting in a hang until the client has finished. This
3867 could be exploited in a Denial Of Service attack.
3869 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3874 * Cache timing vulnerability in RSA Key Generation
3876 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3877 a cache timing side channel attack. An attacker with sufficient access to
3878 mount cache timing attacks during the RSA key generation process could
3879 recover the private key.
3881 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3882 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3887 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3888 parameter is no longer accepted, as it leads to a corrupt table. NULL
3889 pem_str is reserved for alias entries only.
3893 * Revert blinding in ECDSA sign and instead make problematic addition
3894 length-invariant. Switch even to fixed-length Montgomery multiplication.
3898 * Change generating and checking of primes so that the error rate of not
3899 being prime depends on the intended use based on the size of the input.
3900 For larger primes this will result in more rounds of Miller-Rabin.
3901 The maximal error rate for primes with more than 1080 bits is lowered
3904 *Kurt Roeckx, Annie Yousar*
3906 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3910 * Add blinding to ECDSA and DSA signatures to protect against side channel
3911 attacks discovered by Keegan Ryan (NCC Group).
3915 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3916 now allow empty (zero character) pass phrases.
3920 * Certificate time validation (X509_cmp_time) enforces stricter
3921 compliance with RFC 5280. Fractional seconds and timezone offsets
3922 are no longer allowed.
3926 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3928 * Constructed ASN.1 types with a recursive definition could exceed the stack
3930 Constructed ASN.1 types with a recursive definition (such as can be found
3931 in PKCS7) could eventually exceed the stack given malicious input with
3932 excessive recursion. This could result in a Denial Of Service attack. There
3933 are no such structures used within SSL/TLS that come from untrusted sources
3934 so this is considered safe.
3936 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3942 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3944 * Read/write after SSL object in error state
3946 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3947 mechanism. The intent was that if a fatal error occurred during a handshake
3948 then OpenSSL would move into the error state and would immediately fail if
3949 you attempted to continue the handshake. This works as designed for the
3950 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3951 SSL_connect()), however due to a bug it does not work correctly if
3952 SSL_read() or SSL_write() is called directly. In that scenario, if the
3953 handshake fails then a fatal error will be returned in the initial function
3954 call. If SSL_read()/SSL_write() is subsequently called by the application
3955 for the same SSL object then it will succeed and the data is passed without
3956 being decrypted/encrypted directly from the SSL/TLS record layer.
3958 In order to exploit this issue an application bug would have to be present
3959 that resulted in a call to SSL_read()/SSL_write() being issued after having
3960 already received a fatal error.
3962 This issue was reported to OpenSSL by David Benjamin (Google).
3967 * rsaz_1024_mul_avx2 overflow bug on x86_64
3969 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3970 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3971 Analysis suggests that attacks against RSA and DSA as a result of this
3972 defect would be very difficult to perform and are not believed likely.
3973 Attacks against DH1024 are considered just feasible, because most of the
3974 work necessary to deduce information about a private key may be performed
3975 offline. The amount of resources required for such an attack would be
3976 significant. However, for an attack on TLS to be meaningful, the server
3977 would have to share the DH1024 private key among multiple clients, which is
3978 no longer an option since CVE-2016-0701.
3980 This only affects processors that support the AVX2 but not ADX extensions
3981 like Intel Haswell (4th generation).
3983 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3984 was originally found via the OSS-Fuzz project.
3989 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
3991 * bn_sqrx8x_internal carry bug on x86_64
3993 There is a carry propagating bug in the x86_64 Montgomery squaring
3994 procedure. No EC algorithms are affected. Analysis suggests that attacks
3995 against RSA and DSA as a result of this defect would be very difficult to
3996 perform and are not believed likely. Attacks against DH are considered just
3997 feasible (although very difficult) because most of the work necessary to
3998 deduce information about a private key may be performed offline. The amount
3999 of resources required for such an attack would be very significant and
4000 likely only accessible to a limited number of attackers. An attacker would
4001 additionally need online access to an unpatched system using the target
4002 private key in a scenario with persistent DH parameters and a private
4003 key that is shared between multiple clients.
4005 This only affects processors that support the BMI1, BMI2 and ADX extensions
4006 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4008 This issue was reported to OpenSSL by the OSS-Fuzz project.
4013 * Malformed X.509 IPAddressFamily could cause OOB read
4015 If an X.509 certificate has a malformed IPAddressFamily extension,
4016 OpenSSL could do a one-byte buffer overread. The most likely result
4017 would be an erroneous display of the certificate in text format.
4019 This issue was reported to OpenSSL by the OSS-Fuzz project.
4024 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4026 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4027 platform rather than 'mingw'.
4031 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4033 * Truncated packet could crash via OOB read
4035 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4036 cipher is being used, then a truncated packet can cause that host to
4037 perform an out-of-bounds read, usually resulting in a crash.
4039 This issue was reported to OpenSSL by Robert Święcki of Google.
4044 * BN_mod_exp may produce incorrect results on x86_64
4046 There is a carry propagating bug in the x86_64 Montgomery squaring
4047 procedure. No EC algorithms are affected. Analysis suggests that attacks
4048 against RSA and DSA as a result of this defect would be very difficult to
4049 perform and are not believed likely. Attacks against DH are considered just
4050 feasible (although very difficult) because most of the work necessary to
4051 deduce information about a private key may be performed offline. The amount
4052 of resources required for such an attack would be very significant and
4053 likely only accessible to a limited number of attackers. An attacker would
4054 additionally need online access to an unpatched system using the target
4055 private key in a scenario with persistent DH parameters and a private
4056 key that is shared between multiple clients. For example this can occur by
4057 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4058 similar to CVE-2015-3193 but must be treated as a separate problem.
4060 This issue was reported to OpenSSL by the OSS-Fuzz project.
4065 * Montgomery multiplication may produce incorrect results
4067 There is a carry propagating bug in the Broadwell-specific Montgomery
4068 multiplication procedure that handles input lengths divisible by, but
4069 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4070 and DH private keys are impossible. This is because the subroutine in
4071 question is not used in operations with the private key itself and an input
4072 of the attacker's direct choice. Otherwise the bug can manifest itself as
4073 transient authentication and key negotiation failures or reproducible
4074 erroneous outcome of public-key operations with specially crafted input.
4075 Among EC algorithms only Brainpool P-512 curves are affected and one
4076 presumably can attack ECDH key negotiation. Impact was not analyzed in
4077 detail, because pre-requisites for attack are considered unlikely. Namely
4078 multiple clients have to choose the curve in question and the server has to
4079 share the private key among them, neither of which is default behaviour.
4080 Even then only clients that chose the curve will be affected.
4082 This issue was publicly reported as transient failures and was not
4083 initially recognized as a security issue. Thanks to Richard Morgan for
4084 providing reproducible case.
4089 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4090 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4091 prevent issues where no progress is being made and the peer continually
4092 sends unrecognised record types, using up resources processing them.
4096 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4098 * Missing CRL sanity check
4100 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4101 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4102 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4104 This issue only affects the OpenSSL 1.0.2i
4109 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4111 * OCSP Status Request extension unbounded memory growth
4113 A malicious client can send an excessively large OCSP Status Request
4114 extension. If that client continually requests renegotiation, sending a
4115 large OCSP Status Request extension each time, then there will be unbounded
4116 memory growth on the server. This will eventually lead to a Denial Of
4117 Service attack through memory exhaustion. Servers with a default
4118 configuration are vulnerable even if they do not support OCSP. Builds using
4119 the "no-ocsp" build time option are not affected.
4121 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4126 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4129 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4135 * OOB write in MDC2_Update()
4137 An overflow can occur in MDC2_Update() either if called directly or
4138 through the EVP_DigestUpdate() function using MDC2. If an attacker
4139 is able to supply very large amounts of input data after a previous
4140 call to EVP_EncryptUpdate() with a partial block then a length check
4141 can overflow resulting in a heap corruption.
4143 The amount of data needed is comparable to SIZE_MAX which is impractical
4146 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4151 * Malformed SHA512 ticket DoS
4153 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4154 DoS attack where a malformed ticket will result in an OOB read which will
4157 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4158 a custom server callback and ticket lookup mechanism.
4160 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4165 * OOB write in BN_bn2dec()
4167 The function BN_bn2dec() does not check the return value of BN_div_word().
4168 This can cause an OOB write if an application uses this function with an
4169 overly large BIGNUM. This could be a problem if an overly large certificate
4170 or CRL is printed out from an untrusted source. TLS is not affected because
4171 record limits will reject an oversized certificate before it is parsed.
4173 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4178 * OOB read in TS_OBJ_print_bio()
4180 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4181 the total length the OID text representation would use and not the amount
4182 of data written. This will result in OOB reads when large OIDs are
4185 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4190 * Pointer arithmetic undefined behaviour
4192 Avoid some undefined pointer arithmetic
4194 A common idiom in the codebase is to check limits in the following manner:
4197 Where "p" points to some malloc'd data of SIZE bytes and
4200 "len" here could be from some externally supplied data (e.g. from a TLS
4203 The rules of C pointer arithmetic are such that "p + len" is only well
4204 defined where len <= SIZE. Therefore the above idiom is actually
4205 undefined behaviour.
4207 For example this could cause problems if some malloc implementation
4208 provides an address for "p" such that "p + len" actually overflows for
4209 values of len that are too big and therefore p + len < limit.
4211 This issue was reported to OpenSSL by Guido Vranken
4216 * Constant time flag not preserved in DSA signing
4218 Operations in the DSA signing algorithm should run in constant time in
4219 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4220 implementation means that a non-constant time codepath is followed for
4221 certain operations. This has been demonstrated through a cache-timing
4222 attack to be sufficient for an attacker to recover the private DSA key.
4224 This issue was reported by César Pereida (Aalto University), Billy Brumley
4225 (Tampere University of Technology), and Yuval Yarom (The University of
4226 Adelaide and NICTA).
4231 * DTLS buffered message DoS
4233 In a DTLS connection where handshake messages are delivered out-of-order
4234 those messages that OpenSSL is not yet ready to process will be buffered
4235 for later use. Under certain circumstances, a flaw in the logic means that
4236 those messages do not get removed from the buffer even though the handshake
4237 has been completed. An attacker could force up to approx. 15 messages to
4238 remain in the buffer when they are no longer required. These messages will
4239 be cleared when the DTLS connection is closed. The default maximum size for
4240 a message is 100k. Therefore the attacker could force an additional 1500k
4241 to be consumed per connection. By opening many simulataneous connections an
4242 attacker could cause a DoS attack through memory exhaustion.
4244 This issue was reported to OpenSSL by Quan Luo.
4249 * DTLS replay protection DoS
4251 A flaw in the DTLS replay attack protection mechanism means that records
4252 that arrive for future epochs update the replay protection "window" before
4253 the MAC for the record has been validated. This could be exploited by an
4254 attacker by sending a record for the next epoch (which does not have to
4255 decrypt or have a valid MAC), with a very large sequence number. This means
4256 that all subsequent legitimate packets are dropped causing a denial of
4257 service for a specific DTLS connection.
4259 This issue was reported to OpenSSL by the OCAP audit team.
4264 * Certificate message OOB reads
4266 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4267 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4268 theoretical DoS risk but this has not been observed in practice on common
4271 The messages affected are client certificate, client certificate request
4272 and server certificate. As a result the attack can only be performed
4273 against a client or a server which enables client authentication.
4275 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4280 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4282 * Prevent padding oracle in AES-NI CBC MAC check
4284 A MITM attacker can use a padding oracle attack to decrypt traffic
4285 when the connection uses an AES CBC cipher and the server support
4288 This issue was introduced as part of the fix for Lucky 13 padding
4289 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4290 constant time by making sure that always the same bytes are read and
4291 compared against either the MAC or padding bytes. But it no longer
4292 checked that there was enough data to have both the MAC and padding
4295 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4300 * Fix EVP_EncodeUpdate overflow
4302 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4303 Base64 encoding of binary data. If an attacker is able to supply very large
4304 amounts of input data then a length check can overflow resulting in a heap
4307 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4308 the `PEM_write_bio*` family of functions. These are mainly used within the
4309 OpenSSL command line applications, so any application which processes data
4310 from an untrusted source and outputs it as a PEM file should be considered
4311 vulnerable to this issue. User applications that call these APIs directly
4312 with large amounts of untrusted data may also be vulnerable.
4314 This issue was reported by Guido Vranken.
4319 * Fix EVP_EncryptUpdate overflow
4321 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4322 is able to supply very large amounts of input data after a previous call to
4323 EVP_EncryptUpdate() with a partial block then a length check can overflow
4324 resulting in a heap corruption. Following an analysis of all OpenSSL
4325 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4326 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4327 the first called function after an EVP_EncryptInit(), and therefore that
4328 specific call must be safe. The second form is where the length passed to
4329 EVP_EncryptUpdate() can be seen from the code to be some small value and
4330 therefore there is no possibility of an overflow. Since all instances are
4331 one of these two forms, it is believed that there can be no overflows in
4332 internal code due to this problem. It should be noted that
4333 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4334 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4335 of these calls have also been analysed too and it is believed there are no
4336 instances in internal usage where an overflow could occur.
4338 This issue was reported by Guido Vranken.
4343 * Prevent ASN.1 BIO excessive memory allocation
4345 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4346 a short invalid encoding can cause allocation of large amounts of memory
4347 potentially consuming excessive resources or exhausting memory.
4349 Any application parsing untrusted data through d2i BIO functions is
4350 affected. The memory based functions such as d2i_X509() are *not* affected.
4351 Since the memory based functions are used by the TLS library, TLS
4352 applications are not affected.
4354 This issue was reported by Brian Carpenter.
4361 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4362 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4363 in arbitrary stack data being returned in the buffer.
4365 This issue was reported by Guido Vranken.
4370 * Modify behavior of ALPN to invoke callback after SNI/servername
4371 callback, such that updates to the SSL_CTX affect ALPN.
4375 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4380 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4381 methods are enabled and ssl2 is disabled the methods return NULL.
4385 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4387 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4388 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4389 provide any "EXPORT" or "LOW" strength ciphers.
4393 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4394 is by default disabled at build-time. Builds that are not configured with
4395 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4396 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4397 will need to explicitly call either of:
4399 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4401 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4403 as appropriate. Even if either of those is used, or the application
4404 explicitly uses the version-specific SSLv2_method() or its client and
4405 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4406 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4407 ciphers, and SSLv2 56-bit DES are no longer available.
4412 * Fix a double-free in DSA code
4414 A double free bug was discovered when OpenSSL parses malformed DSA private
4415 keys and could lead to a DoS attack or memory corruption for applications
4416 that receive DSA private keys from untrusted sources. This scenario is
4419 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4425 * Disable SRP fake user seed to address a server memory leak.
4427 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4429 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4430 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4431 was changed to ignore the "fake user" SRP seed, even if the seed
4434 Users should use SRP_VBASE_get1_by_user instead. Note that in
4435 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4436 also that even though configuring the SRP seed attempts to hide
4437 invalid usernames by continuing the handshake with fake
4438 credentials, this behaviour is not constant time and no strong
4439 guarantees are made that the handshake is indistinguishable from
4440 that of a valid user.
4445 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4447 In the BN_hex2bn function the number of hex digits is calculated using an
4448 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4449 large values of |i| this can result in |bn_expand| not allocating any
4450 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4451 field as NULL leading to a subsequent NULL ptr deref. For very large values
4452 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4453 In this case memory is allocated to the internal BIGNUM data field, but it
4454 is insufficiently sized leading to heap corruption. A similar issue exists
4455 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4456 is ever called by user applications with very large untrusted hex/dec data.
4457 This is anticipated to be a rare occurrence.
4459 All OpenSSL internal usage of these functions use data that is not expected
4460 to be untrusted, e.g. config file data or application command line
4461 arguments. If user developed applications generate config file data based
4462 on untrusted data then it is possible that this could also lead to security
4463 consequences. This is also anticipated to be rare.
4465 This issue was reported to OpenSSL by Guido Vranken.
4470 * Fix memory issues in `BIO_*printf` functions
4472 The internal |fmtstr| function used in processing a "%s" format string in
4473 the `BIO_*printf` functions could overflow while calculating the length of a
4474 string and cause an OOB read when printing very long strings.
4476 Additionally the internal |doapr_outch| function can attempt to write to an
4477 OOB memory location (at an offset from the NULL pointer) in the event of a
4478 memory allocation failure. In 1.0.2 and below this could be caused where
4479 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4480 could be in processing a very long "%s" format string. Memory leaks can
4483 The first issue may mask the second issue dependent on compiler behaviour.
4484 These problems could enable attacks where large amounts of untrusted data
4485 is passed to the `BIO_*printf` functions. If applications use these functions
4486 in this way then they could be vulnerable. OpenSSL itself uses these
4487 functions when printing out human-readable dumps of ASN.1 data. Therefore
4488 applications that print this data could be vulnerable if the data is from
4489 untrusted sources. OpenSSL command line applications could also be
4490 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4491 as command line arguments.
4493 Libssl is not considered directly vulnerable. Additionally certificates etc
4494 received via remote connections via libssl are also unlikely to be able to
4495 trigger these issues because of message size limits enforced within libssl.
4497 This issue was reported to OpenSSL Guido Vranken.
4502 * Side channel attack on modular exponentiation
4504 A side-channel attack was found which makes use of cache-bank conflicts on
4505 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4506 of RSA keys. The ability to exploit this issue is limited as it relies on
4507 an attacker who has control of code in a thread running on the same
4508 hyper-threaded core as the victim thread which is performing decryptions.
4510 This issue was reported to OpenSSL by Yuval Yarom, The University of
4511 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4512 Nadia Heninger, University of Pennsylvania with more information at
4513 <http://cachebleed.info>.
4518 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4519 if no keysize is specified with default_bits. This fixes an
4520 omission in an earlier change that changed all RSA/DSA key generation
4521 apps to use 2048 bits by default.
4525 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4527 * DH small subgroups
4529 Historically OpenSSL only ever generated DH parameters based on "safe"
4530 primes. More recently (in version 1.0.2) support was provided for
4531 generating X9.42 style parameter files such as those required for RFC 5114
4532 support. The primes used in such files may not be "safe". Where an
4533 application is using DH configured with parameters based on primes that are
4534 not "safe" then an attacker could use this fact to find a peer's private
4535 DH exponent. This attack requires that the attacker complete multiple
4536 handshakes in which the peer uses the same private DH exponent. For example
4537 this could be used to discover a TLS server's private DH exponent if it's
4538 reusing the private DH exponent or it's using a static DH ciphersuite.
4540 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4541 TLS. It is not on by default. If the option is not set then the server
4542 reuses the same private DH exponent for the life of the server process and
4543 would be vulnerable to this attack. It is believed that many popular
4544 applications do set this option and would therefore not be at risk.
4546 The fix for this issue adds an additional check where a "q" parameter is
4547 available (as is the case in X9.42 based parameters). This detects the
4548 only known attack, and is the only possible defense for static DH
4549 ciphersuites. This could have some performance impact.
4551 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4552 default and cannot be disabled. This could have some performance impact.
4554 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4559 * SSLv2 doesn't block disabled ciphers
4561 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4562 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4563 been disabled, provided that the SSLv2 protocol was not also disabled via
4566 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4567 and Sebastian Schinzel.
4572 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4574 * BN_mod_exp may produce incorrect results on x86_64
4576 There is a carry propagating bug in the x86_64 Montgomery squaring
4577 procedure. No EC algorithms are affected. Analysis suggests that attacks
4578 against RSA and DSA as a result of this defect would be very difficult to
4579 perform and are not believed likely. Attacks against DH are considered just
4580 feasible (although very difficult) because most of the work necessary to
4581 deduce information about a private key may be performed offline. The amount
4582 of resources required for such an attack would be very significant and
4583 likely only accessible to a limited number of attackers. An attacker would
4584 additionally need online access to an unpatched system using the target
4585 private key in a scenario with persistent DH parameters and a private
4586 key that is shared between multiple clients. For example this can occur by
4587 default in OpenSSL DHE based SSL/TLS ciphersuites.
4589 This issue was reported to OpenSSL by Hanno Böck.
4594 * Certificate verify crash with missing PSS parameter
4596 The signature verification routines will crash with a NULL pointer
4597 dereference if presented with an ASN.1 signature using the RSA PSS
4598 algorithm and absent mask generation function parameter. Since these
4599 routines are used to verify certificate signature algorithms this can be
4600 used to crash any certificate verification operation and exploited in a
4601 DoS attack. Any application which performs certificate verification is
4602 vulnerable including OpenSSL clients and servers which enable client
4605 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4610 * X509_ATTRIBUTE memory leak
4612 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4613 memory. This structure is used by the PKCS#7 and CMS routines so any
4614 application which reads PKCS#7 or CMS data from untrusted sources is
4615 affected. SSL/TLS is not affected.
4617 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4623 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4624 This changes the decoding behaviour for some invalid messages,
4625 though the change is mostly in the more lenient direction, and
4626 legacy behaviour is preserved as much as possible.
4630 * In DSA_generate_parameters_ex, if the provided seed is too short,
4633 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4635 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4637 * Alternate chains certificate forgery
4639 During certificate verification, OpenSSL will attempt to find an
4640 alternative certificate chain if the first attempt to build such a chain
4641 fails. An error in the implementation of this logic can mean that an
4642 attacker could cause certain checks on untrusted certificates to be
4643 bypassed, such as the CA flag, enabling them to use a valid leaf
4644 certificate to act as a CA and "issue" an invalid certificate.
4646 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4651 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4653 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4654 incompatibility in the handling of HMAC. The previous ABI has now been
4659 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4661 * Malformed ECParameters causes infinite loop
4663 When processing an ECParameters structure OpenSSL enters an infinite loop
4664 if the curve specified is over a specially malformed binary polynomial
4667 This can be used to perform denial of service against any
4668 system which processes public keys, certificate requests or
4669 certificates. This includes TLS clients and TLS servers with
4670 client authentication enabled.
4672 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4677 * Exploitable out-of-bounds read in X509_cmp_time
4679 X509_cmp_time does not properly check the length of the ASN1_TIME
4680 string and can read a few bytes out of bounds. In addition,
4681 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4684 An attacker can use this to craft malformed certificates and CRLs of
4685 various sizes and potentially cause a segmentation fault, resulting in
4686 a DoS on applications that verify certificates or CRLs. TLS clients
4687 that verify CRLs are affected. TLS clients and servers with client
4688 authentication enabled may be affected if they use custom verification
4691 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4692 independently by Hanno Böck.
4697 * PKCS7 crash with missing EnvelopedContent
4699 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4700 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4701 with missing content and trigger a NULL pointer dereference on parsing.
4703 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4704 structures from untrusted sources are affected. OpenSSL clients and
4705 servers are not affected.
4707 This issue was reported to OpenSSL by Michal Zalewski (Google).
4712 * CMS verify infinite loop with unknown hash function
4714 When verifying a signedData message the CMS code can enter an infinite loop
4715 if presented with an unknown hash function OID. This can be used to perform
4716 denial of service against any system which verifies signedData messages using
4718 This issue was reported to OpenSSL by Johannes Bauer.
4723 * Race condition handling NewSessionTicket
4725 If a NewSessionTicket is received by a multi-threaded client when attempting to
4726 reuse a previous ticket then a race condition can occur potentially leading to
4727 a double free of the ticket data.
4732 * Only support 256-bit or stronger elliptic curves with the
4733 'ecdh_auto' setting (server) or by default (client). Of supported
4734 curves, prefer P-256 (both).
4738 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4740 * ClientHello sigalgs DoS fix
4742 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4743 invalid signature algorithms extension a NULL pointer dereference will
4744 occur. This can be exploited in a DoS attack against the server.
4746 This issue was was reported to OpenSSL by David Ramos of Stanford
4750 *Stephen Henson and Matt Caswell*
4752 * Multiblock corrupted pointer fix
4754 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4755 feature only applies on 64 bit x86 architecture platforms that support AES
4756 NI instructions. A defect in the implementation of "multiblock" can cause
4757 OpenSSL's internal write buffer to become incorrectly set to NULL when
4758 using non-blocking IO. Typically, when the user application is using a
4759 socket BIO for writing, this will only result in a failed connection.
4760 However if some other BIO is used then it is likely that a segmentation
4761 fault will be triggered, thus enabling a potential DoS attack.
4763 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4768 * Segmentation fault in DTLSv1_listen fix
4770 The DTLSv1_listen function is intended to be stateless and processes the
4771 initial ClientHello from many peers. It is common for user code to loop
4772 over the call to DTLSv1_listen until a valid ClientHello is received with
4773 an associated cookie. A defect in the implementation of DTLSv1_listen means
4774 that state is preserved in the SSL object from one invocation to the next
4775 that can lead to a segmentation fault. Errors processing the initial
4776 ClientHello can trigger this scenario. An example of such an error could be
4777 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4780 This issue was reported to OpenSSL by Per Allansson.
4785 * Segmentation fault in ASN1_TYPE_cmp fix
4787 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4788 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4789 certificate signature algorithm consistency this can be used to crash any
4790 certificate verification operation and exploited in a DoS attack. Any
4791 application which performs certificate verification is vulnerable including
4792 OpenSSL clients and servers which enable client authentication.
4797 * Segmentation fault for invalid PSS parameters fix
4799 The signature verification routines will crash with a NULL pointer
4800 dereference if presented with an ASN.1 signature using the RSA PSS
4801 algorithm and invalid parameters. Since these routines are used to verify
4802 certificate signature algorithms this can be used to crash any
4803 certificate verification operation and exploited in a DoS attack. Any
4804 application which performs certificate verification is vulnerable including
4805 OpenSSL clients and servers which enable client authentication.
4807 This issue was was reported to OpenSSL by Brian Carpenter.
4812 * ASN.1 structure reuse memory corruption fix
4814 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4815 memory corruption via an invalid write. Such reuse is and has been
4816 strongly discouraged and is believed to be rare.
4818 Applications that parse structures containing CHOICE or ANY DEFINED BY
4819 components may be affected. Certificate parsing (d2i_X509 and related
4820 functions) are however not affected. OpenSSL clients and servers are
4826 * PKCS7 NULL pointer dereferences fix
4828 The PKCS#7 parsing code does not handle missing outer ContentInfo
4829 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4830 missing content and trigger a NULL pointer dereference on parsing.
4832 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4833 otherwise parse PKCS#7 structures from untrusted sources are
4834 affected. OpenSSL clients and servers are not affected.
4836 This issue was reported to OpenSSL by Michal Zalewski (Google).
4841 * DoS via reachable assert in SSLv2 servers fix
4843 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4844 servers that both support SSLv2 and enable export cipher suites by sending
4845 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4847 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4848 (OpenSSL development team).
4853 * Empty CKE with client auth and DHE fix
4855 If client auth is used then a server can seg fault in the event of a DHE
4856 ciphersuite being selected and a zero length ClientKeyExchange message
4857 being sent by the client. This could be exploited in a DoS attack.
4862 * Handshake with unseeded PRNG fix
4864 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4865 with an unseeded PRNG. The conditions are:
4866 - The client is on a platform where the PRNG has not been seeded
4867 automatically, and the user has not seeded manually
4868 - A protocol specific client method version has been used (i.e. not
4869 SSL_client_methodv23)
4870 - A ciphersuite is used that does not require additional random data from
4871 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4873 If the handshake succeeds then the client random that has been used will
4874 have been generated from a PRNG with insufficient entropy and therefore the
4875 output may be predictable.
4877 For example using the following command with an unseeded openssl will
4878 succeed on an unpatched platform:
4880 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4885 * Use After Free following d2i_ECPrivatekey error fix
4887 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4888 could cause a use after free condition. This, in turn, could cause a double
4889 free in several private key parsing functions (such as d2i_PrivateKey
4890 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4891 for applications that receive EC private keys from untrusted
4892 sources. This scenario is considered rare.
4894 This issue was discovered by the BoringSSL project and fixed in their
4900 * X509_to_X509_REQ NULL pointer deref fix
4902 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4903 the certificate key is invalid. This function is rarely used in practice.
4905 This issue was discovered by Brian Carpenter.
4910 * Removed the export ciphers from the DEFAULT ciphers
4914 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4916 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4917 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4918 So far those who have to target multiple platforms would compromise
4919 and argue that binary targeting say ARMv5 would still execute on
4920 ARMv8. "Universal" build resolves this compromise by providing
4921 near-optimal performance even on newer platforms.
4925 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4926 (other platforms pending).
4928 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4930 * Add support for the SignedCertificateTimestampList certificate and
4931 OCSP response extensions from RFC6962.
4935 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4936 for corner cases. (Certain input points at infinity could lead to
4937 bogus results, with non-infinity inputs mapped to infinity too.)
4941 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4942 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4943 common cases are optimized and there still is room for further
4944 improvements. Vector Permutation AES for Altivec is also added.
4948 * Add support for little-endian ppc64 Linux target.
4950 *Marcelo Cerri (IBM)*
4952 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4953 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4954 are optimized and there still is room for further improvements.
4955 Both 32- and 64-bit modes are supported.
4957 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4959 * Improved ARMv7 NEON support.
4963 * Support for SPARC Architecture 2011 crypto extensions, first
4964 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4965 SHA256/512, MD5, GHASH and modular exponentiation.
4967 *Andy Polyakov, David Miller*
4969 * Accelerated modular exponentiation for Intel processors, a.k.a.
4972 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4974 * Support for new and upcoming Intel processors, including AVX2,
4975 BMI and SHA ISA extensions. This includes additional "stitched"
4976 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4979 This work was sponsored by Intel Corp.
4983 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4984 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4985 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4989 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4990 this fixes a limitation in previous versions of OpenSSL.
4994 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4995 MGF1 digest and OAEP label.
4999 * Add EVP support for key wrapping algorithms, to avoid problems with
5000 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5001 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5002 algorithms and include tests cases.
5006 * Add functions to allocate and set the fields of an ECDSA_METHOD
5009 *Douglas E. Engert, Steve Henson*
5011 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5012 difference in days and seconds between two tm or ASN1_TIME structures.
5016 * Add -rev test option to s_server to just reverse order of characters
5017 received by client and send back to server. Also prints an abbreviated
5018 summary of the connection parameters.
5022 * New option -brief for s_client and s_server to print out a brief summary
5023 of connection parameters.
5027 * Add callbacks for arbitrary TLS extensions.
5029 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5031 * New option -crl_download in several openssl utilities to download CRLs
5032 from CRLDP extension in certificates.
5036 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5040 * New function X509_CRL_diff to generate a delta CRL from the difference
5041 of two full CRLs. Add support to "crl" utility.
5045 * New functions to set lookup_crls function and to retrieve
5046 X509_STORE from X509_STORE_CTX.
5050 * Print out deprecated issuer and subject unique ID fields in
5055 * Extend OCSP I/O functions so they can be used for simple general purpose
5056 HTTP as well as OCSP. New wrapper function which can be used to download
5057 CRLs using the OCSP API.
5061 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5065 * `SSL_CONF*` functions. These provide a common framework for application
5066 configuration using configuration files or command lines.
5070 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5071 message callback and prints the results. Needs compile time option
5072 "enable-ssl-trace". New options to s_client and s_server to enable
5077 * New ctrl and macro to retrieve supported points extensions.
5078 Print out extension in s_server and s_client.
5082 * New functions to retrieve certificate signature and signature
5087 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5092 * New Suite B modes for TLS code. These use and enforce the requirements
5093 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5094 only use Suite B curves. The Suite B modes can be set by using the
5095 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5099 * New chain verification flags for Suite B levels of security. Check
5100 algorithms are acceptable when flags are set in X509_verify_cert.
5104 * Make tls1_check_chain return a set of flags indicating checks passed
5105 by a certificate chain. Add additional tests to handle client
5106 certificates: checks for matching certificate type and issuer name
5111 * If an attempt is made to use a signature algorithm not in the peer
5112 preference list abort the handshake. If client has no suitable
5113 signature algorithms in response to a certificate request do not
5114 use the certificate.
5118 * If server EC tmp key is not in client preference list abort handshake.
5122 * Add support for certificate stores in CERT structure. This makes it
5123 possible to have different stores per SSL structure or one store in
5124 the parent SSL_CTX. Include distinct stores for certificate chain
5125 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5126 to build and store a certificate chain in CERT structure: returning
5127 an error if the chain cannot be built: this will allow applications
5128 to test if a chain is correctly configured.
5130 Note: if the CERT based stores are not set then the parent SSL_CTX
5131 store is used to retain compatibility with existing behaviour.
5135 * New function ssl_set_client_disabled to set a ciphersuite disabled
5136 mask based on the current session, check mask when sending client
5137 hello and checking the requested ciphersuite.
5141 * New ctrls to retrieve and set certificate types in a certificate
5142 request message. Print out received values in s_client. If certificate
5143 types is not set with custom values set sensible values based on
5144 supported signature algorithms.
5148 * Support for distinct client and server supported signature algorithms.
5152 * Add certificate callback. If set this is called whenever a certificate
5153 is required by client or server. An application can decide which
5154 certificate chain to present based on arbitrary criteria: for example
5155 supported signature algorithms. Add very simple example to s_server.
5156 This fixes many of the problems and restrictions of the existing client
5157 certificate callback: for example you can now clear an existing
5158 certificate and specify the whole chain.
5162 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5163 the certificate can be used for (if anything). Set valid_flags field
5164 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5165 to have similar checks in it.
5167 Add new "cert_flags" field to CERT structure and include a "strict mode".
5168 This enforces some TLS certificate requirements (such as only permitting
5169 certificate signature algorithms contained in the supported algorithms
5170 extension) which some implementations ignore: this option should be used
5171 with caution as it could cause interoperability issues.
5175 * Update and tidy signature algorithm extension processing. Work out
5176 shared signature algorithms based on preferences and peer algorithms
5177 and print them out in s_client and s_server. Abort handshake if no
5178 shared signature algorithms.
5182 * Add new functions to allow customised supported signature algorithms
5183 for SSL and SSL_CTX structures. Add options to s_client and s_server
5188 * New function SSL_certs_clear() to delete all references to certificates
5189 from an SSL structure. Before this once a certificate had been added
5190 it couldn't be removed.
5194 * Integrate hostname, email address and IP address checking with certificate
5195 verification. New verify options supporting checking in openssl utility.
5199 * Fixes and wildcard matching support to hostname and email checking
5200 functions. Add manual page.
5202 *Florian Weimer (Red Hat Product Security Team)*
5204 * New functions to check a hostname email or IP address against a
5205 certificate. Add options x509 utility to print results of checks against
5210 * Fix OCSP checking.
5212 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5214 * Initial experimental support for explicitly trusted non-root CAs.
5215 OpenSSL still tries to build a complete chain to a root but if an
5216 intermediate CA has a trust setting included that is used. The first
5217 setting is used: whether to trust (e.g., -addtrust option to the x509
5222 * Add -trusted_first option which attempts to find certificates in the
5223 trusted store even if an untrusted chain is also supplied.
5227 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5228 platform support for Linux and Android.
5232 * Support for linux-x32, ILP32 environment in x86_64 framework.
5236 * Experimental multi-implementation support for FIPS capable OpenSSL.
5237 When in FIPS mode the approved implementations are used as normal,
5238 when not in FIPS mode the internal unapproved versions are used instead.
5239 This means that the FIPS capable OpenSSL isn't forced to use the
5240 (often lower performance) FIPS implementations outside FIPS mode.
5244 * Transparently support X9.42 DH parameters when calling
5245 PEM_read_bio_DHparameters. This means existing applications can handle
5246 the new parameter format automatically.
5250 * Initial experimental support for X9.42 DH parameter format: mainly
5251 to support use of 'q' parameter for RFC5114 parameters.
5255 * Add DH parameters from RFC5114 including test data to dhtest.
5259 * Support for automatic EC temporary key parameter selection. If enabled
5260 the most preferred EC parameters are automatically used instead of
5261 hardcoded fixed parameters. Now a server just has to call:
5262 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5263 support ECDH and use the most appropriate parameters.
5267 * Enhance and tidy EC curve and point format TLS extension code. Use
5268 static structures instead of allocation if default values are used.
5269 New ctrls to set curves we wish to support and to retrieve shared curves.
5270 Print out shared curves in s_server. New options to s_server and s_client
5271 to set list of supported curves.
5275 * New ctrls to retrieve supported signature algorithms and
5276 supported curve values as an array of NIDs. Extend openssl utility
5277 to print out received values.
5281 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5282 between NIDs and the more common NIST names such as "P-256". Enhance
5283 ecparam utility and ECC method to recognise the NIST names for curves.
5287 * Enhance SSL/TLS certificate chain handling to support different
5288 chains for each certificate instead of one chain in the parent SSL_CTX.
5292 * Support for fixed DH ciphersuite client authentication: where both
5293 server and client use DH certificates with common parameters.
5297 * Support for fixed DH ciphersuites: those requiring DH server
5302 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5304 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5305 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5306 X509_CINF_get_signature were reverted post internal team review.
5311 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5313 * OCSP Status Request extension unbounded memory growth
5315 A malicious client can send an excessively large OCSP Status Request
5316 extension. If that client continually requests renegotiation, sending a
5317 large OCSP Status Request extension each time, then there will be unbounded
5318 memory growth on the server. This will eventually lead to a Denial Of
5319 Service attack through memory exhaustion. Servers with a default
5320 configuration are vulnerable even if they do not support OCSP. Builds using
5321 the "no-ocsp" build time option are not affected.
5323 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5328 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5331 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5337 * OOB write in MDC2_Update()
5339 An overflow can occur in MDC2_Update() either if called directly or
5340 through the EVP_DigestUpdate() function using MDC2. If an attacker
5341 is able to supply very large amounts of input data after a previous
5342 call to EVP_EncryptUpdate() with a partial block then a length check
5343 can overflow resulting in a heap corruption.
5345 The amount of data needed is comparable to SIZE_MAX which is impractical
5348 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5353 * Malformed SHA512 ticket DoS
5355 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5356 DoS attack where a malformed ticket will result in an OOB read which will
5359 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5360 a custom server callback and ticket lookup mechanism.
5362 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5367 * OOB write in BN_bn2dec()
5369 The function BN_bn2dec() does not check the return value of BN_div_word().
5370 This can cause an OOB write if an application uses this function with an
5371 overly large BIGNUM. This could be a problem if an overly large certificate
5372 or CRL is printed out from an untrusted source. TLS is not affected because
5373 record limits will reject an oversized certificate before it is parsed.
5375 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5380 * OOB read in TS_OBJ_print_bio()
5382 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5383 the total length the OID text representation would use and not the amount
5384 of data written. This will result in OOB reads when large OIDs are
5387 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5392 * Pointer arithmetic undefined behaviour
5394 Avoid some undefined pointer arithmetic
5396 A common idiom in the codebase is to check limits in the following manner:
5399 Where "p" points to some malloc'd data of SIZE bytes and
5402 "len" here could be from some externally supplied data (e.g. from a TLS
5405 The rules of C pointer arithmetic are such that "p + len" is only well
5406 defined where len <= SIZE. Therefore the above idiom is actually
5407 undefined behaviour.
5409 For example this could cause problems if some malloc implementation
5410 provides an address for "p" such that "p + len" actually overflows for
5411 values of len that are too big and therefore p + len < limit.
5413 This issue was reported to OpenSSL by Guido Vranken
5418 * Constant time flag not preserved in DSA signing
5420 Operations in the DSA signing algorithm should run in constant time in
5421 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5422 implementation means that a non-constant time codepath is followed for
5423 certain operations. This has been demonstrated through a cache-timing
5424 attack to be sufficient for an attacker to recover the private DSA key.
5426 This issue was reported by César Pereida (Aalto University), Billy Brumley
5427 (Tampere University of Technology), and Yuval Yarom (The University of
5428 Adelaide and NICTA).
5433 * DTLS buffered message DoS
5435 In a DTLS connection where handshake messages are delivered out-of-order
5436 those messages that OpenSSL is not yet ready to process will be buffered
5437 for later use. Under certain circumstances, a flaw in the logic means that
5438 those messages do not get removed from the buffer even though the handshake
5439 has been completed. An attacker could force up to approx. 15 messages to
5440 remain in the buffer when they are no longer required. These messages will
5441 be cleared when the DTLS connection is closed. The default maximum size for
5442 a message is 100k. Therefore the attacker could force an additional 1500k
5443 to be consumed per connection. By opening many simulataneous connections an
5444 attacker could cause a DoS attack through memory exhaustion.
5446 This issue was reported to OpenSSL by Quan Luo.
5451 * DTLS replay protection DoS
5453 A flaw in the DTLS replay attack protection mechanism means that records
5454 that arrive for future epochs update the replay protection "window" before
5455 the MAC for the record has been validated. This could be exploited by an
5456 attacker by sending a record for the next epoch (which does not have to
5457 decrypt or have a valid MAC), with a very large sequence number. This means
5458 that all subsequent legitimate packets are dropped causing a denial of
5459 service for a specific DTLS connection.
5461 This issue was reported to OpenSSL by the OCAP audit team.
5466 * Certificate message OOB reads
5468 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5469 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5470 theoretical DoS risk but this has not been observed in practice on common
5473 The messages affected are client certificate, client certificate request
5474 and server certificate. As a result the attack can only be performed
5475 against a client or a server which enables client authentication.
5477 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5482 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5484 * Prevent padding oracle in AES-NI CBC MAC check
5486 A MITM attacker can use a padding oracle attack to decrypt traffic
5487 when the connection uses an AES CBC cipher and the server support
5490 This issue was introduced as part of the fix for Lucky 13 padding
5491 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5492 constant time by making sure that always the same bytes are read and
5493 compared against either the MAC or padding bytes. But it no longer
5494 checked that there was enough data to have both the MAC and padding
5497 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5502 * Fix EVP_EncodeUpdate overflow
5504 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5505 Base64 encoding of binary data. If an attacker is able to supply very large
5506 amounts of input data then a length check can overflow resulting in a heap
5509 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5510 the `PEM_write_bio*` family of functions. These are mainly used within the
5511 OpenSSL command line applications, so any application which processes data
5512 from an untrusted source and outputs it as a PEM file should be considered
5513 vulnerable to this issue. User applications that call these APIs directly
5514 with large amounts of untrusted data may also be vulnerable.
5516 This issue was reported by Guido Vranken.
5521 * Fix EVP_EncryptUpdate overflow
5523 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5524 is able to supply very large amounts of input data after a previous call to
5525 EVP_EncryptUpdate() with a partial block then a length check can overflow
5526 resulting in a heap corruption. Following an analysis of all OpenSSL
5527 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5528 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5529 the first called function after an EVP_EncryptInit(), and therefore that
5530 specific call must be safe. The second form is where the length passed to
5531 EVP_EncryptUpdate() can be seen from the code to be some small value and
5532 therefore there is no possibility of an overflow. Since all instances are
5533 one of these two forms, it is believed that there can be no overflows in
5534 internal code due to this problem. It should be noted that
5535 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5536 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5537 of these calls have also been analysed too and it is believed there are no
5538 instances in internal usage where an overflow could occur.
5540 This issue was reported by Guido Vranken.
5545 * Prevent ASN.1 BIO excessive memory allocation
5547 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5548 a short invalid encoding can casuse allocation of large amounts of memory
5549 potentially consuming excessive resources or exhausting memory.
5551 Any application parsing untrusted data through d2i BIO functions is
5552 affected. The memory based functions such as d2i_X509() are *not* affected.
5553 Since the memory based functions are used by the TLS library, TLS
5554 applications are not affected.
5556 This issue was reported by Brian Carpenter.
5563 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5564 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5565 in arbitrary stack data being returned in the buffer.
5567 This issue was reported by Guido Vranken.
5572 * Modify behavior of ALPN to invoke callback after SNI/servername
5573 callback, such that updates to the SSL_CTX affect ALPN.
5577 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5582 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5583 methods are enabled and ssl2 is disabled the methods return NULL.
5587 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5589 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5590 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5591 provide any "EXPORT" or "LOW" strength ciphers.
5595 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5596 is by default disabled at build-time. Builds that are not configured with
5597 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5598 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5599 will need to explicitly call either of:
5601 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5603 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5605 as appropriate. Even if either of those is used, or the application
5606 explicitly uses the version-specific SSLv2_method() or its client and
5607 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5608 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5609 ciphers, and SSLv2 56-bit DES are no longer available.
5614 * Fix a double-free in DSA code
5616 A double free bug was discovered when OpenSSL parses malformed DSA private
5617 keys and could lead to a DoS attack or memory corruption for applications
5618 that receive DSA private keys from untrusted sources. This scenario is
5621 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5627 * Disable SRP fake user seed to address a server memory leak.
5629 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5631 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5632 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5633 was changed to ignore the "fake user" SRP seed, even if the seed
5636 Users should use SRP_VBASE_get1_by_user instead. Note that in
5637 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5638 also that even though configuring the SRP seed attempts to hide
5639 invalid usernames by continuing the handshake with fake
5640 credentials, this behaviour is not constant time and no strong
5641 guarantees are made that the handshake is indistinguishable from
5642 that of a valid user.
5647 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5649 In the BN_hex2bn function the number of hex digits is calculated using an
5650 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5651 large values of |i| this can result in |bn_expand| not allocating any
5652 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5653 field as NULL leading to a subsequent NULL ptr deref. For very large values
5654 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5655 In this case memory is allocated to the internal BIGNUM data field, but it
5656 is insufficiently sized leading to heap corruption. A similar issue exists
5657 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5658 is ever called by user applications with very large untrusted hex/dec data.
5659 This is anticipated to be a rare occurrence.
5661 All OpenSSL internal usage of these functions use data that is not expected
5662 to be untrusted, e.g. config file data or application command line
5663 arguments. If user developed applications generate config file data based
5664 on untrusted data then it is possible that this could also lead to security
5665 consequences. This is also anticipated to be rare.
5667 This issue was reported to OpenSSL by Guido Vranken.
5672 * Fix memory issues in `BIO_*printf` functions
5674 The internal |fmtstr| function used in processing a "%s" format string in
5675 the `BIO_*printf` functions could overflow while calculating the length of a
5676 string and cause an OOB read when printing very long strings.
5678 Additionally the internal |doapr_outch| function can attempt to write to an
5679 OOB memory location (at an offset from the NULL pointer) in the event of a
5680 memory allocation failure. In 1.0.2 and below this could be caused where
5681 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5682 could be in processing a very long "%s" format string. Memory leaks can
5685 The first issue may mask the second issue dependent on compiler behaviour.
5686 These problems could enable attacks where large amounts of untrusted data
5687 is passed to the `BIO_*printf` functions. If applications use these functions
5688 in this way then they could be vulnerable. OpenSSL itself uses these
5689 functions when printing out human-readable dumps of ASN.1 data. Therefore
5690 applications that print this data could be vulnerable if the data is from
5691 untrusted sources. OpenSSL command line applications could also be
5692 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5693 as command line arguments.
5695 Libssl is not considered directly vulnerable. Additionally certificates etc
5696 received via remote connections via libssl are also unlikely to be able to
5697 trigger these issues because of message size limits enforced within libssl.
5699 This issue was reported to OpenSSL Guido Vranken.
5704 * Side channel attack on modular exponentiation
5706 A side-channel attack was found which makes use of cache-bank conflicts on
5707 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5708 of RSA keys. The ability to exploit this issue is limited as it relies on
5709 an attacker who has control of code in a thread running on the same
5710 hyper-threaded core as the victim thread which is performing decryptions.
5712 This issue was reported to OpenSSL by Yuval Yarom, The University of
5713 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5714 Nadia Heninger, University of Pennsylvania with more information at
5715 <http://cachebleed.info>.
5720 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5721 if no keysize is specified with default_bits. This fixes an
5722 omission in an earlier change that changed all RSA/DSA key generation
5723 apps to use 2048 bits by default.
5727 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5729 * Protection for DH small subgroup attacks
5731 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5732 switched on by default and cannot be disabled. This could have some
5737 * SSLv2 doesn't block disabled ciphers
5739 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5740 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5741 been disabled, provided that the SSLv2 protocol was not also disabled via
5744 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5745 and Sebastian Schinzel.
5750 * Reject DH handshakes with parameters shorter than 1024 bits.
5754 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5756 * Certificate verify crash with missing PSS parameter
5758 The signature verification routines will crash with a NULL pointer
5759 dereference if presented with an ASN.1 signature using the RSA PSS
5760 algorithm and absent mask generation function parameter. Since these
5761 routines are used to verify certificate signature algorithms this can be
5762 used to crash any certificate verification operation and exploited in a
5763 DoS attack. Any application which performs certificate verification is
5764 vulnerable including OpenSSL clients and servers which enable client
5767 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5772 * X509_ATTRIBUTE memory leak
5774 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5775 memory. This structure is used by the PKCS#7 and CMS routines so any
5776 application which reads PKCS#7 or CMS data from untrusted sources is
5777 affected. SSL/TLS is not affected.
5779 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5785 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5786 This changes the decoding behaviour for some invalid messages,
5787 though the change is mostly in the more lenient direction, and
5788 legacy behaviour is preserved as much as possible.
5792 * In DSA_generate_parameters_ex, if the provided seed is too short,
5793 use a random seed, as already documented.
5795 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5797 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5799 * Alternate chains certificate forgery
5801 During certificate verfification, OpenSSL will attempt to find an
5802 alternative certificate chain if the first attempt to build such a chain
5803 fails. An error in the implementation of this logic can mean that an
5804 attacker could cause certain checks on untrusted certificates to be
5805 bypassed, such as the CA flag, enabling them to use a valid leaf
5806 certificate to act as a CA and "issue" an invalid certificate.
5808 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5814 * Race condition handling PSK identify hint
5816 If PSK identity hints are received by a multi-threaded client then
5817 the values are wrongly updated in the parent SSL_CTX structure. This can
5818 result in a race condition potentially leading to a double free of the
5824 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5826 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5827 incompatibility in the handling of HMAC. The previous ABI has now been
5830 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5832 * Malformed ECParameters causes infinite loop
5834 When processing an ECParameters structure OpenSSL enters an infinite loop
5835 if the curve specified is over a specially malformed binary polynomial
5838 This can be used to perform denial of service against any
5839 system which processes public keys, certificate requests or
5840 certificates. This includes TLS clients and TLS servers with
5841 client authentication enabled.
5843 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5848 * Exploitable out-of-bounds read in X509_cmp_time
5850 X509_cmp_time does not properly check the length of the ASN1_TIME
5851 string and can read a few bytes out of bounds. In addition,
5852 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5855 An attacker can use this to craft malformed certificates and CRLs of
5856 various sizes and potentially cause a segmentation fault, resulting in
5857 a DoS on applications that verify certificates or CRLs. TLS clients
5858 that verify CRLs are affected. TLS clients and servers with client
5859 authentication enabled may be affected if they use custom verification
5862 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5863 independently by Hanno Böck.
5868 * PKCS7 crash with missing EnvelopedContent
5870 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5871 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5872 with missing content and trigger a NULL pointer dereference on parsing.
5874 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5875 structures from untrusted sources are affected. OpenSSL clients and
5876 servers are not affected.
5878 This issue was reported to OpenSSL by Michal Zalewski (Google).
5883 * CMS verify infinite loop with unknown hash function
5885 When verifying a signedData message the CMS code can enter an infinite loop
5886 if presented with an unknown hash function OID. This can be used to perform
5887 denial of service against any system which verifies signedData messages using
5889 This issue was reported to OpenSSL by Johannes Bauer.
5894 * Race condition handling NewSessionTicket
5896 If a NewSessionTicket is received by a multi-threaded client when attempting to
5897 reuse a previous ticket then a race condition can occur potentially leading to
5898 a double free of the ticket data.
5903 * Reject DH handshakes with parameters shorter than 768 bits.
5905 *Kurt Roeckx and Emilia Kasper*
5907 * dhparam: generate 2048-bit parameters by default.
5909 *Kurt Roeckx and Emilia Kasper*
5911 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5913 * Segmentation fault in ASN1_TYPE_cmp fix
5915 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5916 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5917 certificate signature algorithm consistency this can be used to crash any
5918 certificate verification operation and exploited in a DoS attack. Any
5919 application which performs certificate verification is vulnerable including
5920 OpenSSL clients and servers which enable client authentication.
5925 * ASN.1 structure reuse memory corruption fix
5927 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5928 memory corruption via an invalid write. Such reuse is and has been
5929 strongly discouraged and is believed to be rare.
5931 Applications that parse structures containing CHOICE or ANY DEFINED BY
5932 components may be affected. Certificate parsing (d2i_X509 and related
5933 functions) are however not affected. OpenSSL clients and servers are
5939 * PKCS7 NULL pointer dereferences fix
5941 The PKCS#7 parsing code does not handle missing outer ContentInfo
5942 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5943 missing content and trigger a NULL pointer dereference on parsing.
5945 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5946 otherwise parse PKCS#7 structures from untrusted sources are
5947 affected. OpenSSL clients and servers are not affected.
5949 This issue was reported to OpenSSL by Michal Zalewski (Google).
5954 * DoS via reachable assert in SSLv2 servers fix
5956 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5957 servers that both support SSLv2 and enable export cipher suites by sending
5958 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5960 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5961 (OpenSSL development team).
5966 * Use After Free following d2i_ECPrivatekey error fix
5968 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5969 could cause a use after free condition. This, in turn, could cause a double
5970 free in several private key parsing functions (such as d2i_PrivateKey
5971 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5972 for applications that receive EC private keys from untrusted
5973 sources. This scenario is considered rare.
5975 This issue was discovered by the BoringSSL project and fixed in their
5981 * X509_to_X509_REQ NULL pointer deref fix
5983 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5984 the certificate key is invalid. This function is rarely used in practice.
5986 This issue was discovered by Brian Carpenter.
5991 * Removed the export ciphers from the DEFAULT ciphers
5995 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
5997 * Build fixes for the Windows and OpenVMS platforms
5999 *Matt Caswell and Richard Levitte*
6001 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6003 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6004 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6005 dereference. This could lead to a Denial Of Service attack. Thanks to
6006 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6011 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6012 dtls1_buffer_record function under certain conditions. In particular this
6013 could occur if an attacker sent repeated DTLS records with the same
6014 sequence number but for the next epoch. The memory leak could be exploited
6015 by an attacker in a Denial of Service attack through memory exhaustion.
6016 Thanks to Chris Mueller for reporting this issue.
6021 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6022 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6023 method would be set to NULL which could later result in a NULL pointer
6024 dereference. Thanks to Frank Schmirler for reporting this issue.
6029 * Abort handshake if server key exchange message is omitted for ephemeral
6032 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6033 reporting this issue.
6038 * Remove non-export ephemeral RSA code on client and server. This code
6039 violated the TLS standard by allowing the use of temporary RSA keys in
6040 non-export ciphersuites and could be used by a server to effectively
6041 downgrade the RSA key length used to a value smaller than the server
6042 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6043 INRIA or reporting this issue.
6048 * Fixed issue where DH client certificates are accepted without verification.
6049 An OpenSSL server will accept a DH certificate for client authentication
6050 without the certificate verify message. This effectively allows a client to
6051 authenticate without the use of a private key. This only affects servers
6052 which trust a client certificate authority which issues certificates
6053 containing DH keys: these are extremely rare and hardly ever encountered.
6054 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6060 * Ensure that the session ID context of an SSL is updated when its
6061 SSL_CTX is updated via SSL_set_SSL_CTX.
6063 The session ID context is typically set from the parent SSL_CTX,
6064 and can vary with the CTX.
6068 * Fix various certificate fingerprint issues.
6070 By using non-DER or invalid encodings outside the signed portion of a
6071 certificate the fingerprint can be changed without breaking the signature.
6072 Although no details of the signed portion of the certificate can be changed
6073 this can cause problems with some applications: e.g. those using the
6074 certificate fingerprint for blacklists.
6076 1. Reject signatures with non zero unused bits.
6078 If the BIT STRING containing the signature has non zero unused bits reject
6079 the signature. All current signature algorithms require zero unused bits.
6081 2. Check certificate algorithm consistency.
6083 Check the AlgorithmIdentifier inside TBS matches the one in the
6084 certificate signature. NB: this will result in signature failure
6085 errors for some broken certificates.
6087 Thanks to Konrad Kraszewski from Google for reporting this issue.
6089 3. Check DSA/ECDSA signatures use DER.
6091 Re-encode DSA/ECDSA signatures and compare with the original received
6092 signature. Return an error if there is a mismatch.
6094 This will reject various cases including garbage after signature
6095 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6096 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6097 (negative or with leading zeroes).
6099 Further analysis was conducted and fixes were developed by Stephen Henson
6100 of the OpenSSL core team.
6106 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6107 results on some platforms, including x86_64. This bug occurs at random
6108 with a very low probability, and is not known to be exploitable in any
6109 way, though its exact impact is difficult to determine. Thanks to Pieter
6110 Wuille (Blockstream) who reported this issue and also suggested an initial
6111 fix. Further analysis was conducted by the OpenSSL development team and
6112 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6113 the OpenSSL core team.
6118 * Do not resume sessions on the server if the negotiated protocol
6119 version does not match the session's version. Resuming with a different
6120 version, while not strictly forbidden by the RFC, is of questionable
6121 sanity and breaks all known clients.
6123 *David Benjamin, Emilia Käsper*
6125 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6126 early CCS messages during renegotiation. (Note that because
6127 renegotiation is encrypted, this early CCS was not exploitable.)
6131 * Tighten client-side session ticket handling during renegotiation:
6132 ensure that the client only accepts a session ticket if the server sends
6133 the extension anew in the ServerHello. Previously, a TLS client would
6134 reuse the old extension state and thus accept a session ticket if one was
6135 announced in the initial ServerHello.
6137 Similarly, ensure that the client requires a session ticket if one
6138 was advertised in the ServerHello. Previously, a TLS client would
6139 ignore a missing NewSessionTicket message.
6143 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6147 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6148 sends a carefully crafted handshake message, to cause OpenSSL to fail
6149 to free up to 64k of memory causing a memory leak. This could be
6150 exploited in a Denial Of Service attack. This issue affects OpenSSL
6151 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6152 whether SRTP is used or configured. Implementations of OpenSSL that
6153 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6155 The fix was developed by the OpenSSL team.
6160 * Session Ticket Memory Leak.
6162 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6163 integrity of that ticket is first verified. In the event of a session
6164 ticket integrity check failing, OpenSSL will fail to free memory
6165 causing a memory leak. By sending a large number of invalid session
6166 tickets an attacker could exploit this issue in a Denial Of Service
6172 * Build option no-ssl3 is incomplete.
6174 When OpenSSL is configured with "no-ssl3" as a build option, servers
6175 could accept and complete a SSL 3.0 handshake, and clients could be
6176 configured to send them.
6179 *Akamai and the OpenSSL team*
6181 * Add support for TLS_FALLBACK_SCSV.
6182 Client applications doing fallback retries should call
6183 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6186 *Adam Langley, Bodo Moeller*
6188 * Add additional DigestInfo checks.
6190 Re-encode DigestInto in DER and check against the original when
6191 verifying RSA signature: this will reject any improperly encoded
6192 DigestInfo structures.
6194 Note: this is a precautionary measure and no attacks are currently known.
6198 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6200 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6201 SRP code can be overrun an internal buffer. Add sanity check that
6202 g, A, B < N to SRP code.
6204 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6205 Group for discovering this issue.
6210 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6211 TLS 1.0 instead of higher protocol versions when the ClientHello message
6212 is badly fragmented. This allows a man-in-the-middle attacker to force a
6213 downgrade to TLS 1.0 even if both the server and the client support a
6214 higher protocol version, by modifying the client's TLS records.
6216 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6217 researching this issue.
6222 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6223 to a denial of service attack. A malicious server can crash the client
6224 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6225 ciphersuite and sending carefully crafted handshake messages.
6227 Thanks to Felix Gröbert (Google) for discovering and researching this
6233 * By sending carefully crafted DTLS packets an attacker could cause openssl
6234 to leak memory. This can be exploited through a Denial of Service attack.
6235 Thanks to Adam Langley for discovering and researching this issue.
6240 * An attacker can force openssl to consume large amounts of memory whilst
6241 processing DTLS handshake messages. This can be exploited through a
6242 Denial of Service attack.
6243 Thanks to Adam Langley for discovering and researching this issue.
6248 * An attacker can force an error condition which causes openssl to crash
6249 whilst processing DTLS packets due to memory being freed twice. This
6250 can be exploited through a Denial of Service attack.
6251 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6257 * If a multithreaded client connects to a malicious server using a resumed
6258 session and the server sends an ec point format extension it could write
6259 up to 255 bytes to freed memory.
6261 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6267 * A malicious server can crash an OpenSSL client with a null pointer
6268 dereference (read) by specifying an SRP ciphersuite even though it was not
6269 properly negotiated with the client. This can be exploited through a
6270 Denial of Service attack.
6272 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6273 discovering and researching this issue.
6278 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6279 X509_name_oneline, X509_name_print_ex et al. to leak some information
6280 from the stack. Applications may be affected if they echo pretty printing
6281 output to the attacker.
6283 Thanks to Ivan Fratric (Google) for discovering this issue.
6286 *Emilia Käsper, and Steve Henson*
6288 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6289 for corner cases. (Certain input points at infinity could lead to
6290 bogus results, with non-infinity inputs mapped to infinity too.)
6294 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6296 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6297 handshake can force the use of weak keying material in OpenSSL
6298 SSL/TLS clients and servers.
6300 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6301 researching this issue. [CVE-2014-0224][]
6303 *KIKUCHI Masashi, Steve Henson*
6305 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6306 OpenSSL DTLS client the code can be made to recurse eventually crashing
6309 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6312 *Imre Rad, Steve Henson*
6314 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6315 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6316 client or server. This is potentially exploitable to run arbitrary
6317 code on a vulnerable client or server.
6319 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6321 *Jüri Aedla, Steve Henson*
6323 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6324 are subject to a denial of service attack.
6326 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6327 this issue. [CVE-2014-3470][]
6329 *Felix Gröbert, Ivan Fratric, Steve Henson*
6331 * Harmonize version and its documentation. -f flag is used to display
6334 *mancha <mancha1@zoho.com>*
6336 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6337 in i2d_ECPrivateKey.
6339 *mancha <mancha1@zoho.com>*
6341 * Fix some double frees. These are not thought to be exploitable.
6343 *mancha <mancha1@zoho.com>*
6345 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6347 * A missing bounds check in the handling of the TLS heartbeat extension
6348 can be used to reveal up to 64k of memory to a connected client or
6351 Thanks for Neel Mehta of Google Security for discovering this bug and to
6352 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6353 preparing the fix [CVE-2014-0160][]
6355 *Adam Langley, Bodo Moeller*
6357 * Fix for the attack described in the paper "Recovering OpenSSL
6358 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6359 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6360 <http://eprint.iacr.org/2014/140>
6362 Thanks to Yuval Yarom and Naomi Benger for discovering this
6363 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6365 *Yuval Yarom and Naomi Benger*
6367 * TLS pad extension: draft-agl-tls-padding-03
6369 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6370 TLS client Hello record length value would otherwise be > 255 and
6371 less that 512 pad with a dummy extension containing zeroes so it
6372 is at least 512 bytes long.
6374 *Adam Langley, Steve Henson*
6376 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6378 * Fix for TLS record tampering bug. A carefully crafted invalid
6379 handshake could crash OpenSSL with a NULL pointer exception.
6380 Thanks to Anton Johansson for reporting this issues.
6383 * Keep original DTLS digest and encryption contexts in retransmission
6384 structures so we can use the previous session parameters if they need
6385 to be resent. [CVE-2013-6450][]
6389 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6390 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6391 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6392 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6393 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6394 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6396 *Rob Stradling, Adam Langley*
6398 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6400 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6401 supporting platforms or when small records were transferred.
6403 *Andy Polyakov, Steve Henson*
6405 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6407 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6409 This addresses the flaw in CBC record processing discovered by
6410 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6411 at: <http://www.isg.rhul.ac.uk/tls/>
6413 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6414 Security Group at Royal Holloway, University of London
6415 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6416 Emilia Käsper for the initial patch.
6419 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6421 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6422 ciphersuites which can be exploited in a denial of service attack.
6423 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6424 and detecting this bug and to Wolfgang Ettlinger
6425 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6430 * Return an error when checking OCSP signatures when key is NULL.
6431 This fixes a DoS attack. [CVE-2013-0166][]
6435 * Make openssl verify return errors.
6437 *Chris Palmer <palmer@google.com> and Ben Laurie*
6439 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6440 the right response is stapled. Also change SSL_get_certificate()
6441 so it returns the certificate actually sent.
6442 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6444 *Rob Stradling <rob.stradling@comodo.com>*
6446 * Fix possible deadlock when decoding public keys.
6450 * Don't use TLS 1.0 record version number in initial client hello
6455 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6457 * Sanity check record length before skipping explicit IV in TLS
6458 1.2, 1.1 and DTLS to fix DoS attack.
6460 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6461 fuzzing as a service testing platform.
6466 * Initialise tkeylen properly when encrypting CMS messages.
6467 Thanks to Solar Designer of Openwall for reporting this issue.
6471 * In FIPS mode don't try to use composite ciphers as they are not
6476 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6478 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6479 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6480 mean any application compiled against OpenSSL 1.0.0 headers setting
6481 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6482 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6483 0x10000000L Any application which was previously compiled against
6484 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6485 will need to be recompiled as a result. Letting be results in
6486 inability to disable specifically TLS 1.1 and in client context,
6487 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6491 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6492 disable just protocol X, but all protocols above X *if* there are
6493 protocols *below* X still enabled. In more practical terms it means
6494 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6495 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6496 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6501 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6503 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6504 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6505 in CRYPTO_realloc_clean.
6507 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6508 issue and to Adam Langley <agl@chromium.org> for fixing it.
6511 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6513 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6517 * Workarounds for some broken servers that "hang" if a client hello
6518 record length exceeds 255 bytes.
6520 1. Do not use record version number > TLS 1.0 in initial client
6521 hello: some (but not all) hanging servers will now work.
6522 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6523 the number of ciphers sent in the client hello. This should be
6524 set to an even number, such as 50, for example by passing:
6525 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6526 Most broken servers should now work.
6527 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6528 TLS 1.2 client support entirely.
6532 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6536 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6538 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6539 STRING form instead of a DigestInfo.
6543 * The format used for MDC2 RSA signatures is inconsistent between EVP
6544 and the RSA_sign/RSA_verify functions. This was made more apparent when
6545 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6546 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6547 the correct format in RSA_verify so both forms transparently work.
6551 * Some servers which support TLS 1.0 can choke if we initially indicate
6552 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6553 encrypted premaster secret. As a workaround use the maximum permitted
6554 client version in client hello, this should keep such servers happy
6555 and still work with previous versions of OpenSSL.
6559 * Add support for TLS/DTLS heartbeats.
6561 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6563 * Add support for SCTP.
6565 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6567 * Improved PRNG seeding for VOS.
6569 *Paul Green <Paul.Green@stratus.com>*
6571 * Extensive assembler packs updates, most notably:
6573 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6574 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6575 - x86_64: bit-sliced AES implementation;
6576 - ARM: NEON support, contemporary platforms optimizations;
6577 - s390x: z196 support;
6578 - `*`: GHASH and GF(2^m) multiplication implementations;
6582 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6583 (removal of unnecessary code)
6585 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6587 * Add TLS key material exporter from RFC 5705.
6591 * Add DTLS-SRTP negotiation from RFC 5764.
6595 * Add Next Protocol Negotiation,
6596 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6597 disabled with a no-npn flag to config or Configure. Code donated
6600 *Adam Langley <agl@google.com> and Ben Laurie*
6602 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6603 NIST-P256, NIST-P521, with constant-time single point multiplication on
6604 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6605 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6606 Code made available under Apache License version 2.0.
6608 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6609 line to include this in your build of OpenSSL, and run "make depend" (or
6610 "make update"). This enables the following EC_METHODs:
6612 EC_GFp_nistp224_method()
6613 EC_GFp_nistp256_method()
6614 EC_GFp_nistp521_method()
6616 EC_GROUP_new_by_curve_name() will automatically use these (while
6617 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6620 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6622 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6623 all platforms. Move ssize_t definition from e_os.h to the public
6624 header file e_os2.h as it now appears in public header file cms.h
6628 * New -sigopt option to the ca, req and x509 utilities. Additional
6629 signature parameters can be passed using this option and in
6634 * Add RSA PSS signing function. This will generate and set the
6635 appropriate AlgorithmIdentifiers for PSS based on those in the
6636 corresponding EVP_MD_CTX structure. No application support yet.
6640 * Support for companion algorithm specific ASN1 signing routines.
6641 New function ASN1_item_sign_ctx() signs a pre-initialised
6642 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6643 the appropriate parameters.
6647 * Add new algorithm specific ASN1 verification initialisation function
6648 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6649 handling will be the same no matter what EVP_PKEY_METHOD is used.
6650 Add a PSS handler to support verification of PSS signatures: checked
6651 against a number of sample certificates.
6655 * Add signature printing for PSS. Add PSS OIDs.
6657 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6659 * Add algorithm specific signature printing. An individual ASN1 method
6660 can now print out signatures instead of the standard hex dump.
6662 More complex signatures (e.g. PSS) can print out more meaningful
6663 information. Include DSA version that prints out the signature
6668 * Password based recipient info support for CMS library: implementing
6673 * Split password based encryption into PBES2 and PBKDF2 functions. This
6674 neatly separates the code into cipher and PBE sections and is required
6675 for some algorithms that split PBES2 into separate pieces (such as
6676 password based CMS).
6680 * Session-handling fixes:
6681 - Fix handling of connections that are resuming with a session ID,
6682 but also support Session Tickets.
6683 - Fix a bug that suppressed issuing of a new ticket if the client
6684 presented a ticket with an expired session.
6685 - Try to set the ticket lifetime hint to something reasonable.
6686 - Make tickets shorter by excluding irrelevant information.
6687 - On the client side, don't ignore renewed tickets.
6689 *Adam Langley, Bodo Moeller (Google)*
6691 * Fix PSK session representation.
6695 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6697 This work was sponsored by Intel.
6701 * Add GCM support to TLS library. Some custom code is needed to split
6702 the IV between the fixed (from PRF) and explicit (from TLS record)
6703 portions. This adds all GCM ciphersuites supported by RFC5288 and
6704 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6705 add a special AESGCM string for GCM only.
6709 * Expand range of ctrls for AES GCM. Permit setting invocation
6710 field on decrypt and retrieval of invocation field only on encrypt.
6714 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6715 As required by RFC5289 these ciphersuites cannot be used if for
6716 versions of TLS earlier than 1.2.
6720 * For FIPS capable OpenSSL interpret a NULL default public key method
6721 as unset and return the appropriate default but do *not* set the default.
6722 This means we can return the appropriate method in applications that
6723 switch between FIPS and non-FIPS modes.
6727 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6728 ENGINE is used then we cannot handle that in the FIPS module so we
6729 keep original code iff non-FIPS operations are allowed.
6733 * Add -attime option to openssl utilities.
6735 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6737 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6741 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6742 FIPS EC methods unconditionally for now.
6746 * New build option no-ec2m to disable characteristic 2 code.
6750 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6751 all cases can be covered as some introduce binary incompatibilities.
6755 * Redirect RSA operations to FIPS module including keygen,
6756 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6760 * Add similar low level API blocking to ciphers.
6764 * Low level digest APIs are not approved in FIPS mode: any attempt
6765 to use these will cause a fatal error. Applications that *really* want
6766 to use them can use the `private_*` version instead.
6770 * Redirect cipher operations to FIPS module for FIPS builds.
6774 * Redirect digest operations to FIPS module for FIPS builds.
6778 * Update build system to add "fips" flag which will link in fipscanister.o
6779 for static and shared library builds embedding a signature if needed.
6783 * Output TLS supported curves in preference order instead of numerical
6784 order. This is currently hardcoded for the highest order curves first.
6785 This should be configurable so applications can judge speed vs strength.
6789 * Add TLS v1.2 server support for client authentication.
6793 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6798 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6799 FIPS modules versions.
6803 * Add TLS v1.2 client side support for client authentication. Keep cache
6804 of handshake records longer as we don't know the hash algorithm to use
6805 until after the certificate request message is received.
6809 * Initial TLS v1.2 client support. Add a default signature algorithms
6810 extension including all the algorithms we support. Parse new signature
6811 format in client key exchange. Relax some ECC signing restrictions for
6812 TLS v1.2 as indicated in RFC5246.
6816 * Add server support for TLS v1.2 signature algorithms extension. Switch
6817 to new signature format when needed using client digest preference.
6818 All server ciphersuites should now work correctly in TLS v1.2. No client
6819 support yet and no support for client certificates.
6823 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6824 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6825 ciphersuites. At present only RSA key exchange ciphersuites work with
6826 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6827 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6828 and version checking.
6832 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6833 with this defined it will not be affected by any changes to ssl internal
6834 structures. Add several utility functions to allow openssl application
6835 to work with OPENSSL_NO_SSL_INTERN defined.
6839 * A long standing patch to add support for SRP from EdelWeb (Peter
6840 Sylvester and Christophe Renou) was integrated.
6841 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6842 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6845 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6849 * Permit abbreviated handshakes when renegotiating using the function
6850 SSL_renegotiate_abbreviated().
6852 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6854 * Add call to ENGINE_register_all_complete() to
6855 ENGINE_load_builtin_engines(), so some implementations get used
6856 automatically instead of needing explicit application support.
6860 * Add support for TLS key exporter as described in RFC5705.
6862 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6864 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6865 a few changes are required:
6867 Add SSL_OP_NO_TLSv1_1 flag.
6868 Add TLSv1_1 methods.
6869 Update version checking logic to handle version 1.1.
6870 Add explicit IV handling (ported from DTLS code).
6871 Add command line options to s_client/s_server.
6878 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6880 * X509_ATTRIBUTE memory leak
6882 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6883 memory. This structure is used by the PKCS#7 and CMS routines so any
6884 application which reads PKCS#7 or CMS data from untrusted sources is
6885 affected. SSL/TLS is not affected.
6887 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6893 * Race condition handling PSK identify hint
6895 If PSK identity hints are received by a multi-threaded client then
6896 the values are wrongly updated in the parent SSL_CTX structure. This can
6897 result in a race condition potentially leading to a double free of the
6903 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6905 * Malformed ECParameters causes infinite loop
6907 When processing an ECParameters structure OpenSSL enters an infinite loop
6908 if the curve specified is over a specially malformed binary polynomial
6911 This can be used to perform denial of service against any
6912 system which processes public keys, certificate requests or
6913 certificates. This includes TLS clients and TLS servers with
6914 client authentication enabled.
6916 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6921 * Exploitable out-of-bounds read in X509_cmp_time
6923 X509_cmp_time does not properly check the length of the ASN1_TIME
6924 string and can read a few bytes out of bounds. In addition,
6925 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6928 An attacker can use this to craft malformed certificates and CRLs of
6929 various sizes and potentially cause a segmentation fault, resulting in
6930 a DoS on applications that verify certificates or CRLs. TLS clients
6931 that verify CRLs are affected. TLS clients and servers with client
6932 authentication enabled may be affected if they use custom verification
6935 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6936 independently by Hanno Böck.
6941 * PKCS7 crash with missing EnvelopedContent
6943 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6944 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6945 with missing content and trigger a NULL pointer dereference on parsing.
6947 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6948 structures from untrusted sources are affected. OpenSSL clients and
6949 servers are not affected.
6951 This issue was reported to OpenSSL by Michal Zalewski (Google).
6956 * CMS verify infinite loop with unknown hash function
6958 When verifying a signedData message the CMS code can enter an infinite loop
6959 if presented with an unknown hash function OID. This can be used to perform
6960 denial of service against any system which verifies signedData messages using
6962 This issue was reported to OpenSSL by Johannes Bauer.
6967 * Race condition handling NewSessionTicket
6969 If a NewSessionTicket is received by a multi-threaded client when attempting to
6970 reuse a previous ticket then a race condition can occur potentially leading to
6971 a double free of the ticket data.
6976 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
6978 * Segmentation fault in ASN1_TYPE_cmp fix
6980 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6981 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6982 certificate signature algorithm consistency this can be used to crash any
6983 certificate verification operation and exploited in a DoS attack. Any
6984 application which performs certificate verification is vulnerable including
6985 OpenSSL clients and servers which enable client authentication.
6990 * ASN.1 structure reuse memory corruption fix
6992 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6993 memory corruption via an invalid write. Such reuse is and has been
6994 strongly discouraged and is believed to be rare.
6996 Applications that parse structures containing CHOICE or ANY DEFINED BY
6997 components may be affected. Certificate parsing (d2i_X509 and related
6998 functions) are however not affected. OpenSSL clients and servers are
7004 * PKCS7 NULL pointer dereferences fix
7006 The PKCS#7 parsing code does not handle missing outer ContentInfo
7007 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7008 missing content and trigger a NULL pointer dereference on parsing.
7010 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7011 otherwise parse PKCS#7 structures from untrusted sources are
7012 affected. OpenSSL clients and servers are not affected.
7014 This issue was reported to OpenSSL by Michal Zalewski (Google).
7019 * DoS via reachable assert in SSLv2 servers fix
7021 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7022 servers that both support SSLv2 and enable export cipher suites by sending
7023 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7025 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7026 (OpenSSL development team).
7031 * Use After Free following d2i_ECPrivatekey error fix
7033 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7034 could cause a use after free condition. This, in turn, could cause a double
7035 free in several private key parsing functions (such as d2i_PrivateKey
7036 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7037 for applications that receive EC private keys from untrusted
7038 sources. This scenario is considered rare.
7040 This issue was discovered by the BoringSSL project and fixed in their
7046 * X509_to_X509_REQ NULL pointer deref fix
7048 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7049 the certificate key is invalid. This function is rarely used in practice.
7051 This issue was discovered by Brian Carpenter.
7056 * Removed the export ciphers from the DEFAULT ciphers
7060 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7062 * Build fixes for the Windows and OpenVMS platforms
7064 *Matt Caswell and Richard Levitte*
7066 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7068 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7069 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7070 dereference. This could lead to a Denial Of Service attack. Thanks to
7071 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7076 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7077 dtls1_buffer_record function under certain conditions. In particular this
7078 could occur if an attacker sent repeated DTLS records with the same
7079 sequence number but for the next epoch. The memory leak could be exploited
7080 by an attacker in a Denial of Service attack through memory exhaustion.
7081 Thanks to Chris Mueller for reporting this issue.
7086 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7087 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7088 method would be set to NULL which could later result in a NULL pointer
7089 dereference. Thanks to Frank Schmirler for reporting this issue.
7094 * Abort handshake if server key exchange message is omitted for ephemeral
7097 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7098 reporting this issue.
7103 * Remove non-export ephemeral RSA code on client and server. This code
7104 violated the TLS standard by allowing the use of temporary RSA keys in
7105 non-export ciphersuites and could be used by a server to effectively
7106 downgrade the RSA key length used to a value smaller than the server
7107 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7108 INRIA or reporting this issue.
7113 * Fixed issue where DH client certificates are accepted without verification.
7114 An OpenSSL server will accept a DH certificate for client authentication
7115 without the certificate verify message. This effectively allows a client to
7116 authenticate without the use of a private key. This only affects servers
7117 which trust a client certificate authority which issues certificates
7118 containing DH keys: these are extremely rare and hardly ever encountered.
7119 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7125 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7126 results on some platforms, including x86_64. This bug occurs at random
7127 with a very low probability, and is not known to be exploitable in any
7128 way, though its exact impact is difficult to determine. Thanks to Pieter
7129 Wuille (Blockstream) who reported this issue and also suggested an initial
7130 fix. Further analysis was conducted by the OpenSSL development team and
7131 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7132 the OpenSSL core team.
7137 * Fix various certificate fingerprint issues.
7139 By using non-DER or invalid encodings outside the signed portion of a
7140 certificate the fingerprint can be changed without breaking the signature.
7141 Although no details of the signed portion of the certificate can be changed
7142 this can cause problems with some applications: e.g. those using the
7143 certificate fingerprint for blacklists.
7145 1. Reject signatures with non zero unused bits.
7147 If the BIT STRING containing the signature has non zero unused bits reject
7148 the signature. All current signature algorithms require zero unused bits.
7150 2. Check certificate algorithm consistency.
7152 Check the AlgorithmIdentifier inside TBS matches the one in the
7153 certificate signature. NB: this will result in signature failure
7154 errors for some broken certificates.
7156 Thanks to Konrad Kraszewski from Google for reporting this issue.
7158 3. Check DSA/ECDSA signatures use DER.
7160 Reencode DSA/ECDSA signatures and compare with the original received
7161 signature. Return an error if there is a mismatch.
7163 This will reject various cases including garbage after signature
7164 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7165 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7166 (negative or with leading zeroes).
7168 Further analysis was conducted and fixes were developed by Stephen Henson
7169 of the OpenSSL core team.
7175 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7177 * Session Ticket Memory Leak.
7179 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7180 integrity of that ticket is first verified. In the event of a session
7181 ticket integrity check failing, OpenSSL will fail to free memory
7182 causing a memory leak. By sending a large number of invalid session
7183 tickets an attacker could exploit this issue in a Denial Of Service
7189 * Build option no-ssl3 is incomplete.
7191 When OpenSSL is configured with "no-ssl3" as a build option, servers
7192 could accept and complete a SSL 3.0 handshake, and clients could be
7193 configured to send them.
7196 *Akamai and the OpenSSL team*
7198 * Add support for TLS_FALLBACK_SCSV.
7199 Client applications doing fallback retries should call
7200 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7203 *Adam Langley, Bodo Moeller*
7205 * Add additional DigestInfo checks.
7207 Reencode DigestInto in DER and check against the original when
7208 verifying RSA signature: this will reject any improperly encoded
7209 DigestInfo structures.
7211 Note: this is a precautionary measure and no attacks are currently known.
7215 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7217 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7218 to a denial of service attack. A malicious server can crash the client
7219 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7220 ciphersuite and sending carefully crafted handshake messages.
7222 Thanks to Felix Gröbert (Google) for discovering and researching this
7228 * By sending carefully crafted DTLS packets an attacker could cause openssl
7229 to leak memory. This can be exploited through a Denial of Service attack.
7230 Thanks to Adam Langley for discovering and researching this issue.
7235 * An attacker can force openssl to consume large amounts of memory whilst
7236 processing DTLS handshake messages. This can be exploited through a
7237 Denial of Service attack.
7238 Thanks to Adam Langley for discovering and researching this issue.
7243 * An attacker can force an error condition which causes openssl to crash
7244 whilst processing DTLS packets due to memory being freed twice. This
7245 can be exploited through a Denial of Service attack.
7246 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7252 * If a multithreaded client connects to a malicious server using a resumed
7253 session and the server sends an ec point format extension it could write
7254 up to 255 bytes to freed memory.
7256 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7262 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7263 X509_name_oneline, X509_name_print_ex et al. to leak some information
7264 from the stack. Applications may be affected if they echo pretty printing
7265 output to the attacker.
7267 Thanks to Ivan Fratric (Google) for discovering this issue.
7270 *Emilia Käsper, and Steve Henson*
7272 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7273 for corner cases. (Certain input points at infinity could lead to
7274 bogus results, with non-infinity inputs mapped to infinity too.)
7278 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7280 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7281 handshake can force the use of weak keying material in OpenSSL
7282 SSL/TLS clients and servers.
7284 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7285 researching this issue. [CVE-2014-0224][]
7287 *KIKUCHI Masashi, Steve Henson*
7289 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7290 OpenSSL DTLS client the code can be made to recurse eventually crashing
7293 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7296 *Imre Rad, Steve Henson*
7298 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7299 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7300 client or server. This is potentially exploitable to run arbitrary
7301 code on a vulnerable client or server.
7303 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7305 *Jüri Aedla, Steve Henson*
7307 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7308 are subject to a denial of service attack.
7310 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7311 this issue. [CVE-2014-3470][]
7313 *Felix Gröbert, Ivan Fratric, Steve Henson*
7315 * Harmonize version and its documentation. -f flag is used to display
7318 *mancha <mancha1@zoho.com>*
7320 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7321 in i2d_ECPrivateKey.
7323 *mancha <mancha1@zoho.com>*
7325 * Fix some double frees. These are not thought to be exploitable.
7327 *mancha <mancha1@zoho.com>*
7329 * Fix for the attack described in the paper "Recovering OpenSSL
7330 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7331 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7332 <http://eprint.iacr.org/2014/140>
7334 Thanks to Yuval Yarom and Naomi Benger for discovering this
7335 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7337 *Yuval Yarom and Naomi Benger*
7339 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7341 * Keep original DTLS digest and encryption contexts in retransmission
7342 structures so we can use the previous session parameters if they need
7343 to be resent. [CVE-2013-6450][]
7347 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7348 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7349 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7350 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7351 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7352 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7354 *Rob Stradling, Adam Langley*
7356 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7358 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7360 This addresses the flaw in CBC record processing discovered by
7361 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7362 at: <http://www.isg.rhul.ac.uk/tls/>
7364 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7365 Security Group at Royal Holloway, University of London
7366 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7367 Emilia Käsper for the initial patch.
7370 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7372 * Return an error when checking OCSP signatures when key is NULL.
7373 This fixes a DoS attack. [CVE-2013-0166][]
7377 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7378 the right response is stapled. Also change SSL_get_certificate()
7379 so it returns the certificate actually sent.
7380 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7381 (This is a backport)
7383 *Rob Stradling <rob.stradling@comodo.com>*
7385 * Fix possible deadlock when decoding public keys.
7389 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7391 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7394 * Sanity check record length before skipping explicit IV in DTLS
7397 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7398 fuzzing as a service testing platform.
7403 * Initialise tkeylen properly when encrypting CMS messages.
7404 Thanks to Solar Designer of Openwall for reporting this issue.
7408 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7410 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7411 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7412 in CRYPTO_realloc_clean.
7414 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7415 issue and to Adam Langley <agl@chromium.org> for fixing it.
7418 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7420 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7422 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7423 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7424 content decryption and always return the same error. Note: this attack
7425 needs on average 2^20 messages so it only affects automated senders. The
7426 old behaviour can be re-enabled in the CMS code by setting the
7427 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7428 an MMA defence is not necessary.
7429 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7430 this issue. [CVE-2012-0884][]
7434 * Fix CVE-2011-4619: make sure we really are receiving a
7435 client hello before rejecting multiple SGC restarts. Thanks to
7436 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7440 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7442 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7443 Thanks to Antonio Martin, Enterprise Secure Access Research and
7444 Development, Cisco Systems, Inc. for discovering this bug and
7445 preparing a fix. [CVE-2012-0050][]
7449 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7451 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7452 of the Vaudenay padding oracle attack on CBC mode encryption
7453 which enables an efficient plaintext recovery attack against
7454 the OpenSSL implementation of DTLS. Their attack exploits timing
7455 differences arising during decryption processing. A research
7456 paper describing this attack can be found at:
7457 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7458 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7459 Security Group at Royal Holloway, University of London
7460 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7461 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7462 for preparing the fix. [CVE-2011-4108][]
7464 *Robin Seggelmann, Michael Tuexen*
7466 * Clear bytes used for block padding of SSL 3.0 records.
7469 *Adam Langley (Google)*
7471 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7472 Kadianakis <desnacked@gmail.com> for discovering this issue and
7473 Adam Langley for preparing the fix. [CVE-2011-4619][]
7475 *Adam Langley (Google)*
7477 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7479 *Andrey Kulikov <amdeich@gmail.com>*
7481 * Prevent malformed RFC3779 data triggering an assertion failure.
7482 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7483 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7485 *Rob Austein <sra@hactrn.net>*
7487 * Improved PRNG seeding for VOS.
7489 *Paul Green <Paul.Green@stratus.com>*
7491 * Fix ssl_ciph.c set-up race.
7493 *Adam Langley (Google)*
7495 * Fix spurious failures in ecdsatest.c.
7497 *Emilia Käsper (Google)*
7499 * Fix the BIO_f_buffer() implementation (which was mixing different
7500 interpretations of the `..._len` fields).
7502 *Adam Langley (Google)*
7504 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7505 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7506 threads won't reuse the same blinding coefficients.
7508 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7509 lock to call BN_BLINDING_invert_ex, and avoids one use of
7510 BN_BLINDING_update for each BN_BLINDING structure (previously,
7511 the last update always remained unused).
7513 *Emilia Käsper (Google)*
7515 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7517 *Bob Buckholz (Google)*
7519 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7521 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7522 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7524 *Kaspar Brand <ossl@velox.ch>*
7526 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7527 for multi-threaded use of ECDH. [CVE-2011-3210][]
7529 *Adam Langley (Google)*
7531 * Fix x509_name_ex_d2i memory leak on bad inputs.
7535 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7536 signature public key algorithm by using OID xref utilities instead.
7537 Before this you could only use some ECC ciphersuites with SHA1 only.
7541 * Add protection against ECDSA timing attacks as mentioned in the paper
7542 by Billy Bob Brumley and Nicola Tuveri, see:
7543 <http://eprint.iacr.org/2011/232.pdf>
7545 *Billy Bob Brumley and Nicola Tuveri*
7547 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7549 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7551 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7553 * Fix bug in string printing code: if *any* escaping is enabled we must
7554 escape the escape character (backslash) or the resulting string is
7559 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7561 * Disable code workaround for ancient and obsolete Netscape browsers
7562 and servers: an attacker can use it in a ciphersuite downgrade attack.
7563 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7567 * Fixed J-PAKE implementation error, originally discovered by
7568 Sebastien Martini, further info and confirmation from Stefan
7569 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7573 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7575 * Fix extension code to avoid race conditions which can result in a buffer
7576 overrun vulnerability: resumed sessions must not be modified as they can
7577 be shared by multiple threads. CVE-2010-3864
7581 * Fix WIN32 build system to correctly link an ENGINE directory into
7586 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7588 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7591 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7593 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7595 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7596 context. The operation can be customised via the ctrl mechanism in
7597 case ENGINEs want to include additional functionality.
7601 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7605 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7606 output hashes compatible with older versions of OpenSSL.
7608 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7610 * Fix compression algorithm handling: if resuming a session use the
7611 compression algorithm of the resumed session instead of determining
7612 it from client hello again. Don't allow server to change algorithm.
7616 * Add load_crls() function to apps tidying load_certs() too. Add option
7617 to verify utility to allow additional CRLs to be included.
7621 * Update OCSP request code to permit adding custom headers to the request:
7622 some responders need this.
7626 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7629 *Julia Lawall <julia@diku.dk>*
7631 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7632 needlessly dereferenced structures, used obsolete functions and
7633 didn't handle all updated verify codes correctly.
7637 * Disable MD2 in the default configuration.
7641 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7642 indicate the initial BIO being pushed or popped. This makes it possible
7643 to determine whether the BIO is the one explicitly called or as a result
7644 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7645 it handles reference counts correctly and doesn't zero out the I/O bio
7646 when it is not being explicitly popped. WARNING: applications which
7647 included workarounds for the old buggy behaviour will need to be modified
7648 or they could free up already freed BIOs.
7652 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7653 renaming to all platforms (within the 0.9.8 branch, this was
7654 done conditionally on Netware platforms to avoid a name clash).
7656 *Guenter <lists@gknw.net>*
7658 * Add ECDHE and PSK support to DTLS.
7660 *Michael Tuexen <tuexen@fh-muenster.de>*
7662 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7667 * Add "missing" function EVP_MD_flags() (without this the only way to
7668 retrieve a digest flags is by accessing the structure directly. Update
7669 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7670 or cipher is registered as in the "from" argument. Print out all
7671 registered digests in the dgst usage message instead of manually
7672 attempting to work them out.
7676 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7677 this allows the use of compression and extensions. Change default cipher
7678 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7679 by default unless an application cipher string requests it.
7683 * Alter match criteria in PKCS12_parse(). It used to try to use local
7684 key ids to find matching certificates and keys but some PKCS#12 files
7685 don't follow the (somewhat unwritten) rules and this strategy fails.
7686 Now just gather all certificates together and the first private key
7687 then look for the first certificate that matches the key.
7691 * Support use of registered digest and cipher names for dgst and cipher
7692 commands instead of having to add each one as a special case. So now
7699 openssl dgst -sha256 foo
7701 and this works for ENGINE based algorithms too.
7705 * Update Gost ENGINE to support parameter files.
7707 *Victor B. Wagner <vitus@cryptocom.ru>*
7709 * Support GeneralizedTime in ca utility.
7711 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7713 * Enhance the hash format used for certificate directory links. The new
7714 form uses the canonical encoding (meaning equivalent names will work
7715 even if they aren't identical) and uses SHA1 instead of MD5. This form
7716 is incompatible with the older format and as a result c_rehash should
7717 be used to rebuild symbolic links.
7721 * Make PKCS#8 the default write format for private keys, replacing the
7722 traditional format. This form is standardised, more secure and doesn't
7723 include an implicit MD5 dependency.
7727 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7728 committed to OpenSSL should pass this lot as a minimum.
7732 * Add session ticket override functionality for use by EAP-FAST.
7734 *Jouni Malinen <j@w1.fi>*
7736 * Modify HMAC functions to return a value. Since these can be implemented
7737 in an ENGINE errors can occur.
7741 * Type-checked OBJ_bsearch_ex.
7745 * Type-checked OBJ_bsearch. Also some constification necessitated
7746 by type-checking. Still to come: TXT_DB, bsearch(?),
7747 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7752 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7753 seconds to a tm structure directly, instead of going through OS
7754 specific date routines. This avoids any issues with OS routines such
7755 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7756 and X509_time_adj_ex() to cover the extended range. The existing
7757 X509_time_adj() is still usable and will no longer have any date issues.
7761 * Delta CRL support. New use deltas option which will attempt to locate
7762 and search any appropriate delta CRLs available.
7764 This work was sponsored by Google.
7768 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7769 code and add additional score elements. Validate alternate CRL paths
7770 as part of the CRL checking and indicate a new error "CRL path validation
7771 error" in this case. Applications wanting additional details can use
7772 the verify callback and check the new "parent" field. If this is not
7773 NULL CRL path validation is taking place. Existing applications won't
7774 see this because it requires extended CRL support which is off by
7777 This work was sponsored by Google.
7781 * Support for freshest CRL extension.
7783 This work was sponsored by Google.
7787 * Initial indirect CRL support. Currently only supported in the CRLs
7788 passed directly and not via lookup. Process certificate issuer
7789 CRL entry extension and lookup CRL entries by bother issuer name
7790 and serial number. Check and process CRL issuer entry in IDP extension.
7792 This work was sponsored by Google.
7796 * Add support for distinct certificate and CRL paths. The CRL issuer
7797 certificate is validated separately in this case. Only enabled if
7798 an extended CRL support flag is set: this flag will enable additional
7799 CRL functionality in future.
7801 This work was sponsored by Google.
7805 * Add support for policy mappings extension.
7807 This work was sponsored by Google.
7811 * Fixes to pathlength constraint, self issued certificate handling,
7812 policy processing to align with RFC3280 and PKITS tests.
7814 This work was sponsored by Google.
7818 * Support for name constraints certificate extension. DN, email, DNS
7819 and URI types are currently supported.
7821 This work was sponsored by Google.
7825 * To cater for systems that provide a pointer-based thread ID rather
7826 than numeric, deprecate the current numeric thread ID mechanism and
7827 replace it with a structure and associated callback type. This
7828 mechanism allows a numeric "hash" to be extracted from a thread ID in
7829 either case, and on platforms where pointers are larger than 'long',
7830 mixing is done to help ensure the numeric 'hash' is usable even if it
7831 can't be guaranteed unique. The default mechanism is to use "&errno"
7832 as a pointer-based thread ID to distinguish between threads.
7834 Applications that want to provide their own thread IDs should now use
7835 CRYPTO_THREADID_set_callback() to register a callback that will call
7836 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7838 Note that ERR_remove_state() is now deprecated, because it is tied
7839 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7840 to free the current thread's error state should be replaced by
7841 ERR_remove_thread_state(NULL).
7843 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7844 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7845 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7846 application was previously providing a numeric thread callback that
7847 was inappropriate for distinguishing threads, then uniqueness might
7848 have been obtained with &errno that happened immediately in the
7849 intermediate development versions of OpenSSL; this is no longer the
7850 case, the numeric thread callback will now override the automatic use
7853 *Geoff Thorpe, with help from Bodo Moeller*
7855 * Initial support for different CRL issuing certificates. This covers a
7856 simple case where the self issued certificates in the chain exist and
7857 the real CRL issuer is higher in the existing chain.
7859 This work was sponsored by Google.
7863 * Removed effectively defunct crypto/store from the build.
7867 * Revamp of STACK to provide stronger type-checking. Still to come:
7868 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7869 ASN1_STRING, CONF_VALUE.
7873 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7874 RAM on SSL connections. This option can save about 34k per idle SSL.
7878 * Revamp of LHASH to provide stronger type-checking. Still to come:
7879 STACK, TXT_DB, bsearch, qsort.
7883 * Initial support for Cryptographic Message Syntax (aka CMS) based
7884 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7885 support for data, signedData, compressedData, digestedData and
7886 encryptedData, envelopedData types included. Scripts to check against
7887 RFC4134 examples draft and interop and consistency checks of many
7888 content types and variants.
7892 * Add options to enc utility to support use of zlib compression BIO.
7896 * Extend mk1mf to support importing of options and assembly language
7897 files from Configure script, currently only included in VC-WIN32.
7898 The assembly language rules can now optionally generate the source
7899 files from the associated perl scripts.
7903 * Implement remaining functionality needed to support GOST ciphersuites.
7904 Interop testing has been performed using CryptoPro implementations.
7906 *Victor B. Wagner <vitus@cryptocom.ru>*
7908 * s390x assembler pack.
7912 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7917 * Implement Opaque PRF Input TLS extension as specified in
7918 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7919 official specification yet and no extension type assignment by
7920 IANA exists, this extension (for now) will have to be explicitly
7921 enabled when building OpenSSL by providing the extension number
7922 to use. For example, specify an option
7924 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7926 to the "config" or "Configure" script to enable the extension,
7927 assuming extension number 0x9527 (which is a completely arbitrary
7928 and unofficial assignment based on the MD5 hash of the Internet
7929 Draft). Note that by doing so, you potentially lose
7930 interoperability with other TLS implementations since these might
7931 be using the same extension number for other purposes.
7933 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7934 opaque PRF input value to use in the handshake. This will create
7935 an internal copy of the length-'len' string at 'src', and will
7936 return non-zero for success.
7938 To get more control and flexibility, provide a callback function
7941 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7942 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7946 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7949 Callback function 'cb' will be called in handshakes, and is
7950 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7951 Argument 'arg' is for application purposes (the value as given to
7952 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7953 be provided to the callback function). The callback function
7954 has to return non-zero to report success: usually 1 to use opaque
7955 PRF input just if possible, or 2 to enforce use of the opaque PRF
7956 input. In the latter case, the library will abort the handshake
7957 if opaque PRF input is not successfully negotiated.
7959 Arguments 'peerinput' and 'len' given to the callback function
7960 will always be NULL and 0 in the case of a client. A server will
7961 see the client's opaque PRF input through these variables if
7962 available (NULL and 0 otherwise). Note that if the server
7963 provides an opaque PRF input, the length must be the same as the
7964 length of the client's opaque PRF input.
7966 Note that the callback function will only be called when creating
7967 a new session (session resumption can resume whatever was
7968 previously negotiated), and will not be called in SSL 2.0
7969 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7970 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7971 for applications that need to enforce opaque PRF input.
7975 * Update ssl code to support digests other than SHA1+MD5 for handshake
7978 *Victor B. Wagner <vitus@cryptocom.ru>*
7980 * Add RFC4507 support to OpenSSL. This includes the corrections in
7981 RFC4507bis. The encrypted ticket format is an encrypted encoded
7982 SSL_SESSION structure, that way new session features are automatically
7985 If a client application caches session in an SSL_SESSION structure
7986 support is transparent because tickets are now stored in the encoded
7989 The SSL_CTX structure automatically generates keys for ticket
7990 protection in servers so again support should be possible
7991 with no application modification.
7993 If a client or server wishes to disable RFC4507 support then the option
7994 SSL_OP_NO_TICKET can be set.
7996 Add a TLS extension debugging callback to allow the contents of any client
7997 or server extensions to be examined.
7999 This work was sponsored by Google.
8003 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8004 OpenSSL should now compile cleanly on gcc 4.2
8006 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8008 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8009 support including streaming MAC support: this is required for GOST
8010 ciphersuite support.
8012 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8014 * Add option -stream to use PKCS#7 streaming in smime utility. New
8015 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8016 to output in BER and PEM format.
8020 * Experimental support for use of HMAC via EVP_PKEY interface. This
8021 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8022 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8023 ENGINE support for HMAC keys which are unextractable. New -mac and
8024 -macopt options to dgst utility.
8028 * New option -sigopt to dgst utility. Update dgst to use
8029 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8030 alternative signing parameters such as X9.31 or PSS in the dgst
8035 * Change ssl_cipher_apply_rule(), the internal function that does
8036 the work each time a ciphersuite string requests enabling
8037 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8038 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8039 the order of disabled ciphersuites such that those ciphersuites
8040 that most recently went from enabled to disabled not only stay
8041 in order with respect to each other, but also have higher priority
8042 than other disabled ciphersuites the next time ciphersuites are
8045 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8046 the same ciphersuites as with "HIGH" alone, but in a specific
8047 order where the PSK ciphersuites come first (since they are the
8048 most recently disabled ciphersuites when "HIGH" is parsed).
8050 Also, change ssl_create_cipher_list() (using this new
8051 functionality) such that between otherwise identical
8052 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8057 * Change ssl_create_cipher_list() so that it automatically
8058 arranges the ciphersuites in reasonable order before starting
8059 to process the rule string. Thus, the definition for "DEFAULT"
8060 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8061 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8062 This makes it much easier to arrive at a reasonable default order
8063 in applications for which anonymous ciphers are OK (meaning
8064 that you can't actually use DEFAULT).
8066 *Bodo Moeller; suggested by Victor Duchovni*
8068 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8069 processing) into multiple integers instead of setting
8070 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8071 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8072 (These masks as well as the individual bit definitions are hidden
8073 away into the non-exported interface ssl/ssl_locl.h, so this
8074 change to the definition of the SSL_CIPHER structure shouldn't
8075 affect applications.) This give us more bits for each of these
8076 categories, so there is no longer a need to coagulate AES128 and
8077 AES256 into a single algorithm bit, and to coagulate Camellia128
8078 and Camellia256 into a single algorithm bit, which has led to all
8081 Thus, among other things, the kludge introduced in 0.9.7m and
8082 0.9.8e for masking out AES256 independently of AES128 or masking
8083 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8085 With the change, we also introduce new ciphersuite aliases that
8086 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8091 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8092 Use the leftmost N bytes of the signature input if the input is
8093 larger than the prime q (with N being the size in bytes of q).
8097 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8098 it yet and it is largely untested.
8102 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8106 * Initial incomplete changes to avoid need for function casts in OpenSSL
8107 some compilers (gcc 4.2 and later) reject their use. Safestack is
8108 reimplemented. Update ASN1 to avoid use of legacy functions.
8112 * Win32/64 targets are linked with Winsock2.
8116 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8117 to external functions. This can be used to increase CRL handling
8118 efficiency especially when CRLs are very large by (for example) storing
8119 the CRL revoked certificates in a database.
8123 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8124 new CRLs added to a directory can be used. New command line option
8125 -verify_return_error to s_client and s_server. This causes real errors
8126 to be returned by the verify callback instead of carrying on no matter
8127 what. This reflects the way a "real world" verify callback would behave.
8131 * GOST engine, supporting several GOST algorithms and public key formats.
8132 Kindly donated by Cryptocom.
8136 * Partial support for Issuing Distribution Point CRL extension. CRLs
8137 partitioned by DP are handled but no indirect CRL or reason partitioning
8138 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8139 selected via a scoring technique which handles IDP and AKID in CRLs.
8143 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8144 will ultimately be used for all verify operations: this will remove the
8145 X509_STORE dependency on certificate verification and allow alternative
8146 lookup methods. X509_STORE based implementations of these two callbacks.
8150 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8151 Modify get_crl() to find a valid (unexpired) CRL if possible.
8155 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8156 this would be called X509_CRL_cmp() but that name is already used by
8157 a function that just compares CRL issuer names. Cache several CRL
8158 extensions in X509_CRL structure and cache CRLDP in X509.
8162 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8163 this maps equivalent X509_NAME structures into a consistent structure.
8164 Name comparison can then be performed rapidly using memcmp().
8168 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8173 * Allow digests to supply their own micalg string for S/MIME type using
8174 the ctrl EVP_MD_CTRL_MICALG.
8178 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8179 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8180 ctrl. It can then customise the structure before and/or after signing
8185 * New function OBJ_add_sigid() to allow application defined signature OIDs
8186 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8187 to free up any added signature OIDs.
8191 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8192 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8193 digest and cipher tables. New options added to openssl utility:
8194 list-message-digest-algorithms and list-cipher-algorithms.
8198 * Change the array representation of binary polynomials: the list
8199 of degrees of non-zero coefficients is now terminated with -1.
8200 Previously it was terminated with 0, which was also part of the
8201 value; thus, the array representation was not applicable to
8202 polynomials where t^0 has coefficient zero. This change makes
8203 the array representation useful in a more general context.
8207 * Various modifications and fixes to SSL/TLS cipher string
8208 handling. For ECC, the code now distinguishes between fixed ECDH
8209 with RSA certificates on the one hand and with ECDSA certificates
8210 on the other hand, since these are separate ciphersuites. The
8211 unused code for Fortezza ciphersuites has been removed.
8213 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8214 (not "ECDHE"). For consistency with the code for DH
8215 certificates, use of ECDH certificates is now considered ECDH
8216 authentication, not RSA or ECDSA authentication (the latter is
8217 merely the CA's signing algorithm and not actively used in the
8220 The temporary ciphersuite alias "ECCdraft" is no longer
8221 available, and ECC ciphersuites are no longer excluded from "ALL"
8222 and "DEFAULT". The following aliases now exist for RFC 4492
8223 ciphersuites, most of these by analogy with the DH case:
8225 kECDHr - ECDH cert, signed with RSA
8226 kECDHe - ECDH cert, signed with ECDSA
8227 kECDH - ECDH cert (signed with either RSA or ECDSA)
8228 kEECDH - ephemeral ECDH
8229 ECDH - ECDH cert or ephemeral ECDH
8235 AECDH - anonymous ECDH
8236 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8240 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8241 Use correct micalg parameters depending on digest(s) in signed message.
8245 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8246 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8250 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8251 an engine to register a method. Add ENGINE lookups for methods and
8252 functional reference processing.
8256 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8257 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8262 * New -resign option to smime utility. This adds one or more signers
8263 to an existing PKCS#7 signedData structure. Also -md option to use an
8264 alternative message digest algorithm for signing.
8268 * Tidy up PKCS#7 routines and add new functions to make it easier to
8269 create PKCS7 structures containing multiple signers. Update smime
8270 application to support multiple signers.
8274 * New -macalg option to pkcs12 utility to allow setting of an alternative
8279 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8280 Reorganize PBE internals to lookup from a static table using NIDs,
8281 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8282 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8283 PRF which will be automatically used with PBES2.
8287 * Replace the algorithm specific calls to generate keys in "req" with the
8292 * Update PKCS#7 enveloped data routines to use new API. This is now
8293 supported by any public key method supporting the encrypt operation. A
8294 ctrl is added to allow the public key algorithm to examine or modify
8295 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8300 * Add a ctrl to asn1 method to allow a public key algorithm to express
8301 a default digest type to use. In most cases this will be SHA1 but some
8302 algorithms (such as GOST) need to specify an alternative digest. The
8303 return value indicates how strong the preference is 1 means optional and
8304 2 is mandatory (that is it is the only supported type). Modify
8305 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8306 use the default md. Update openssl utilities to use the default digest
8307 type for signing if it is not explicitly indicated.
8311 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8312 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8313 signing method from the key type. This effectively removes the link
8314 between digests and public key types.
8318 * Add an OID cross reference table and utility functions. Its purpose is to
8319 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8320 rsaEncryption. This will allow some of the algorithm specific hackery
8321 needed to use the correct OID to be removed.
8325 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8326 structures for PKCS7_sign(). They are now set up by the relevant public
8331 * Add provisional EC pkey method with support for ECDSA and ECDH.
8335 * Add support for key derivation (agreement) in the API, DH method and
8340 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8341 public and private key formats. As a side effect these add additional
8342 command line functionality not previously available: DSA signatures can be
8343 generated and verified using pkeyutl and DH key support and generation in
8350 *Oliver Tappe <zooey@hirschkaefer.de>*
8352 * New make target "install_html_docs" installs HTML renditions of the
8355 *Oliver Tappe <zooey@hirschkaefer.de>*
8357 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8358 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8359 support key and parameter generation and add initial key generation
8360 functionality for RSA.
8364 * Add functions for main EVP_PKEY_method operations. The undocumented
8365 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8366 `EVP_PKEY_{encrypt,decrypt}_old`.
8370 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8371 key API, doesn't do much yet.
8375 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8376 public key algorithms. New option to openssl utility:
8377 "list-public-key-algorithms" to print out info.
8381 * Implement the Supported Elliptic Curves Extension for
8382 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8386 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8387 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8391 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8392 utilities such as rsa, dsa, dsaparam etc except they process any key
8397 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8398 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8399 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8404 * Initial support for pluggable public key ASN1.
8405 De-spaghettify the public key ASN1 handling. Move public and private
8406 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8407 algorithm specific handling to a single module within the relevant
8408 algorithm directory. Add functions to allow (near) opaque processing
8409 of public and private key structures.
8413 * Implement the Supported Point Formats Extension for
8414 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8418 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8419 for the psk identity [hint] and the psk callback functions to the
8420 SSL_SESSION, SSL and SSL_CTX structure.
8423 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8427 SSL_CTX_use_psk_identity_hint
8428 SSL_get_psk_identity_hint
8429 SSL_get_psk_identity
8430 SSL_use_psk_identity_hint
8432 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8434 * Add RFC 3161 compliant time stamp request creation, response generation
8435 and response verification functionality.
8437 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8439 * Add initial support for TLS extensions, specifically for the server_name
8440 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8441 have new members for a host name. The SSL data structure has an
8442 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8443 stored in that context to allow for session resumption, even after the
8444 SSL has been switched to a new SSL_CTX in reaction to a client's
8445 server_name extension.
8447 New functions (subject to change):
8449 SSL_get_servername()
8450 SSL_get_servername_type()
8453 New CTRL codes and macros (subject to change):
8455 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8456 - SSL_CTX_set_tlsext_servername_callback()
8457 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8458 - SSL_CTX_set_tlsext_servername_arg()
8459 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8461 openssl s_client has a new '-servername ...' option.
8463 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8464 '-key2 ...', '-servername_fatal' (subject to change). This allows
8465 testing the HostName extension for a specific single host name ('-cert'
8466 and '-key' remain fallbacks for handshakes without HostName
8467 negotiation). If the unrecognized_name alert has to be sent, this by
8468 default is a warning; it becomes fatal with the '-servername_fatal'
8471 *Peter Sylvester, Remy Allais, Christophe Renou*
8473 * Whirlpool hash implementation is added.
8477 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8478 bn(64,32). Because of instruction set limitations it doesn't have
8479 any negative impact on performance. This was done mostly in order
8480 to make it possible to share assembler modules, such as bn_mul_mont
8481 implementations, between 32- and 64-bit builds without hassle.
8485 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8486 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8491 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8492 dedicated Montgomery multiplication procedure, is introduced.
8493 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8494 "64-bit" performance on certain 32-bit targets.
8498 * New option SSL_OP_NO_COMP to disable use of compression selectively
8499 in SSL structures. New SSL ctrl to set maximum send fragment size.
8500 Save memory by setting the I/O buffer sizes dynamically instead of
8501 using the maximum available value.
8505 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8506 in addition to the text details.
8510 * Very, very preliminary EXPERIMENTAL support for printing of general
8511 ASN1 structures. This currently produces rather ugly output and doesn't
8512 handle several customised structures at all.
8516 * Integrated support for PVK file format and some related formats such
8517 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8518 these in the 'rsa' and 'dsa' utilities.
8522 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8526 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8527 place for the (very old) "NETSCAPE" format certificates which are now
8528 handled using new ASN1 code equivalents.
8532 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8533 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8534 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8538 * Modify CRL distribution points extension code to print out previously
8539 unsupported fields. Enhance extension setting code to allow setting of
8544 * Add print and set support for Issuing Distribution Point CRL extension.
8548 * Change 'Configure' script to enable Camellia by default.
8555 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8557 * When rejecting SSL/TLS records due to an incorrect version number, never
8558 update s->server with a new major version number. As of
8559 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8560 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8561 the previous behavior could result in a read attempt at NULL when
8562 receiving specific incorrect SSL/TLS records once record payload
8563 protection is active. [CVE-2010-0740][]
8565 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8567 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8568 could be crashed if the relevant tables were not present (e.g. chrooted).
8570 *Tomas Hoger <thoger@redhat.com>*
8572 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8574 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8576 *Martin Olsson, Neel Mehta*
8578 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8579 accommodate for stack sorting, always a write lock!).
8583 * On some versions of WIN32 Heap32Next is very slow. This can cause
8584 excessive delays in the RAND_poll(): over a minute. As a workaround
8585 include a time check in the inner Heap32Next loop too.
8589 * The code that handled flushing of data in SSL/TLS originally used the
8590 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8591 the problem outlined in PR#1949. The fix suggested there however can
8592 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8593 of Apache). So instead simplify the code to flush unconditionally.
8594 This should be fine since flushing with no data to flush is a no op.
8598 * Handle TLS versions 2.0 and later properly and correctly use the
8599 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8600 off ancient servers have a habit of sticking around for a while...
8604 * Modify compression code so it frees up structures without using the
8605 ex_data callbacks. This works around a problem where some applications
8606 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8607 restarting) then use compression (e.g. SSL with compression) later.
8608 This results in significant per-connection memory leaks and
8609 has caused some security issues including CVE-2008-1678 and
8614 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8615 change when encrypting or decrypting.
8619 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8620 connect and renegotiate with servers which do not support RI.
8621 Until RI is more widely deployed this option is enabled by default.
8625 * Add "missing" ssl ctrls to clear options and mode.
8629 * If client attempts to renegotiate and doesn't support RI respond with
8630 a no_renegotiation alert as required by RFC5746. Some renegotiating
8631 TLS clients will continue a connection gracefully when they receive
8632 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8633 waiting for a server hello which it will never receive. Now we treat a
8634 received no_renegotiation alert as a fatal error. This is because
8635 applications requesting a renegotiation might well expect it to succeed
8636 and would have no code in place to handle the server denying it so the
8637 only safe thing to do is to terminate the connection.
8641 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8642 peer supports secure renegotiation and 0 otherwise. Print out peer
8643 renegotiation support in s_client/s_server.
8647 * Replace the highly broken and deprecated SPKAC certification method with
8648 the updated NID creation version. This should correctly handle UTF8.
8652 * Implement RFC5746. Re-enable renegotiation but require the extension
8653 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8654 turns out to be a bad idea. It has been replaced by
8655 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8656 SSL_CTX_set_options(). This is really not recommended unless you
8657 know what you are doing.
8659 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8661 * Fixes to stateless session resumption handling. Use initial_ctx when
8662 issuing and attempting to decrypt tickets in case it has changed during
8663 servername handling. Use a non-zero length session ID when attempting
8664 stateless session resumption: this makes it possible to determine if
8665 a resumption has occurred immediately after receiving server hello
8666 (several places in OpenSSL subtly assume this) instead of later in
8671 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8672 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8673 fixes for a few places where the return code is not checked
8676 *Julia Lawall <julia@diku.dk>*
8678 * Add --strict-warnings option to Configure script to include devteam
8679 warnings in other configurations.
8683 * Add support for --libdir option and LIBDIR variable in makefiles. This
8684 makes it possible to install openssl libraries in locations which
8685 have names other than "lib", for example "/usr/lib64" which some
8688 *Steve Henson, based on patch from Jeremy Utley*
8690 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8691 X690 8.9.12 and can produce some misleading textual output of OIDs.
8693 *Steve Henson, reported by Dan Kaminsky*
8695 * Delete MD2 from algorithm tables. This follows the recommendation in
8696 several standards that it is not used in new applications due to
8697 several cryptographic weaknesses. For binary compatibility reasons
8698 the MD2 API is still compiled in by default.
8702 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8707 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8708 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8711 *Guenter <lists@gknw.net>*
8713 * Fix the server certificate chain building code to use X509_verify_cert(),
8714 it used to have an ad-hoc builder which was unable to cope with anything
8715 other than a simple chain.
8717 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8719 * Don't check self signed certificate signatures in X509_verify_cert()
8720 by default (a flag can override this): it just wastes time without
8721 adding any security. As a useful side effect self signed root CAs
8722 with non-FIPS digests are now usable in FIPS mode.
8726 * In dtls1_process_out_of_seq_message() the check if the current message
8727 is already buffered was missing. For every new message was memory
8728 allocated, allowing an attacker to perform an denial of service attack
8729 with sending out of seq handshake messages until there is no memory
8730 left. Additionally every future message was buffered, even if the
8731 sequence number made no sense and would be part of another handshake.
8732 So only messages with sequence numbers less than 10 in advance will be
8733 buffered. [CVE-2009-1378][]
8735 *Robin Seggelmann, discovered by Daniel Mentz*
8737 * Records are buffered if they arrive with a future epoch to be
8738 processed after finishing the corresponding handshake. There is
8739 currently no limitation to this buffer allowing an attacker to perform
8740 a DOS attack with sending records with future epochs until there is no
8741 memory left. This patch adds the pqueue_size() function to determine
8742 the size of a buffer and limits the record buffer to 100 entries.
8745 *Robin Seggelmann, discovered by Daniel Mentz*
8747 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8748 parent structure is freed. [CVE-2009-1379][]
8752 * Handle non-blocking I/O properly in SSL_shutdown() call.
8754 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8756 * Add `2.5.4.*` OIDs
8758 *Ilya O. <vrghost@gmail.com>*
8760 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8762 * Disable renegotiation completely - this fixes a severe security
8763 problem [CVE-2009-3555][] at the cost of breaking all
8764 renegotiation. Renegotiation can be re-enabled by setting
8765 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8766 run-time. This is really not recommended unless you know what
8771 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8773 * Don't set val to NULL when freeing up structures, it is freed up by
8774 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8775 zeroing past the valid field. [CVE-2009-0789][]
8777 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8779 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8780 checked correctly. This would allow some invalid signed attributes to
8781 appear to verify correctly. [CVE-2009-0591][]
8783 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8785 * Reject UniversalString and BMPString types with invalid lengths. This
8786 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8787 a legal length. [CVE-2009-0590][]
8791 * Set S/MIME signing as the default purpose rather than setting it
8792 unconditionally. This allows applications to override it at the store
8797 * Permit restricted recursion of ASN1 strings. This is needed in practice
8798 to handle some structures.
8802 * Improve efficiency of mem_gets: don't search whole buffer each time
8805 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8807 * New -hex option for openssl rand.
8811 * Print out UTF8String and NumericString when parsing ASN1.
8815 * Support NumericString type for name components.
8819 * Allow CC in the environment to override the automatically chosen
8820 compiler. Note that nothing is done to ensure flags work with the
8825 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8827 * Properly check EVP_VerifyFinal() and similar return values
8830 *Ben Laurie, Bodo Moeller, Google Security Team*
8832 * Enable TLS extensions by default.
8836 * Allow the CHIL engine to be loaded, whether the application is
8837 multithreaded or not. (This does not release the developer from the
8838 obligation to set up the dynamic locking callbacks.)
8840 *Sander Temme <sander@temme.net>*
8842 * Use correct exit code if there is an error in dgst command.
8844 *Steve Henson; problem pointed out by Roland Dirlewanger*
8846 * Tweak Configure so that you need to say "experimental-jpake" to enable
8847 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8851 * Add experimental JPAKE support, including demo authentication in
8852 s_client and s_server.
8856 * Set the comparison function in v3_addr_canonize().
8858 *Rob Austein <sra@hactrn.net>*
8860 * Add support for XMPP STARTTLS in s_client.
8862 *Philip Paeps <philip@freebsd.org>*
8864 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8865 to ensure that even with this option, only ciphersuites in the
8866 server's preference list will be accepted. (Note that the option
8867 applies only when resuming a session, so the earlier behavior was
8868 just about the algorithm choice for symmetric cryptography.)
8872 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8874 * Fix NULL pointer dereference if a DTLS server received
8875 ChangeCipherSpec as first record [CVE-2009-1386][].
8879 * Fix a state transition in s3_srvr.c and d1_srvr.c
8880 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8884 * The fix in 0.9.8c that supposedly got rid of unsafe
8885 double-checked locking was incomplete for RSA blinding,
8886 addressing just one layer of what turns out to have been
8887 doubly unsafe triple-checked locking.
8889 So now fix this for real by retiring the MONT_HELPER macro
8890 in crypto/rsa/rsa_eay.c.
8892 *Bodo Moeller; problem pointed out by Marius Schilder*
8894 * Various precautionary measures:
8896 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8898 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8899 (NB: This would require knowledge of the secret session ticket key
8900 to exploit, in which case you'd be SOL either way.)
8902 - Change bn_nist.c so that it will properly handle input BIGNUMs
8903 outside the expected range.
8905 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8908 *Neel Mehta, Bodo Moeller*
8910 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8911 the load fails. Useful for distros.
8913 *Ben Laurie and the FreeBSD team*
8915 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8919 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8923 * Expand ENGINE to support engine supplied SSL client certificate functions.
8925 This work was sponsored by Logica.
8929 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8930 keystores. Support for SSL/TLS client authentication too.
8931 Not compiled unless enable-capieng specified to Configure.
8933 This work was sponsored by Logica.
8937 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8938 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8939 attribute creation routines such as certificate requests and PKCS#12
8944 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8946 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8947 handshake which could lead to a client crash as found using the
8948 Codenomicon TLS test suite [CVE-2008-1672][]
8950 *Steve Henson, Mark Cox*
8952 * Fix double free in TLS server name extensions which could lead to
8953 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8957 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8959 Clear the error queue to ensure that error entries left from
8960 older function calls do not interfere with the correct operation.
8962 *Lutz Jaenicke, Erik de Castro Lopo*
8964 * Remove root CA certificates of commercial CAs:
8966 The OpenSSL project does not recommend any specific CA and does not
8967 have any policy with respect to including or excluding any CA.
8968 Therefore it does not make any sense to ship an arbitrary selection
8969 of root CA certificates with the OpenSSL software.
8973 * RSA OAEP patches to fix two separate invalid memory reads.
8974 The first one involves inputs when 'lzero' is greater than
8975 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8976 before the beginning of from). The second one involves inputs where
8977 the 'db' section contains nothing but zeroes (there is a one-byte
8978 invalid read after the end of 'db').
8980 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8982 * Partial backport from 0.9.9-dev:
8984 Introduce bn_mul_mont (dedicated Montgomery multiplication
8985 procedure) as a candidate for BIGNUM assembler implementation.
8986 While 0.9.9-dev uses assembler for various architectures, only
8987 x86_64 is available by default here in the 0.9.8 branch, and
8988 32-bit x86 is available through a compile-time setting.
8990 To try the 32-bit x86 assembler implementation, use Configure
8991 option "enable-montasm" (which exists only for this backport).
8993 As "enable-montasm" for 32-bit x86 disclaims code stability
8994 anyway, in this constellation we activate additional code
8995 backported from 0.9.9-dev for further performance improvements,
8996 namely BN_from_montgomery_word. (To enable this otherwise,
8997 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8999 *Andy Polyakov (backport partially by Bodo Moeller)*
9001 * Add TLS session ticket callback. This allows an application to set
9002 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9003 values. This is useful for key rollover for example where several key
9004 sets may exist with different names.
9008 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9009 This was broken until now in 0.9.8 releases, such that the only way
9010 a registered ENGINE could be used (assuming it initialises
9011 successfully on the host) was to explicitly set it as the default
9012 for the relevant algorithms. This is in contradiction with 0.9.7
9013 behaviour and the documentation. With this fix, when an ENGINE is
9014 registered into a given algorithm's table of implementations, the
9015 'uptodate' flag is reset so that auto-discovery will be used next
9016 time a new context for that algorithm attempts to select an
9019 *Ian Lister (tweaked by Geoff Thorpe)*
9021 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9022 implementation in the following ways:
9024 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9027 Lack of BER streaming support means one pass streaming processing is
9028 only supported if data is detached: setting the streaming flag is
9029 ignored for embedded content.
9031 CMS support is disabled by default and must be explicitly enabled
9032 with the enable-cms configuration option.
9036 * Update the GMP engine glue to do direct copies between BIGNUM and
9037 mpz_t when openssl and GMP use the same limb size. Otherwise the
9038 existing "conversion via a text string export" trick is still used.
9040 *Paul Sheer <paulsheer@gmail.com>*
9042 * Zlib compression BIO. This is a filter BIO which compressed and
9043 uncompresses any data passed through it.
9047 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9048 RFC3394 compatible AES key wrapping.
9052 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9053 sets string data without copying. X509_ALGOR_set0() and
9054 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9055 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9056 from an X509_ATTRIBUTE structure optionally checking it occurs only
9057 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9062 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9063 to get the expected BN_FLG_CONSTTIME behavior.
9065 *Bodo Moeller (Google)*
9069 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9070 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9071 - added some more tests to do_tests.pl
9072 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9073 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9074 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9075 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9076 - various changes to netware.pl to enable gcc-cross builds on Win32
9078 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9079 - various changes to fix missing prototype warnings
9080 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9081 - added AES, WHIRLPOOL and CPUID assembler code to build files
9082 - added missing AES assembler make rules to mk1mf.pl
9083 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9085 *Guenter Knauf <eflash@gmx.net>*
9087 * Implement certificate status request TLS extension defined in RFC3546.
9088 A client can set the appropriate parameters and receive the encoded
9089 OCSP response via a callback. A server can query the supplied parameters
9090 and set the encoded OCSP response in the callback. Add simplified examples
9091 to s_client and s_server.
9095 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9098 + Binary incompatibility of ssl_ctx_st structure
9099 + DTLS interoperation with non-compliant servers
9100 + Don't call get_session_cb() without proposed session
9101 + Fix ia64 assembler code
9103 *Andy Polyakov, Steve Henson*
9105 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9107 * DTLS Handshake overhaul. There were longstanding issues with
9108 OpenSSL DTLS implementation, which were making it impossible for
9109 RFC 4347 compliant client to communicate with OpenSSL server.
9110 Unfortunately just fixing these incompatibilities would "cut off"
9111 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9112 server keeps tolerating non RFC compliant syntax. The opposite is
9113 not true, 0.9.8f client can not communicate with earlier server.
9114 This update even addresses CVE-2007-4995.
9118 * Changes to avoid need for function casts in OpenSSL: some compilers
9119 (gcc 4.2 and later) reject their use.
9120 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9123 * Add RFC4507 support to OpenSSL. This includes the corrections in
9124 RFC4507bis. The encrypted ticket format is an encrypted encoded
9125 SSL_SESSION structure, that way new session features are automatically
9128 If a client application caches session in an SSL_SESSION structure
9129 support is transparent because tickets are now stored in the encoded
9132 The SSL_CTX structure automatically generates keys for ticket
9133 protection in servers so again support should be possible
9134 with no application modification.
9136 If a client or server wishes to disable RFC4507 support then the option
9137 SSL_OP_NO_TICKET can be set.
9139 Add a TLS extension debugging callback to allow the contents of any client
9140 or server extensions to be examined.
9142 This work was sponsored by Google.
9146 * Add initial support for TLS extensions, specifically for the server_name
9147 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9148 have new members for a host name. The SSL data structure has an
9149 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9150 stored in that context to allow for session resumption, even after the
9151 SSL has been switched to a new SSL_CTX in reaction to a client's
9152 server_name extension.
9154 New functions (subject to change):
9156 SSL_get_servername()
9157 SSL_get_servername_type()
9160 New CTRL codes and macros (subject to change):
9162 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9163 - SSL_CTX_set_tlsext_servername_callback()
9164 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9165 - SSL_CTX_set_tlsext_servername_arg()
9166 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9168 openssl s_client has a new '-servername ...' option.
9170 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9171 '-key2 ...', '-servername_fatal' (subject to change). This allows
9172 testing the HostName extension for a specific single host name ('-cert'
9173 and '-key' remain fallbacks for handshakes without HostName
9174 negotiation). If the unrecognized_name alert has to be sent, this by
9175 default is a warning; it becomes fatal with the '-servername_fatal'
9178 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9180 * Add AES and SSE2 assembly language support to VC++ build.
9184 * Mitigate attack on final subtraction in Montgomery reduction.
9188 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9189 (which previously caused an internal error).
9193 * Squeeze another 10% out of IGE mode when in != out.
9197 * AES IGE mode speedup.
9199 *Dean Gaudet (Google)*
9201 * Add the Korean symmetric 128-bit cipher SEED (see
9202 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9203 add SEED ciphersuites from RFC 4162:
9205 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9206 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9207 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9208 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9210 To minimize changes between patchlevels in the OpenSSL 0.9.8
9211 series, SEED remains excluded from compilation unless OpenSSL
9212 is configured with 'enable-seed'.
9214 *KISA, Bodo Moeller*
9216 * Mitigate branch prediction attacks, which can be practical if a
9217 single processor is shared, allowing a spy process to extract
9218 information. For detailed background information, see
9219 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9220 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9221 and Necessary Software Countermeasures"). The core of the change
9222 are new versions BN_div_no_branch() and
9223 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9224 respectively, which are slower, but avoid the security-relevant
9225 conditional branches. These are automatically called by BN_div()
9226 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9227 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9228 remove a conditional branch.
9230 BN_FLG_CONSTTIME is the new name for the previous
9231 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9232 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9233 in the exponent causes BN_mod_exp_mont() to use the alternative
9234 implementation in BN_mod_exp_mont_consttime().) The old name
9235 remains as a deprecated alias.
9237 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9238 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9239 constant-time implementations for more than just exponentiation.
9240 Here too the old name is kept as a deprecated alias.
9242 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9243 the BN_BLINDING structure gets an independent copy of the
9244 modulus. This means that the previous `BIGNUM *m` argument to
9245 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9246 essentially becomes `const BIGNUM *m`, although we can't actually
9247 change this in the header file before 0.9.9. It allows
9248 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9249 enable BN_FLG_CONSTTIME.
9251 *Matthew D Wood (Intel Corp)*
9253 * In the SSL/TLS server implementation, be strict about session ID
9254 context matching (which matters if an application uses a single
9255 external cache for different purposes). Previously,
9256 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9257 set. This did ensure strict client verification, but meant that,
9258 with applications using a single external cache for quite
9259 different requirements, clients could circumvent ciphersuite
9260 restrictions for a given session ID context by starting a session
9261 in a different context.
9265 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9266 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9267 authentication-only ciphersuites.
9271 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9272 not complete and could lead to a possible single byte overflow
9273 [CVE-2007-5135][] [Ben Laurie]
9275 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9277 * Since AES128 and AES256 (and similarly Camellia128 and
9278 Camellia256) share a single mask bit in the logic of
9279 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9280 kludge to work properly if AES128 is available and AES256 isn't
9281 (or if Camellia128 is available and Camellia256 isn't).
9285 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9286 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9287 When a point or a seed is encoded in a BIT STRING, we need to
9288 prevent the removal of trailing zero bits to get the proper DER
9289 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9290 of a NamedBitList, for which trailing 0 bits need to be removed.)
9294 * Have SSL/TLS server implementation tolerate "mismatched" record
9295 protocol version while receiving ClientHello even if the
9296 ClientHello is fragmented. (The server can't insist on the
9297 particular protocol version it has chosen before the ServerHello
9298 message has informed the client about his choice.)
9302 * Add RFC 3779 support.
9304 *Rob Austein for ARIN, Ben Laurie*
9306 * Load error codes if they are not already present instead of using a
9307 static variable. This allows them to be cleanly unloaded and reloaded.
9308 Improve header file function name parsing.
9312 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9313 or CAPABILITY handshake as required by RFCs.
9317 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9319 * Introduce limits to prevent malicious keys being able to
9320 cause a denial of service. [CVE-2006-2940][]
9322 *Steve Henson, Bodo Moeller*
9324 * Fix ASN.1 parsing of certain invalid structures that can result
9325 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9327 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9328 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9330 * Fix SSL client code which could crash if connecting to a
9331 malicious SSLv2 server. [CVE-2006-4343][]
9333 *Tavis Ormandy and Will Drewry, Google Security Team*
9335 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9336 match only those. Before that, "AES256-SHA" would be interpreted
9337 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9338 the same strength classification in 0.9.7h) as we currently only
9339 have a single AES bit in the ciphersuite description bitmap.
9340 That change, however, also applied to ciphersuite strings such as
9341 "RC4-MD5" that intentionally matched multiple ciphersuites --
9342 namely, SSL 2.0 ciphersuites in addition to the more common ones
9343 from SSL 3.0/TLS 1.0.
9345 So we change the selection algorithm again: Naming an explicit
9346 ciphersuite selects this one ciphersuite, and any other similar
9347 ciphersuite (same bitmap) from *other* protocol versions.
9348 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9349 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9351 Since SSL 2.0 does not have any ciphersuites for which the
9352 128/256 bit distinction would be relevant, this works for now.
9353 The proper fix will be to use different bits for AES128 and
9354 AES256, which would have avoided the problems from the beginning;
9355 however, bits are scarce, so we can only do this in a new release
9356 (not just a patchlevel) when we can change the SSL_CIPHER
9357 definition to split the single 'unsigned long mask' bitmap into
9358 multiple values to extend the available space.
9362 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9364 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9365 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9367 * Add AES IGE and biIGE modes.
9371 * Change the Unix randomness entropy gathering to use poll() when
9372 possible instead of select(), since the latter has some
9373 undesirable limitations.
9375 *Darryl Miles via Richard Levitte and Bodo Moeller*
9377 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9378 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9379 cannot be implicitly activated as part of, e.g., the "AES" alias.
9380 However, please upgrade to OpenSSL 0.9.9[-dev] for
9381 non-experimental use of the ECC ciphersuites to get TLS extension
9382 support, which is required for curve and point format negotiation
9383 to avoid potential handshake problems.
9387 * Disable rogue ciphersuites:
9389 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9390 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9391 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9393 The latter two were purportedly from
9394 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9397 Also deactivate the remaining ciphersuites from
9398 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9399 unofficial, and the ID has long expired.
9403 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9404 dual-core machines) and other potential thread-safety issues.
9408 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9409 versions), which is now available for royalty-free use
9410 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9411 Also, add Camellia TLS ciphersuites from RFC 4132.
9413 To minimize changes between patchlevels in the OpenSSL 0.9.8
9414 series, Camellia remains excluded from compilation unless OpenSSL
9415 is configured with 'enable-camellia'.
9419 * Disable the padding bug check when compression is in use. The padding
9420 bug check assumes the first packet is of even length, this is not
9421 necessarily true if compression is enabled and can result in false
9422 positives causing handshake failure. The actual bug test is ancient
9423 code so it is hoped that implementations will either have fixed it by
9424 now or any which still have the bug do not support compression.
9428 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9430 * When applying a cipher rule check to see if string match is an explicit
9431 cipher suite and only match that one cipher suite if it is.
9435 * Link in manifests for VC++ if needed.
9437 *Austin Ziegler <halostatue@gmail.com>*
9439 * Update support for ECC-based TLS ciphersuites according to
9440 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9441 TLS extensions, which are supported starting with the 0.9.9
9442 branch, not in the OpenSSL 0.9.8 branch).
9446 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9447 opaque EVP_CIPHER_CTX handling.
9451 * Fixes and enhancements to zlib compression code. We now only use
9452 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9453 to conform with the standards mentioned here:
9454 <http://www.zlib.net/DLL_FAQ.txt>
9455 Static zlib linking now works on Windows and the new --with-zlib-include
9456 --with-zlib-lib options to Configure can be used to supply the location
9457 of the headers and library. Gracefully handle case where zlib library
9462 * Several fixes and enhancements to the OID generation code. The old code
9463 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9464 handle numbers larger than ULONG_MAX, truncated printing and had a
9465 non standard OBJ_obj2txt() behaviour.
9469 * Add support for building of engines under engine/ as shared libraries
9470 under VC++ build system.
9474 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9475 Hopefully, we will not see any false combination of paths any more.
9479 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9481 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9482 (part of SSL_OP_ALL). This option used to disable the
9483 countermeasure against man-in-the-middle protocol-version
9484 rollback in the SSL 2.0 server implementation, which is a bad
9485 idea. [CVE-2005-2969][]
9487 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9488 for Information Security, National Institute of Advanced Industrial
9489 Science and Technology [AIST], Japan)*
9491 * Add two function to clear and return the verify parameter flags.
9495 * Keep cipherlists sorted in the source instead of sorting them at
9496 runtime, thus removing the need for a lock.
9500 * Avoid some small subgroup attacks in Diffie-Hellman.
9502 *Nick Mathewson and Ben Laurie*
9504 * Add functions for well-known primes.
9508 * Extended Windows CE support.
9510 *Satoshi Nakamura and Andy Polyakov*
9512 * Initialize SSL_METHOD structures at compile time instead of during
9513 runtime, thus removing the need for a lock.
9517 * Make PKCS7_decrypt() work even if no certificate is supplied by
9518 attempting to decrypt each encrypted key in turn. Add support to
9523 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9525 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9528 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9532 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9533 key into the same file any more.
9537 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9541 * Add -utf8 command line and config file option to 'ca'.
9543 *Stefan <stf@udoma.org*
9545 * Removed the macro des_crypt(), as it seems to conflict with some
9546 libraries. Use DES_crypt().
9550 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9551 involves renaming the source and generated shared-libs for
9552 both. The engines will accept the corrected or legacy ids
9553 ('ncipher' and '4758_cca' respectively) when binding. NB,
9554 this only applies when building 'shared'.
9556 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9558 * Add attribute functions to EVP_PKEY structure. Modify
9559 PKCS12_create() to recognize a CSP name attribute and
9560 use it. Make -CSP option work again in pkcs12 utility.
9564 * Add new functionality to the bn blinding code:
9565 - automatic re-creation of the BN_BLINDING parameters after
9566 a fixed number of uses (currently 32)
9567 - add new function for parameter creation
9568 - introduce flags to control the update behaviour of the
9569 BN_BLINDING parameters
9570 - hide BN_BLINDING structure
9571 Add a second BN_BLINDING slot to the RSA structure to improve
9572 performance when a single RSA object is shared among several
9577 * Add support for DTLS.
9579 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9581 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9582 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9586 * Remove buggy and incomplete DH cert support from
9587 ssl/ssl_rsa.c and ssl/s3_both.c
9591 * Use SHA-1 instead of MD5 as the default digest algorithm for
9592 the apps/openssl applications.
9596 * Compile clean with "-Wall -Wmissing-prototypes
9597 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9598 DEBUG_SAFESTACK must also be set.
9602 * Change ./Configure so that certain algorithms can be disabled by default.
9603 The new counterpiece to "no-xxx" is "enable-xxx".
9605 The patented RC5 and MDC2 algorithms will now be disabled unless
9606 "enable-rc5" and "enable-mdc2", respectively, are specified.
9608 (IDEA remains enabled despite being patented. This is because IDEA
9609 is frequently required for interoperability, and there is no license
9610 fee for non-commercial use. As before, "no-idea" can be used to
9611 avoid this algorithm.)
9615 * Add processing of proxy certificates (see RFC 3820). This work was
9616 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9617 EGEE (Enabling Grids for E-science in Europe).
9621 * RC4 performance overhaul on modern architectures/implementations, such
9622 as Intel P4, IA-64 and AMD64.
9626 * New utility extract-section.pl. This can be used specify an alternative
9627 section number in a pod file instead of having to treat each file as
9628 a separate case in Makefile. This can be done by adding two lines to the
9631 =for comment openssl_section:XXX
9633 The blank line is mandatory.
9637 * New arguments -certform, -keyform and -pass for s_client and s_server
9638 to allow alternative format key and certificate files and passphrase
9643 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9644 update associated structures and add various utility functions.
9646 Add new policy related verify parameters, include policy checking in
9647 standard verify code. Enhance 'smime' application with extra parameters
9648 to support policy checking and print out.
9652 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9653 Nehemiah processors. These extensions support AES encryption in hardware
9654 as well as RNG (though RNG support is currently disabled).
9656 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9658 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9662 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9664 *Andy Polyakov and a number of other people*
9666 * Improved PowerPC platform support. Most notably BIGNUM assembler
9667 implementation contributed by IBM.
9669 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9671 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9672 exponent rather than 'unsigned long'. There is a corresponding change to
9673 the new 'rsa_keygen' element of the RSA_METHOD structure.
9675 *Jelte Jansen, Geoff Thorpe*
9677 * Functionality for creating the initial serial number file is now
9678 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9680 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9681 number file to 1, which is bound to cause problems. To avoid
9682 the problems while respecting compatibility between different 0.9.7
9683 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9684 CA.pl for serial number initialization. With the new release 0.9.8,
9685 we can fix the problem directly in the 'ca' utility.)
9689 * Reduced header interdependencies by declaring more opaque objects in
9690 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9691 give fewer recursive includes, which could break lazy source code - so
9692 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9693 developers should define this symbol when building and using openssl to
9694 ensure they track the recommended behaviour, interfaces, [etc], but
9695 backwards-compatible behaviour prevails when this isn't defined.
9699 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9703 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9704 This will generate a random key of the appropriate length based on the
9705 cipher context. The EVP_CIPHER can provide its own random key generation
9706 routine to support keys of a specific form. This is used in the des and
9707 3des routines to generate a key of the correct parity. Update S/MIME
9708 code to use new functions and hence generate correct parity DES keys.
9709 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9710 valid (weak or incorrect parity).
9714 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9715 as looking them up. This is useful when the verified structure may contain
9716 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9717 present unless the new PKCS7_NO_CRL flag is asserted.
9721 * Extend ASN1 oid configuration module. It now additionally accepts the
9724 shortName = some long name, 1.2.3.4
9728 * Reimplemented the BN_CTX implementation. There is now no more static
9729 limitation on the number of variables it can handle nor the depth of the
9730 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9731 information can now expand as required, and rather than having a single
9732 static array of bignums, BN_CTX now uses a linked-list of such arrays
9733 allowing it to expand on demand whilst maintaining the usefulness of
9734 BN_CTX's "bundling".
9738 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9739 to allow all RSA operations to function using a single BN_CTX.
9743 * Preliminary support for certificate policy evaluation and checking. This
9744 is initially intended to pass the tests outlined in "Conformance Testing
9745 of Relying Party Client Certificate Path Processing Logic" v1.07.
9749 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9750 remained unused and not that useful. A variety of other little bignum
9751 tweaks and fixes have also been made continuing on from the audit (see
9756 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9757 associated ASN1, EVP and SSL functions and old ASN1 macros.
9761 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9762 and this should never fail. So the return value from the use of
9763 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9764 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9768 * BN_CTX_get() should return zero-valued bignums, providing the same
9769 initialised value as BN_new().
9771 *Geoff Thorpe, suggested by Ulf Möller*
9773 * Support for inhibitAnyPolicy certificate extension.
9777 * An audit of the BIGNUM code is underway, for which debugging code is
9778 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9779 is considered valid when processing BIGNUMs, and causes execution to
9780 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9781 further steps are taken to deliberately pollute unused data in BIGNUM
9782 structures to try and expose faulty code further on. For now, openssl will
9783 (in its default mode of operation) continue to tolerate the inconsistent
9784 forms that it has tolerated in the past, but authors and packagers should
9785 consider trying openssl and their own applications when compiled with
9786 these debugging symbols defined. It will help highlight potential bugs in
9787 their own code, and will improve the test coverage for OpenSSL itself. At
9788 some point, these tighter rules will become openssl's default to improve
9789 maintainability, though the assert()s and other overheads will remain only
9790 in debugging configurations. See bn.h for more details.
9792 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9794 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9795 that can only be obtained through BN_CTX_new() (which implicitly
9796 initialises it). The presence of this function only made it possible
9797 to overwrite an existing structure (and cause memory leaks).
9801 * Because of the callback-based approach for implementing LHASH as a
9802 template type, lh_insert() adds opaque objects to hash-tables and
9803 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9804 to clean up those corresponding objects before destroying the hash table
9805 (and losing the object pointers). So some over-zealous constifications in
9806 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9807 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9808 prototyped to have "const" restrictions on the object pointers they are
9809 given (and so aren't required to cast them away any more).
9813 * The tmdiff.h API was so ugly and minimal that our own timing utility
9814 (speed) prefers to use its own implementation. The two implementations
9815 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9816 its object type properly exposed (MS_TM) instead of casting to/from
9817 `char *`. This may still change yet if someone realises MS_TM and
9819 aren't necessarily the greatest nomenclatures - but this is what was used
9820 internally to the implementation so I've used that for now.
9824 * Ensure that deprecated functions do not get compiled when
9825 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9826 the self-tests were still using deprecated key-generation functions so
9827 these have been updated also.
9831 * Reorganise PKCS#7 code to separate the digest location functionality
9832 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9833 New function PKCS7_set_digest() to set the digest type for PKCS#7
9834 digestedData type. Add additional code to correctly generate the
9835 digestedData type and add support for this type in PKCS7 initialization
9840 * New function PKCS7_set0_type_other() this initializes a PKCS7
9841 structure of type "other".
9845 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9846 sure the loop does correctly stop and breaking ("division by zero")
9847 modulus operations are not performed. The (pre-generated) prime
9848 table crypto/bn/bn_prime.h was already correct, but it could not be
9849 re-generated on some platforms because of the "division by zero"
9850 situation in the script.
9852 *Ralf S. Engelschall*
9854 * Update support for ECC-based TLS ciphersuites according to
9855 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9856 SHA-1 now is only used for "small" curves (where the
9857 representation of a field element takes up to 24 bytes); for
9858 larger curves, the field element resulting from ECDH is directly
9859 used as premaster secret.
9861 *Douglas Stebila (Sun Microsystems Laboratories)*
9863 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9864 curve secp160r1 to the tests.
9866 *Douglas Stebila (Sun Microsystems Laboratories)*
9868 * Add the possibility to load symbols globally with DSO.
9870 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9872 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9873 control of the error stack.
9877 * Add support for STORE in ENGINE.
9881 * Add the STORE type. The intention is to provide a common interface
9882 to certificate and key stores, be they simple file-based stores, or
9883 HSM-type store, or LDAP stores, or...
9884 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9888 * Add a generic structure called OPENSSL_ITEM. This can be used to
9889 pass a list of arguments to any function as well as provide a way
9890 for a function to pass data back to the caller.
9894 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9895 works like BUF_strdup() but can be used to duplicate a portion of
9896 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9901 * Add the function sk_find_ex() which works like sk_find(), but will
9902 return an index to an element even if an exact match couldn't be
9903 found. The index is guaranteed to point at the element where the
9904 searched-for key would be inserted to preserve sorting order.
9908 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9909 takes an extra flags argument for optional functionality. Currently,
9910 the following flags are defined:
9912 OBJ_BSEARCH_VALUE_ON_NOMATCH
9913 This one gets OBJ_bsearch_ex() to return a pointer to the first
9914 element where the comparing function returns a negative or zero
9917 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9918 This one gets OBJ_bsearch_ex() to return a pointer to the first
9919 element where the comparing function returns zero. This is useful
9920 if there are more than one element where the comparing function
9925 * Make it possible to create self-signed certificates with 'openssl ca'
9926 in such a way that the self-signed certificate becomes part of the
9927 CA database and uses the same mechanisms for serial number generation
9928 as all other certificate signing. The new flag '-selfsign' enables
9929 this functionality. Adapt CA.sh and CA.pl.in.
9933 * Add functionality to check the public key of a certificate request
9934 against a given private. This is useful to check that a certificate
9935 request can be signed by that key (self-signing).
9939 * Make it possible to have multiple active certificates with the same
9940 subject in the CA index file. This is done only if the keyword
9941 'unique_subject' is set to 'no' in the main CA section (default
9942 if 'CA_default') of the configuration file. The value is saved
9943 with the database itself in a separate index attribute file,
9944 named like the index file with '.attr' appended to the name.
9948 * Generate multi-valued AVAs using '+' notation in config files for
9953 * Support for nameConstraints certificate extension.
9957 * Support for policyConstraints certificate extension.
9961 * Support for policyMappings certificate extension.
9965 * Make sure the default DSA_METHOD implementation only uses its
9966 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9967 and change its own handlers to be NULL so as to remove unnecessary
9968 indirection. This lets alternative implementations fallback to the
9969 default implementation more easily.
9973 * Support for directoryName in GeneralName related extensions
9978 * Make it possible to link applications using Makefile.shared.
9979 Make that possible even when linking against static libraries!
9983 * Support for single pass processing for S/MIME signing. This now
9984 means that S/MIME signing can be done from a pipe, in addition
9985 cleartext signing (multipart/signed type) is effectively streaming
9986 and the signed data does not need to be all held in memory.
9988 This is done with a new flag PKCS7_STREAM. When this flag is set
9989 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9990 is done after the data is output (and digests calculated) in
9991 SMIME_write_PKCS7().
9995 * Add full support for -rpath/-R, both in shared libraries and
9996 applications, at least on the platforms where it's known how
10001 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10002 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10003 will now compute a table of multiples of the generator that
10004 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10005 faster (notably in the case of a single point multiplication,
10006 scalar * generator).
10008 *Nils Larsch, Bodo Moeller*
10010 * IPv6 support for certificate extensions. The various extensions
10011 which use the IP:a.b.c.d can now take IPv6 addresses using the
10012 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10017 * Added an ENGINE that implements RSA by performing private key
10018 exponentiations with the GMP library. The conversions to and from
10019 GMP's mpz_t format aren't optimised nor are any montgomery forms
10020 cached, and on x86 it appears OpenSSL's own performance has caught up.
10021 However there are likely to be other architectures where GMP could
10022 provide a boost. This ENGINE is not built in by default, but it can be
10023 specified at Configure time and should be accompanied by the necessary
10024 linker additions, eg;
10025 ./config -DOPENSSL_USE_GMP -lgmp
10029 * "openssl engine" will not display ENGINE/DSO load failure errors when
10030 testing availability of engines with "-t" - the old behaviour is
10031 produced by increasing the feature's verbosity with "-tt".
10035 * ECDSA routines: under certain error conditions uninitialized BN objects
10036 could be freed. Solution: make sure initialization is performed early
10037 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10042 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10043 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10044 software implementations. For DSA and DH, parameter generation can
10045 also be overridden by providing the appropriate method callbacks.
10049 * Change the "progress" mechanism used in key-generation and
10050 primality testing to functions that take a new BN_GENCB pointer in
10051 place of callback/argument pairs. The new API functions have `_ex`
10052 postfixes and the older functions are reimplemented as wrappers for
10053 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10054 declarations of the old functions to help (graceful) attempts to
10055 migrate to the new functions. Also, the new key-generation API
10056 functions operate on a caller-supplied key-structure and return
10057 success/failure rather than returning a key or NULL - this is to
10058 help make "keygen" another member function of RSA_METHOD etc.
10060 Example for using the new callback interface:
10062 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10063 void *my_arg = ...;
10066 BN_GENCB_set(&my_cb, my_callback, my_arg);
10068 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10069 /* For the meaning of a, b in calls to my_callback(), see the
10070 * documentation of the function that calls the callback.
10071 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10072 * my_callback should return 1 if it wants BN_is_prime_ex()
10073 * to continue, or 0 to stop.
10078 * Change the ZLIB compression method to be stateful, and make it
10079 available to TLS with the number defined in
10080 draft-ietf-tls-compression-04.txt.
10084 * Add the ASN.1 structures and functions for CertificatePair, which
10085 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10087 CertificatePair ::= SEQUENCE {
10088 forward [0] Certificate OPTIONAL,
10089 reverse [1] Certificate OPTIONAL,
10090 -- at least one of the pair shall be present -- }
10092 Also implement the PEM functions to read and write certificate
10093 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10095 This needed to be defined, mostly for the sake of the LDAP
10096 attribute crossCertificatePair, but may prove useful elsewhere as
10101 * Make it possible to inhibit symlinking of shared libraries in
10102 Makefile.shared, for Cygwin's sake.
10106 * Extend the BIGNUM API by creating a function
10107 void BN_set_negative(BIGNUM *a, int neg);
10108 and a macro that behave like
10109 int BN_is_negative(const BIGNUM *a);
10111 to avoid the need to access 'a->neg' directly in applications.
10115 * Implement fast modular reduction for pseudo-Mersenne primes
10116 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10117 EC_GROUP_new_curve_GFp() will now automatically use this
10120 *Nils Larsch <nla@trustcenter.de>*
10122 * Add new lock type (CRYPTO_LOCK_BN).
10126 * Change the ENGINE framework to automatically load engines
10127 dynamically from specific directories unless they could be
10128 found to already be built in or loaded. Move all the
10129 current engines except for the cryptodev one to a new
10130 directory engines/.
10131 The engines in engines/ are built as shared libraries if
10132 the "shared" options was given to ./Configure or ./config.
10133 Otherwise, they are inserted in libcrypto.a.
10134 /usr/local/ssl/engines is the default directory for dynamic
10135 engines, but that can be overridden at configure time through
10136 the usual use of --prefix and/or --openssldir, and at run
10137 time with the environment variable OPENSSL_ENGINES.
10139 *Geoff Thorpe and Richard Levitte*
10141 * Add Makefile.shared, a helper makefile to build shared
10142 libraries. Adapt Makefile.org.
10146 * Add version info to Win32 DLLs.
10148 *Peter 'Luna' Runestig" <peter@runestig.com>*
10150 * Add new 'medium level' PKCS#12 API. Certificates and keys
10151 can be added using this API to created arbitrary PKCS#12
10152 files while avoiding the low level API.
10154 New options to PKCS12_create(), key or cert can be NULL and
10155 will then be omitted from the output file. The encryption
10156 algorithm NIDs can be set to -1 for no encryption, the mac
10157 iteration count can be set to 0 to omit the mac.
10159 Enhance pkcs12 utility by making the -nokeys and -nocerts
10160 options work when creating a PKCS#12 file. New option -nomac
10161 to omit the mac, NONE can be set for an encryption algorithm.
10162 New code is modified to use the enhanced PKCS12_create()
10163 instead of the low level API.
10167 * Extend ASN1 encoder to support indefinite length constructed
10168 encoding. This can output sequences tags and octet strings in
10169 this form. Modify pk7_asn1.c to support indefinite length
10170 encoding. This is experimental and needs additional code to
10171 be useful, such as an ASN1 bio and some enhanced streaming
10174 Extend template encode functionality so that tagging is passed
10175 down to the template encoder.
10179 * Let 'openssl req' fail if an argument to '-newkey' is not
10180 recognized instead of using RSA as a default.
10184 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10185 As these are not official, they are not included in "ALL";
10186 the "ECCdraft" ciphersuite group alias can be used to select them.
10188 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10190 * Add ECDH engine support.
10192 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10194 * Add ECDH in new directory crypto/ecdh/.
10196 *Douglas Stebila (Sun Microsystems Laboratories)*
10198 * Let BN_rand_range() abort with an error after 100 iterations
10199 without success (which indicates a broken PRNG).
10203 * Change BN_mod_sqrt() so that it verifies that the input value
10204 is really the square of the return value. (Previously,
10205 BN_mod_sqrt would show GIGO behaviour.)
10209 * Add named elliptic curves over binary fields from X9.62, SECG,
10210 and WAP/WTLS; add OIDs that were still missing.
10212 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10214 * Extend the EC library for elliptic curves over binary fields
10215 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10218 EC_GF2m_simple_method
10222 EC_GROUP_new_curve_GF2m
10223 EC_GROUP_set_curve_GF2m
10224 EC_GROUP_get_curve_GF2m
10225 EC_POINT_set_affine_coordinates_GF2m
10226 EC_POINT_get_affine_coordinates_GF2m
10227 EC_POINT_set_compressed_coordinates_GF2m
10229 Point compression for binary fields is disabled by default for
10230 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10233 As binary polynomials are represented as BIGNUMs, various members
10234 of the EC_GROUP and EC_POINT data structures can be shared
10235 between the implementations for prime fields and binary fields;
10236 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10237 are essentially identical to their `..._GFp` counterparts.
10238 (For simplicity, the `..._GFp` prefix has been dropped from
10239 various internal method names.)
10241 An internal 'field_div' method (similar to 'field_mul' and
10242 'field_sqr') has been added; this is used only for binary fields.
10244 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10246 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10247 through methods ('mul', 'precompute_mult').
10249 The generic implementations (now internally called 'ec_wNAF_mul'
10250 and 'ec_wNAF_precomputed_mult') remain the default if these
10251 methods are undefined.
10253 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10255 * New function EC_GROUP_get_degree, which is defined through
10256 EC_METHOD. For curves over prime fields, this returns the bit
10257 length of the modulus.
10259 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10261 * New functions EC_GROUP_dup, EC_POINT_dup.
10262 (These simply call ..._new and ..._copy).
10264 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10266 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10267 Polynomials are represented as BIGNUMs (where the sign bit is not
10268 used) in the following functions [macros]:
10271 BN_GF2m_sub [= BN_GF2m_add]
10272 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10273 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10274 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10276 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10277 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10278 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10279 BN_GF2m_cmp [= BN_ucmp]
10281 (Note that only the 'mod' functions are actually for fields GF(2^m).
10282 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10284 For some functions, an the irreducible polynomial defining a
10285 field can be given as an 'unsigned int[]' with strictly
10286 decreasing elements giving the indices of those bits that are set;
10287 i.e., p[] represents the polynomial
10288 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10290 p[0] > p[1] > ... > p[k] = 0.
10291 This applies to the following functions:
10294 BN_GF2m_mod_mul_arr
10295 BN_GF2m_mod_sqr_arr
10296 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10297 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10298 BN_GF2m_mod_exp_arr
10299 BN_GF2m_mod_sqrt_arr
10300 BN_GF2m_mod_solve_quad_arr
10304 Conversion can be performed by the following functions:
10309 bntest.c has additional tests for binary polynomial arithmetic.
10311 Two implementations for BN_GF2m_mod_div() are available.
10312 The default algorithm simply uses BN_GF2m_mod_inv() and
10313 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10314 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10315 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10317 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10319 * Add new error code 'ERR_R_DISABLED' that can be used when some
10320 functionality is disabled at compile-time.
10322 *Douglas Stebila <douglas.stebila@sun.com>*
10324 * Change default behaviour of 'openssl asn1parse' so that more
10325 information is visible when viewing, e.g., a certificate:
10327 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10328 mode the content of non-printable OCTET STRINGs is output in a
10329 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10330 avoid the appearance of a printable string.
10332 *Nils Larsch <nla@trustcenter.de>*
10334 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10336 EC_GROUP_set_asn1_flag()
10337 EC_GROUP_get_asn1_flag()
10338 EC_GROUP_set_point_conversion_form()
10339 EC_GROUP_get_point_conversion_form()
10340 These control ASN1 encoding details:
10341 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10342 has been set to OPENSSL_EC_NAMED_CURVE.
10343 - Points are encoded in uncompressed form by default; options for
10344 asn1_for are as for point2oct, namely
10345 POINT_CONVERSION_COMPRESSED
10346 POINT_CONVERSION_UNCOMPRESSED
10347 POINT_CONVERSION_HYBRID
10349 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10351 EC_GROUP_set_seed()
10352 EC_GROUP_get0_seed()
10353 EC_GROUP_get_seed_len()
10354 This is used only for ASN1 purposes (so far).
10356 *Nils Larsch <nla@trustcenter.de>*
10358 * Add 'field_type' member to EC_METHOD, which holds the NID
10359 of the appropriate field type OID. The new function
10360 EC_METHOD_get_field_type() returns this value.
10362 *Nils Larsch <nla@trustcenter.de>*
10365 EC_POINT_point2bn()
10366 EC_POINT_bn2point()
10367 EC_POINT_point2hex()
10368 EC_POINT_hex2point()
10369 providing useful interfaces to EC_POINT_point2oct() and
10370 EC_POINT_oct2point().
10372 *Nils Larsch <nla@trustcenter.de>*
10374 * Change internals of the EC library so that the functions
10375 EC_GROUP_set_generator()
10376 EC_GROUP_get_generator()
10377 EC_GROUP_get_order()
10378 EC_GROUP_get_cofactor()
10379 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10380 to methods, which would lead to unnecessary code duplication when
10381 adding different types of curves.
10383 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10385 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10386 arithmetic, and such that modified wNAFs are generated
10387 (which avoid length expansion in many cases).
10391 * Add a function EC_GROUP_check_discriminant() (defined via
10392 EC_METHOD) that verifies that the curve discriminant is non-zero.
10394 Add a function EC_GROUP_check() that makes some sanity tests
10395 on a EC_GROUP, its generator and order. This includes
10396 EC_GROUP_check_discriminant().
10398 *Nils Larsch <nla@trustcenter.de>*
10400 * Add ECDSA in new directory crypto/ecdsa/.
10402 Add applications 'openssl ecparam' and 'openssl ecdsa'
10403 (these are based on 'openssl dsaparam' and 'openssl dsa').
10405 ECDSA support is also included in various other files across the
10406 library. Most notably,
10407 - 'openssl req' now has a '-newkey ecdsa:file' option;
10408 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10409 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10410 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10411 them suitable for ECDSA where domain parameters must be
10412 extracted before the specific public key;
10413 - ECDSA engine support has been added.
10415 *Nils Larsch <nla@trustcenter.de>*
10417 * Include some named elliptic curves, and add OIDs from X9.62,
10418 SECG, and WAP/WTLS. Each curve can be obtained from the new
10420 EC_GROUP_new_by_curve_name(),
10421 and the list of available named curves can be obtained with
10422 EC_get_builtin_curves().
10423 Also add a 'curve_name' member to EC_GROUP objects, which can be
10425 EC_GROUP_set_curve_name()
10426 EC_GROUP_get_curve_name()
10428 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10430 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10431 was actually never needed) and in BN_mul(). The removal in BN_mul()
10432 required a small change in bn_mul_part_recursive() and the addition
10433 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10434 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10435 bn_sub_words() and bn_add_words() except they take arrays with
10440 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10442 * Cleanse PEM buffers before freeing them since they may contain
10445 *Benjamin Bennett <ben@psc.edu>*
10447 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10448 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10449 authentication-only ciphersuites.
10453 * Since AES128 and AES256 share a single mask bit in the logic of
10454 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10455 kludge to work properly if AES128 is available and AES256 isn't.
10459 * Expand security boundary to match 1.1.1 module.
10463 * Remove redundant features: hash file source, editing of test vectors
10464 modify fipsld to use external fips_premain.c signature.
10468 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10469 run algorithm test programs.
10473 * Make algorithm test programs more tolerant of whitespace.
10477 * Have SSL/TLS server implementation tolerate "mismatched" record
10478 protocol version while receiving ClientHello even if the
10479 ClientHello is fragmented. (The server can't insist on the
10480 particular protocol version it has chosen before the ServerHello
10481 message has informed the client about his choice.)
10485 * Load error codes if they are not already present instead of using a
10486 static variable. This allows them to be cleanly unloaded and reloaded.
10490 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10492 * Introduce limits to prevent malicious keys being able to
10493 cause a denial of service. [CVE-2006-2940][]
10495 *Steve Henson, Bodo Moeller*
10497 * Fix ASN.1 parsing of certain invalid structures that can result
10498 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10500 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10501 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10503 * Fix SSL client code which could crash if connecting to a
10504 malicious SSLv2 server. [CVE-2006-4343][]
10506 *Tavis Ormandy and Will Drewry, Google Security Team*
10508 * Change ciphersuite string processing so that an explicit
10509 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10510 will no longer include "AES128-SHA"), and any other similar
10511 ciphersuite (same bitmap) from *other* protocol versions (so that
10512 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10513 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10514 changes from 0.9.8b and 0.9.8d.
10518 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10520 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10521 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10523 * Change the Unix randomness entropy gathering to use poll() when
10524 possible instead of select(), since the latter has some
10525 undesirable limitations.
10527 *Darryl Miles via Richard Levitte and Bodo Moeller*
10529 * Disable rogue ciphersuites:
10531 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10532 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10533 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10535 The latter two were purportedly from
10536 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10539 Also deactivate the remaining ciphersuites from
10540 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10541 unofficial, and the ID has long expired.
10545 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10546 dual-core machines) and other potential thread-safety issues.
10550 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10552 * Adapt fipsld and the build system to link against the validated FIPS
10553 module in FIPS mode.
10557 * Fixes for VC++ 2005 build under Windows.
10561 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10562 from a Windows bash shell such as MSYS. It is autodetected from the
10563 "config" script when run from a VC++ environment. Modify standard VC++
10564 build to use fipscanister.o from the GNU make build.
10568 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10570 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10571 The value now differs depending on if you build for FIPS or not.
10572 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10573 safely run with a non-FIPSed libcrypto, as it may crash because of
10574 the difference induced by this change.
10578 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10580 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10581 (part of SSL_OP_ALL). This option used to disable the
10582 countermeasure against man-in-the-middle protocol-version
10583 rollback in the SSL 2.0 server implementation, which is a bad
10584 idea. [CVE-2005-2969][]
10586 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10587 for Information Security, National Institute of Advanced Industrial
10588 Science and Technology [AIST, Japan)]*
10590 * Minimal support for X9.31 signatures and PSS padding modes. This is
10591 mainly for FIPS compliance and not fully integrated at this stage.
10595 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10596 the exponentiation using a fixed-length exponent. (Otherwise,
10597 the information leaked through timing could expose the secret key
10598 after many signatures; cf. Bleichenbacher's attack on DSA with
10603 * Make a new fixed-window mod_exp implementation the default for
10604 RSA, DSA, and DH private-key operations so that the sequence of
10605 squares and multiplies and the memory access pattern are
10606 independent of the particular secret key. This will mitigate
10607 cache-timing and potential related attacks.
10609 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10610 and this is automatically used by BN_mod_exp_mont() if the new flag
10611 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10612 will use this BN flag for private exponents unless the flag
10613 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10614 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10616 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10618 * Change the client implementation for SSLv23_method() and
10619 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10620 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10621 (Previously, the SSL 2.0 backwards compatible Client Hello
10622 message format would be used even with SSL_OP_NO_SSLv2.)
10626 * Add support for smime-type MIME parameter in S/MIME messages which some
10631 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10632 a threadsafe manner. Modify rsa code to use new function and add calls
10633 to dsa and dh code (which had race conditions before).
10637 * Include the fixed error library code in the C error file definitions
10638 instead of fixing them up at runtime. This keeps the error code
10639 structures constant.
10643 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10645 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10648 * Fixes for newer kerberos headers. NB: the casts are needed because
10649 the 'length' field is signed on one version and unsigned on another
10650 with no (?) obvious way to tell the difference, without these VC++
10651 complains. Also the "definition" of FAR (blank) is no longer included
10652 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10653 some needed definitions.
10657 * Undo Cygwin change.
10661 * Added support for proxy certificates according to RFC 3820.
10662 Because they may be a security thread to unaware applications,
10663 they must be explicitly allowed in run-time. See
10664 docs/HOWTO/proxy_certificates.txt for further information.
10668 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10670 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10671 server and client random values. Previously
10672 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10673 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10675 This change has negligible security impact because:
10677 1. Server and client random values still have 24 bytes of pseudo random
10680 2. Server and client random values are sent in the clear in the initial
10683 3. The master secret is derived using the premaster secret (48 bytes in
10684 size for static RSA ciphersuites) as well as client server and random
10687 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10690 *Stephen Henson, reported by UK NISCC*
10692 * Use Windows randomness collection on Cygwin.
10696 * Fix hang in EGD/PRNGD query when communication socket is closed
10697 prematurely by EGD/PRNGD.
10699 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10701 * Prompt for pass phrases when appropriate for PKCS12 input format.
10705 * Back-port of selected performance improvements from development
10706 branch, as well as improved support for PowerPC platforms.
10710 * Add lots of checks for memory allocation failure, error codes to indicate
10711 failure and freeing up memory if a failure occurs.
10713 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10715 * Add new -passin argument to dgst.
10719 * Perform some character comparisons of different types in X509_NAME_cmp:
10720 this is needed for some certificates that re-encode DNs into UTF8Strings
10721 (in violation of RFC3280) and can't or won't issue name rollover
10726 * Make an explicit check during certificate validation to see that
10727 the CA setting in each certificate on the chain is correct. As a
10728 side effect always do the following basic checks on extensions,
10729 not just when there's an associated purpose to the check:
10731 - if there is an unhandled critical extension (unless the user
10732 has chosen to ignore this fault)
10733 - if the path length has been exceeded (if one is set at all)
10734 - that certain extensions fit the associated purpose (if one has
10739 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10741 * Avoid a race condition when CRLs are checked in a multi threaded
10742 environment. This would happen due to the reordering of the revoked
10743 entries during signature checking and serial number lookup. Now the
10744 encoding is cached and the serial number sort performed under a lock.
10745 Add new STACK function sk_is_sorted().
10749 * Add Delta CRL to the extension code.
10753 * Various fixes to s3_pkt.c so alerts are sent properly.
10755 *David Holmes <d.holmes@f5.com>*
10757 * Reduce the chances of duplicate issuer name and serial numbers (in
10758 violation of RFC3280) using the OpenSSL certificate creation utilities.
10759 This is done by creating a random 64 bit value for the initial serial
10760 number when a serial number file is created or when a self signed
10761 certificate is created using 'openssl req -x509'. The initial serial
10762 number file is created using 'openssl x509 -next_serial' in CA.pl
10763 rather than being initialized to 1.
10767 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10769 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10770 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10772 *Joe Orton, Steve Henson*
10774 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10777 *Joe Orton, Steve Henson*
10779 * Make it possible to have multiple active certificates with the same
10780 subject in the CA index file. This is done only if the keyword
10781 'unique_subject' is set to 'no' in the main CA section (default
10782 if 'CA_default') of the configuration file. The value is saved
10783 with the database itself in a separate index attribute file,
10784 named like the index file with '.attr' appended to the name.
10788 * X509 verify fixes. Disable broken certificate workarounds when
10789 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10790 keyUsage extension present. Don't accept CRLs with unhandled critical
10791 extensions: since verify currently doesn't process CRL extensions this
10792 rejects a CRL with *any* critical extensions. Add new verify error codes
10797 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10798 A clarification of RFC2560 will require the use of OCTET STRINGs and
10799 some implementations cannot handle the current raw format. Since OpenSSL
10800 copies and compares OCSP nonces as opaque blobs without any attempt at
10801 parsing them this should not create any compatibility issues.
10805 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10806 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10807 this HMAC (and other) operations are several times slower than OpenSSL
10812 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10814 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10816 * Use the correct content when signing type "other".
10820 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10822 * Fix various bugs revealed by running the NISCC test suite:
10824 Stop out of bounds reads in the ASN1 code when presented with
10825 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10827 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10829 If verify callback ignores invalid public key errors don't try to check
10830 certificate signature with the NULL public key.
10834 * New -ignore_err option in ocsp application to stop the server
10835 exiting on the first error in a request.
10839 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10840 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10845 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10846 extra data after the compression methods not only for TLS 1.0
10847 but also for SSL 3.0 (as required by the specification).
10849 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10851 * Change X509_certificate_type() to mark the key as exported/exportable
10852 when it's 512 *bits* long, not 512 bytes.
10856 * Change AES_cbc_encrypt() so it outputs exact multiple of
10857 blocks during encryption.
10861 * Various fixes to base64 BIO and non blocking I/O. On write
10862 flushes were not handled properly if the BIO retried. On read
10863 data was not being buffered properly and had various logic bugs.
10864 This also affects blocking I/O when the data being decoded is a
10869 * Various S/MIME bugfixes and compatibility changes:
10870 output correct application/pkcs7 MIME type if
10871 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10872 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10873 of files as .eml work). Correctly handle very long lines in MIME
10878 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10880 * Countermeasure against the Klima-Pokorny-Rosa extension of
10881 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10882 a protocol version number mismatch like a decryption error
10883 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10887 * Turn on RSA blinding by default in the default implementation
10888 to avoid a timing attack. Applications that don't want it can call
10889 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10890 They would be ill-advised to do so in most cases.
10892 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10894 * Change RSA blinding code so that it works when the PRNG is not
10895 seeded (in this case, the secret RSA exponent is abused as
10896 an unpredictable seed -- if it is not unpredictable, there
10897 is no point in blinding anyway). Make RSA blinding thread-safe
10898 by remembering the creator's thread ID in rsa->blinding and
10899 having all other threads use local one-time blinding factors
10900 (this requires more computation than sharing rsa->blinding, but
10901 avoids excessive locking; and if an RSA object is not shared
10902 between threads, blinding will still be very fast).
10906 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10907 ENGINE as defaults for all supported algorithms irrespective of
10908 the 'flags' parameter. 'flags' is now honoured, so applications
10909 should make sure they are passing it correctly.
10913 * Target "mingw" now allows native Windows code to be generated in
10914 the Cygwin environment as well as with the MinGW compiler.
10918 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10920 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10921 via timing by performing a MAC computation even if incorrect
10922 block cipher padding has been found. This is a countermeasure
10923 against active attacks where the attacker has to distinguish
10924 between bad padding and a MAC verification error. [CVE-2003-0078][]
10926 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10927 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10928 Martin Vuagnoux (EPFL, Ilion)*
10930 * Make the no-err option work as intended. The intention with no-err
10931 is not to have the whole error stack handling routines removed from
10932 libcrypto, it's only intended to remove all the function name and
10933 reason texts, thereby removing some of the footprint that may not
10934 be interesting if those errors aren't displayed anyway.
10936 NOTE: it's still possible for any application or module to have its
10937 own set of error texts inserted. The routines are there, just not
10938 used by default when no-err is given.
10942 * Add support for FreeBSD on IA64.
10944 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10946 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10947 Kerberos function mit_des_cbc_cksum(). Before this change,
10948 the value returned by DES_cbc_cksum() was like the one from
10949 mit_des_cbc_cksum(), except the bytes were swapped.
10951 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10953 * Allow an application to disable the automatic SSL chain building.
10954 Before this a rather primitive chain build was always performed in
10955 ssl3_output_cert_chain(): an application had no way to send the
10956 correct chain if the automatic operation produced an incorrect result.
10958 Now the chain builder is disabled if either:
10960 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10962 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10964 The reasoning behind this is that an application would not want the
10965 auto chain building to take place if extra chain certificates are
10966 present and it might also want a means of sending no additional
10967 certificates (for example the chain has two certificates and the
10972 * Add the possibility to build without the ENGINE framework.
10974 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10976 * Under Win32 gmtime() can return NULL: check return value in
10977 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10981 * DSA routines: under certain error conditions uninitialized BN objects
10982 could be freed. Solution: make sure initialization is performed early
10983 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10984 Nils Larsch <nla@trustcenter.de> via PR#459)
10988 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10989 checked on reconnect on the client side, therefore session resumption
10990 could still fail with a "ssl session id is different" error. This
10991 behaviour is masked when SSL_OP_ALL is used due to
10992 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10993 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10994 followup to PR #377.
10998 * IA-32 assembler support enhancements: unified ELF targets, support
10999 for SCO/Caldera platforms, fix for Cygwin shared build.
11003 * Add support for FreeBSD on sparc64. As a consequence, support for
11004 FreeBSD on non-x86 processors is separate from x86 processors on
11005 the config script, much like the NetBSD support.
11007 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11009 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11011 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11014 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11015 code (06) was taken as the first octet of the session ID and the last
11016 octet was ignored consequently. As a result SSLv2 client side session
11017 caching could not have worked due to the session ID mismatch between
11019 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11024 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11025 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11030 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11031 seems that in spite of existing for more than a year, many application
11032 author have done nothing to provide the necessary callbacks, which
11033 means that this particular engine will not work properly anywhere.
11034 This is a very unfortunate situation which forces us, in the name
11035 of usability, to give the hw_ncipher.c a static lock, which is part
11037 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11038 appear in 0.9.8 or later. We EXPECT application authors to have
11039 dealt properly with this when 0.9.8 is released (unless we actually
11040 make such changes in the libcrypto locking code that changes will
11041 have to be made anyway).
11045 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11046 octets have been read, EOF or an error occurs. Without this change
11047 some truncated ASN1 structures will not produce an error.
11051 * Disable Heimdal support, since it hasn't been fully implemented.
11052 Still give the possibility to force the use of Heimdal, but with
11053 warnings and a request that patches get sent to openssl-dev.
11057 * Add the VC-CE target, introduce the WINCE sysname, and add
11058 INSTALL.WCE and appropriate conditionals to make it build.
11060 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11062 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11063 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11064 edit numbers of the version.
11066 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11068 * Introduce safe string copy and catenation functions
11069 (BUF_strlcpy() and BUF_strlcat()).
11071 *Ben Laurie (CHATS) and Richard Levitte*
11073 * Avoid using fixed-size buffers for one-line DNs.
11075 *Ben Laurie (CHATS)*
11077 * Add BUF_MEM_grow_clean() to avoid information leakage when
11078 resizing buffers containing secrets, and use where appropriate.
11080 *Ben Laurie (CHATS)*
11082 * Avoid using fixed size buffers for configuration file location.
11084 *Ben Laurie (CHATS)*
11086 * Avoid filename truncation for various CA files.
11088 *Ben Laurie (CHATS)*
11090 * Use sizeof in preference to magic numbers.
11092 *Ben Laurie (CHATS)*
11094 * Avoid filename truncation in cert requests.
11096 *Ben Laurie (CHATS)*
11098 * Add assertions to check for (supposedly impossible) buffer
11101 *Ben Laurie (CHATS)*
11103 * Don't cache truncated DNS entries in the local cache (this could
11104 potentially lead to a spoofing attack).
11106 *Ben Laurie (CHATS)*
11108 * Fix various buffers to be large enough for hex/decimal
11109 representations in a platform independent manner.
11111 *Ben Laurie (CHATS)*
11113 * Add CRYPTO_realloc_clean() to avoid information leakage when
11114 resizing buffers containing secrets, and use where appropriate.
11116 *Ben Laurie (CHATS)*
11118 * Add BIO_indent() to avoid much slightly worrying code to do
11121 *Ben Laurie (CHATS)*
11123 * Convert sprintf()/BIO_puts() to BIO_printf().
11125 *Ben Laurie (CHATS)*
11127 * buffer_gets() could terminate with the buffer only half
11130 *Ben Laurie (CHATS)*
11132 * Add assertions to prevent user-supplied crypto functions from
11133 overflowing internal buffers by having large block sizes, etc.
11135 *Ben Laurie (CHATS)*
11137 * New OPENSSL_assert() macro (similar to assert(), but enabled
11140 *Ben Laurie (CHATS)*
11142 * Eliminate unused copy of key in RC4.
11144 *Ben Laurie (CHATS)*
11146 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11148 *Ben Laurie (CHATS)*
11150 * Fix off-by-one error in EGD path.
11152 *Ben Laurie (CHATS)*
11154 * If RANDFILE path is too long, ignore instead of truncating.
11156 *Ben Laurie (CHATS)*
11158 * Eliminate unused and incorrectly sized X.509 structure
11161 *Ben Laurie (CHATS)*
11163 * Eliminate unused and dangerous function knumber().
11165 *Ben Laurie (CHATS)*
11167 * Eliminate unused and dangerous structure, KSSL_ERR.
11169 *Ben Laurie (CHATS)*
11171 * Protect against overlong session ID context length in an encoded
11172 session object. Since these are local, this does not appear to be
11175 *Ben Laurie (CHATS)*
11177 * Change from security patch (see 0.9.6e below) that did not affect
11178 the 0.9.6 release series:
11180 Remote buffer overflow in SSL3 protocol - an attacker could
11181 supply an oversized master key in Kerberos-enabled versions.
11184 *Ben Laurie (CHATS)*
11186 * Change the SSL kerb5 codes to match RFC 2712.
11190 * Make -nameopt work fully for req and add -reqopt switch.
11192 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11194 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11196 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11198 * Make sure tests can be performed even if the corresponding algorithms
11199 have been removed entirely. This was also the last step to make
11200 OpenSSL compilable with DJGPP under all reasonable conditions.
11202 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11204 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11205 to allow version independent disabling of normally unselected ciphers,
11206 which may be activated as a side-effect of selecting a single cipher.
11208 (E.g., cipher list string "RSA" enables ciphersuites that are left
11209 out of "ALL" because they do not provide symmetric encryption.
11210 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11212 *Lutz Jaenicke, Bodo Moeller*
11214 * Add appropriate support for separate platform-dependent build
11215 directories. The recommended way to make a platform-dependent
11216 build directory is the following (tested on Linux), maybe with
11219 # Place yourself outside of the OpenSSL source tree. In
11220 # this example, the environment variable OPENSSL_SOURCE
11221 # is assumed to contain the absolute OpenSSL source directory.
11222 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11223 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11224 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11225 mkdir -p `dirname $F`
11226 ln -s $OPENSSL_SOURCE/$F $F
11229 To be absolutely sure not to disturb the source tree, a "make clean"
11230 is a good thing. If it isn't successful, don't worry about it,
11231 it probably means the source directory is very clean.
11235 * Make sure any ENGINE control commands make local copies of string
11236 pointers passed to them whenever necessary. Otherwise it is possible
11237 the caller may have overwritten (or deallocated) the original string
11238 data when a later ENGINE operation tries to use the stored values.
11240 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11242 * Improve diagnostics in file reading and command-line digests.
11244 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11246 * Add AES modes CFB and OFB to the object database. Correct an
11247 error in AES-CFB decryption.
11251 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11252 allows existing EVP_CIPHER_CTX structures to be reused after
11253 calling `EVP_*Final()`. This behaviour is used by encryption
11254 BIOs and some applications. This has the side effect that
11255 applications must explicitly clean up cipher contexts with
11256 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11260 * Check the values of dna and dnb in bn_mul_recursive before calling
11261 bn_mul_comba (a non zero value means the a or b arrays do not contain
11262 n2 elements) and fallback to bn_mul_normal if either is not zero.
11266 * Fix escaping of non-ASCII characters when using the -subj option
11267 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11271 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11272 form for "surname", serialNumber has no short form.
11273 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11274 therefore remove "mail" short name for "internet 7".
11275 The OID for unique identifiers in X509 certificates is
11276 x500UniqueIdentifier, not uniqueIdentifier.
11277 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11281 * Add an "init" command to the ENGINE config module and auto initialize
11282 ENGINEs. Without any "init" command the ENGINE will be initialized
11283 after all ctrl commands have been executed on it. If init=1 the
11284 ENGINE is initialized at that point (ctrls before that point are run
11285 on the uninitialized ENGINE and after on the initialized one). If
11286 init=0 then the ENGINE will not be initialized at all.
11290 * Fix the 'app_verify_callback' interface so that the user-defined
11291 argument is actually passed to the callback: In the
11292 SSL_CTX_set_cert_verify_callback() prototype, the callback
11293 declaration has been changed from
11296 int (*cb)(X509_STORE_CTX *,void *);
11297 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11298 i=s->ctx->app_verify_callback(&ctx)
11299 has been changed into
11300 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11302 To update applications using SSL_CTX_set_cert_verify_callback(),
11303 a dummy argument can be added to their callback functions.
11305 *D. K. Smetters <smetters@parc.xerox.com>*
11307 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11309 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11311 * Add and OPENSSL_LOAD_CONF define which will cause
11312 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11313 This allows older applications to transparently support certain
11314 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11315 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11316 load the config file and OPENSSL_add_all_algorithms_conf() which will
11317 always load it have also been added.
11321 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11322 Adjust NIDs and EVP layer.
11324 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11326 * Config modules support in openssl utility.
11328 Most commands now load modules from the config file,
11329 though in a few (such as version) this isn't done
11330 because it couldn't be used for anything.
11332 In the case of ca and req the config file used is
11333 the same as the utility itself: that is the -config
11334 command line option can be used to specify an
11339 * Move default behaviour from OPENSSL_config(). If appname is NULL
11340 use "openssl_conf" if filename is NULL use default openssl config file.
11344 * Add an argument to OPENSSL_config() to allow the use of an alternative
11345 config section name. Add a new flag to tolerate a missing config file
11346 and move code to CONF_modules_load_file().
11350 * Support for crypto accelerator cards from Accelerated Encryption
11351 Processing, www.aep.ie. (Use engine 'aep')
11352 The support was copied from 0.9.6c [engine] and adapted/corrected
11353 to work with the new engine framework.
11355 *AEP Inc. and Richard Levitte*
11357 * Support for SureWare crypto accelerator cards from Baltimore
11358 Technologies. (Use engine 'sureware')
11359 The support was copied from 0.9.6c [engine] and adapted
11360 to work with the new engine framework.
11364 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11365 make the newer ENGINE framework commands for the CHIL engine work.
11367 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11369 * Make it possible to produce shared libraries on ReliantUNIX.
11371 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11373 * Add the configuration target debug-linux-ppro.
11374 Make 'openssl rsa' use the general key loading routines
11375 implemented in apps.c, and make those routines able to
11376 handle the key format FORMAT_NETSCAPE and the variant
11379 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11381 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11383 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11385 * Add -keyform to rsautl, and document -engine.
11387 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11389 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11390 BIO_R_NO_SUCH_FILE error code rather than the generic
11391 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11395 * Add new functions
11396 ERR_peek_last_error
11397 ERR_peek_last_error_line
11398 ERR_peek_last_error_line_data.
11399 These are similar to
11401 ERR_peek_error_line
11402 ERR_peek_error_line_data,
11403 but report on the latest error recorded rather than the first one
11404 still in the error queue.
11406 *Ben Laurie, Bodo Moeller*
11408 * default_algorithms option in ENGINE config module. This allows things
11410 default_algorithms = ALL
11411 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11415 * Preliminary ENGINE config module.
11419 * New experimental application configuration code.
11423 * Change the AES code to follow the same name structure as all other
11424 symmetric ciphers, and behave the same way. Move everything to
11425 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11427 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11429 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11431 *Ben Laurie and Theo de Raadt*
11433 * Add option to output public keys in req command.
11435 *Massimiliano Pala madwolf@openca.org*
11437 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11438 (up to about 10% better than before for P-192 and P-224).
11442 * New functions/macros
11444 SSL_CTX_set_msg_callback(ctx, cb)
11445 SSL_CTX_set_msg_callback_arg(ctx, arg)
11446 SSL_set_msg_callback(ssl, cb)
11447 SSL_set_msg_callback_arg(ssl, arg)
11449 to request calling a callback function
11451 void cb(int write_p, int version, int content_type,
11452 const void *buf, size_t len, SSL *ssl, void *arg)
11454 whenever a protocol message has been completely received
11455 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11456 protocol version according to which the SSL library interprets
11457 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11458 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11459 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11460 specification (change_cipher_spec(20), alert(21), handshake(22)).
11461 'buf' and 'len' point to the actual message, 'ssl' to the
11462 SSL object, and 'arg' is the application-defined value set by
11463 SSL[_CTX]_set_msg_callback_arg().
11465 'openssl s_client' and 'openssl s_server' have new '-msg' options
11466 to enable a callback that displays all protocol messages.
11470 * Change the shared library support so shared libraries are built as
11471 soon as the corresponding static library is finished, and thereby get
11472 openssl and the test programs linked against the shared library.
11473 This still only happens when the keyword "shard" has been given to
11474 the configuration scripts.
11476 NOTE: shared library support is still an experimental thing, and
11477 backward binary compatibility is still not guaranteed.
11479 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11481 * Add support for Subject Information Access extension.
11483 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11485 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11486 additional bytes when new memory had to be allocated, not just
11487 when reusing an existing buffer.
11491 * New command line and configuration option 'utf8' for the req command.
11492 This allows field values to be specified as UTF8 strings.
11496 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11497 runs for the former and machine-readable output for the latter.
11501 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11502 of the e-mail address in the DN (i.e., it will go into a certificate
11503 extension only). The new configuration file option 'email_in_dn = no'
11504 has the same effect.
11506 *Massimiliano Pala madwolf@openca.org*
11508 * Change all functions with names starting with `des_` to be starting
11509 with `DES_` instead. Add wrappers that are compatible with libdes,
11510 but are named `_ossl_old_des_*`. Finally, add macros that map the
11511 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11512 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11513 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11516 Since we provide two compatibility mappings, the user needs to
11517 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11518 compatibility is desired. The default (i.e., when that macro
11519 isn't defined) is OpenSSL 0.9.6c compatibility.
11521 There are also macros that enable and disable the support of old
11522 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11523 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11524 are defined, the default will apply: to support the old des routines.
11526 In either case, one must include openssl/des.h to get the correct
11527 definitions. Do not try to just include openssl/des_old.h, that
11530 NOTE: This is a major break of an old API into a new one. Software
11531 authors are encouraged to switch to the `DES_` style functions. Some
11532 time in the future, des_old.h and the libdes compatibility functions
11533 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11534 default), and then completely removed.
11538 * Test for certificates which contain unsupported critical extensions.
11539 If such a certificate is found during a verify operation it is
11540 rejected by default: this behaviour can be overridden by either
11541 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11542 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11543 X509_supported_extension() has also been added which returns 1 if a
11544 particular extension is supported.
11548 * Modify the behaviour of EVP cipher functions in similar way to digests
11549 to retain compatibility with existing code.
11553 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11554 compatibility with existing code. In particular the 'ctx' parameter does
11555 not have to be to be initialized before the call to EVP_DigestInit() and
11556 it is tidied up after a call to EVP_DigestFinal(). New function
11557 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11558 EVP_MD_CTX_copy() changed to not require the destination to be
11559 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11560 requires the destination to be valid.
11562 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11563 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11567 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11568 so that complete 'Handshake' protocol structures are kept in memory
11569 instead of overwriting 'msg_type' and 'length' with 'body' data.
11573 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11575 *Massimo Santin via Richard Levitte*
11577 * Major restructuring to the underlying ENGINE code. This includes
11578 reduction of linker bloat, separation of pure "ENGINE" manipulation
11579 (initialisation, etc) from functionality dealing with implementations
11580 of specific crypto interfaces. This change also introduces integrated
11581 support for symmetric ciphers and digest implementations - so ENGINEs
11582 can now accelerate these by providing EVP_CIPHER and EVP_MD
11583 implementations of their own. This is detailed in crypto/engine/README
11584 as it couldn't be adequately described here. However, there are a few
11585 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11586 were changed in the original introduction of ENGINE code have now
11587 reverted back - the hooking from this code to ENGINE is now a good
11588 deal more passive and at run-time, operations deal directly with
11589 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11590 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11591 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11592 they were not being used by the framework as there is no concept of a
11593 BIGNUM_METHOD and they could not be generalised to the new
11594 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11595 ENGINE_cpy() has been removed as it cannot be consistently defined in
11600 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11604 * Change mkdef.pl to sort symbols that get the same entry number,
11605 and make sure the automatically generated functions `ERR_load_*`
11606 become part of libeay.num as well.
11610 * New function SSL_renegotiate_pending(). This returns true once
11611 renegotiation has been requested (either SSL_renegotiate() call
11612 or HelloRequest/ClientHello received from the peer) and becomes
11613 false once a handshake has been completed.
11614 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11615 sends a HelloRequest, but does not ensure that a handshake takes
11616 place. SSL_renegotiate_pending() is useful for checking if the
11617 client has followed the request.)
11621 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11622 By default, clients may request session resumption even during
11623 renegotiation (if session ID contexts permit); with this option,
11624 session resumption is possible only in the first handshake.
11626 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11627 more bits available for options that should not be part of
11628 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11632 * Add some demos for certificate and certificate request creation.
11636 * Make maximum certificate chain size accepted from the peer application
11637 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11638 "Douglas E. Engert" <deengert@anl.gov>.
11642 * Add support for shared libraries for Unixware-7
11643 (Boyd Lynn Gerber <gerberb@zenez.com>).
11647 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11648 be done prior to destruction. Use this to unload error strings from
11649 ENGINEs that load their own error strings. NB: This adds two new API
11650 functions to "get" and "set" this destroy handler in an ENGINE.
11654 * Alter all existing ENGINE implementations (except "openssl" and
11655 "openbsd") to dynamically instantiate their own error strings. This
11656 makes them more flexible to be built both as statically-linked ENGINEs
11657 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11658 Also, add stub code to each that makes building them as self-contained
11659 shared-libraries easier (see README.ENGINE).
11663 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11664 implementations into applications that are completely implemented in
11665 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11666 commands that can be used to configure what shared-library to load and
11667 to control aspects of the way it is handled. Also, made an update to
11668 the README.ENGINE file that brings its information up-to-date and
11669 provides some information and instructions on the "dynamic" ENGINE
11670 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11674 * Make it possible to unload ranges of ERR strings with a new
11675 "ERR_unload_strings" function.
11679 * Add a copy() function to EVP_MD.
11683 * Make EVP_MD routines take a context pointer instead of just the
11684 md_data void pointer.
11688 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11689 that the digest can only process a single chunk of data
11690 (typically because it is provided by a piece of
11691 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11692 is only going to provide a single chunk of data, and hence the
11693 framework needn't accumulate the data for oneshot drivers.
11697 * As with "ERR", make it possible to replace the underlying "ex_data"
11698 functions. This change also alters the storage and management of global
11699 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11700 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11701 index counters. The API functions that use this state have been changed
11702 to take a "class_index" rather than pointers to the class's local STACK
11703 and counter, and there is now an API function to dynamically create new
11704 classes. This centralisation allows us to (a) plug a lot of the
11705 thread-safety problems that existed, and (b) makes it possible to clean
11706 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11707 such data would previously have always leaked in application code and
11708 workarounds were in place to make the memory debugging turn a blind eye
11709 to it. Application code that doesn't use this new function will still
11710 leak as before, but their memory debugging output will announce it now
11711 rather than letting it slide.
11713 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11714 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11715 has a return value to indicate success or failure.
11719 * Make it possible to replace the underlying "ERR" functions such that the
11720 global state (2 LHASH tables and 2 locks) is only used by the "default"
11721 implementation. This change also adds two functions to "get" and "set"
11722 the implementation prior to it being automatically set the first time
11723 any other ERR function takes place. Ie. an application can call "get",
11724 pass the return value to a module it has just loaded, and that module
11725 can call its own "set" function using that value. This means the
11726 module's "ERR" operations will use (and modify) the error state in the
11727 application and not in its own statically linked copy of OpenSSL code.
11731 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11732 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11733 the operation, and provides a more encapsulated way for external code
11734 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11735 to use these functions rather than manually incrementing the counts.
11737 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11741 * Add EVP test program.
11745 * Add symmetric cipher support to ENGINE. Expect the API to change!
11749 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11750 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11751 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11752 These allow a CRL to be built without having to access X509_CRL fields
11753 directly. Modify 'ca' application to use new functions.
11757 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11758 bug workarounds. Rollback attack detection is a security feature.
11759 The problem will only arise on OpenSSL servers when TLSv1 is not
11760 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11761 Software authors not wanting to support TLSv1 will have special reasons
11762 for their choice and can explicitly enable this option.
11764 *Bodo Moeller, Lutz Jaenicke*
11766 * Rationalise EVP so it can be extended: don't include a union of
11767 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11768 (similar to those existing for EVP_CIPHER_CTX).
11773 EVP_MD_CTX_init(&md); /* new function call */
11774 EVP_DigestInit(&md, EVP_sha1());
11775 EVP_DigestUpdate(&md, in, len);
11776 EVP_DigestFinal(&md, out, NULL);
11777 EVP_MD_CTX_cleanup(&md); /* new function call */
11781 * Make DES key schedule conform to the usual scheme, as well as
11782 correcting its structure. This means that calls to DES functions
11783 now have to pass a pointer to a des_key_schedule instead of a
11784 plain des_key_schedule (which was actually always a pointer
11787 des_key_schedule ks;
11789 des_set_key_checked(..., &ks);
11790 des_ncbc_encrypt(..., &ks, ...);
11792 (Note that a later change renames 'des_...' into 'DES_...'.)
11796 * Initial reduction of linker bloat: the use of some functions, such as
11797 PEM causes large amounts of unused functions to be linked in due to
11798 poor organisation. For example pem_all.c contains every PEM function
11799 which has a knock on effect of linking in large amounts of (unused)
11800 ASN1 code. Grouping together similar functions and splitting unrelated
11801 functions prevents this.
11805 * Cleanup of EVP macros.
11809 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11810 correct `_ecb suffix`.
11814 * Add initial OCSP responder support to ocsp application. The
11815 revocation information is handled using the text based index
11816 use by the ca application. The responder can either handle
11817 requests generated internally, supplied in files (for example
11818 via a CGI script) or using an internal minimal server.
11822 * Add configuration choices to get zlib compression for TLS.
11826 * Changes to Kerberos SSL for RFC 2712 compliance:
11827 1. Implemented real KerberosWrapper, instead of just using
11828 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11829 2. Implemented optional authenticator field of KerberosWrapper.
11831 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11832 and authenticator structs; see crypto/krb5/.
11834 Generalized Kerberos calls to support multiple Kerberos libraries.
11835 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11836 via Richard Levitte*
11838 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11839 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11840 values for each of the key sizes rather than having just
11841 parameters (and 'speed' generating keys each time).
11845 * Speed up EVP routines.
11848 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11849 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11850 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11851 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11853 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11854 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11855 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11858 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11860 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11864 * Added the OS2-EMX target.
11866 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11868 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11869 to support NCONF routines in extension code. New function CONF_set_nconf()
11870 to allow functions which take an NCONF to also handle the old LHASH
11871 structure: this means that the old CONF compatible routines can be
11872 retained (in particular wrt extensions) without having to duplicate the
11873 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11877 * Enhance the general user interface with mechanisms for inner control
11878 and with possibilities to have yes/no kind of prompts.
11882 * Change all calls to low level digest routines in the library and
11883 applications to use EVP. Add missing calls to HMAC_cleanup() and
11884 don't assume HMAC_CTX can be copied using memcpy().
11886 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11888 * Add the possibility to control engines through control names but with
11889 arbitrary arguments instead of just a string.
11890 Change the key loaders to take a UI_METHOD instead of a callback
11891 function pointer. NOTE: this breaks binary compatibility with earlier
11892 versions of OpenSSL [engine].
11893 Adapt the nCipher code for these new conditions and add a card insertion
11898 * Enhance the general user interface with mechanisms to better support
11899 dialog box interfaces, application-defined prompts, the possibility
11900 to use defaults (for example default passwords from somewhere else)
11901 and interrupts/cancellations.
11905 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11906 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11910 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11911 tidy up some unnecessarily weird code in 'sk_new()').
11913 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11915 * Change the key loading routines for ENGINEs to use the same kind
11916 callback (pem_password_cb) as all other routines that need this
11921 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11922 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11923 than this minimum value is recommended.
11927 * New random seeder for OpenVMS, using the system process statistics
11928 that are easily reachable.
11932 * Windows apparently can't transparently handle global
11933 variables defined in DLLs. Initialisations such as:
11935 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11937 won't compile. This is used by the any applications that need to
11938 declare their own ASN1 modules. This was fixed by adding the option
11939 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11940 needed for static libraries under Win32.
11944 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11945 setting of purpose and trust fields. New X509_STORE trust and
11946 purpose functions and tidy up setting in other SSL functions.
11950 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11951 structure. These are inherited by X509_STORE_CTX when it is
11952 initialised. This allows various defaults to be set in the
11953 X509_STORE structure (such as flags for CRL checking and custom
11954 purpose or trust settings) for functions which only use X509_STORE_CTX
11955 internally such as S/MIME.
11957 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11958 trust settings if they are not set in X509_STORE. This allows X509_STORE
11959 purposes and trust (in S/MIME for example) to override any set by default.
11961 Add command line options for CRL checking to smime, s_client and s_server
11966 * Initial CRL based revocation checking. If the CRL checking flag(s)
11967 are set then the CRL is looked up in the X509_STORE structure and
11968 its validity and signature checked, then if the certificate is found
11969 in the CRL the verify fails with a revoked error.
11971 Various new CRL related callbacks added to X509_STORE_CTX structure.
11973 Command line options added to 'verify' application to support this.
11975 This needs some additional work, such as being able to handle multiple
11976 CRLs with different times, extension based lookup (rather than just
11977 by subject name) and ultimately more complete V2 CRL extension
11982 * Add a general user interface API (crypto/ui/). This is designed
11983 to replace things like des_read_password and friends (backward
11984 compatibility functions using this new API are provided).
11985 The purpose is to remove prompting functions from the DES code
11986 section as well as provide for prompting through dialog boxes in
11987 a window system and the like.
11991 * Add "ex_data" support to ENGINE so implementations can add state at a
11992 per-structure level rather than having to store it globally.
11996 * Make it possible for ENGINE structures to be copied when retrieved by
11997 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11998 This causes the "original" ENGINE structure to act like a template,
11999 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12000 operational state can be localised to each ENGINE structure, despite the
12001 fact they all share the same "methods". New ENGINE structures returned in
12002 this case have no functional references and the return value is the single
12003 structural reference. This matches the single structural reference returned
12004 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12009 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12010 needs to match any other type at all we need to manually clear the
12015 * Changes to the "openssl engine" utility to include;
12016 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12017 about an ENGINE's available control commands.
12018 - executing control commands from command line arguments using the
12019 '-pre' and '-post' switches. '-post' is only used if '-t' is
12020 specified and the ENGINE is successfully initialised. The syntax for
12021 the individual commands are colon-separated, for example;
12022 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12026 * New dynamic control command support for ENGINEs. ENGINEs can now
12027 declare their own commands (numbers), names (strings), descriptions,
12028 and input types for run-time discovery by calling applications. A
12029 subset of these commands are implicitly classed as "executable"
12030 depending on their input type, and only these can be invoked through
12031 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12032 can be based on user input, config files, etc). The distinction is
12033 that "executable" commands cannot return anything other than a boolean
12034 result and can only support numeric or string input, whereas some
12035 discoverable commands may only be for direct use through
12036 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12037 pointers, or other custom uses. The "executable" commands are to
12038 support parameterisations of ENGINE behaviour that can be
12039 unambiguously defined by ENGINEs and used consistently across any
12040 OpenSSL-based application. Commands have been added to all the
12041 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12042 control over shared-library paths without source code alterations.
12046 * Changed all ENGINE implementations to dynamically allocate their
12047 ENGINEs rather than declaring them statically. Apart from this being
12048 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12049 this also allows the implementations to compile without using the
12050 internal engine_int.h header.
12054 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12055 'const' value. Any code that should be able to modify a RAND_METHOD
12056 should already have non-const pointers to it (ie. they should only
12057 modify their own ones).
12061 * Made a variety of little tweaks to the ENGINE code.
12062 - "atalla" and "ubsec" string definitions were moved from header files
12063 to C code. "nuron" string definitions were placed in variables
12064 rather than hard-coded - allowing parameterisation of these values
12065 later on via ctrl() commands.
12066 - Removed unused "#if 0"'d code.
12067 - Fixed engine list iteration code so it uses ENGINE_free() to release
12068 structural references.
12069 - Constified the RAND_METHOD element of ENGINE structures.
12070 - Constified various get/set functions as appropriate and added
12071 missing functions (including a catch-all ENGINE_cpy that duplicates
12072 all ENGINE values onto a new ENGINE except reference counts/state).
12073 - Removed NULL parameter checks in get/set functions. Setting a method
12074 or function to NULL is a way of cancelling out a previously set
12075 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12076 and doesn't justify the extra error symbols and code.
12077 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12078 flags from engine_int.h to engine.h.
12079 - Changed prototypes for ENGINE handler functions (init(), finish(),
12080 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12084 * Implement binary inversion algorithm for BN_mod_inverse in addition
12085 to the algorithm using long division. The binary algorithm can be
12086 used only if the modulus is odd. On 32-bit systems, it is faster
12087 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12088 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12089 up to 450 bits. In 64-bit environments, the binary algorithm
12090 appears to be advantageous for much longer moduli; here we use it
12091 for moduli up to 2048 bits.
12095 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12096 could not support the combine flag in choice fields.
12100 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12101 extensions from a certificate request to the certificate.
12105 * Allow multiple 'certopt' and 'nameopt' options to be separated
12106 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12107 file: this allows the display of the certificate about to be
12108 signed to be customised, to allow certain fields to be included
12109 or excluded and extension details. The old system didn't display
12110 multicharacter strings properly, omitted fields not in the policy
12111 and couldn't display additional details such as extensions.
12115 * Function EC_POINTs_mul for multiple scalar multiplication
12116 of an arbitrary number of elliptic curve points
12117 \sum scalars[i]*points[i],
12118 optionally including the generator defined for the EC_GROUP:
12119 scalar*generator + \sum scalars[i]*points[i].
12121 EC_POINT_mul is a simple wrapper function for the typical case
12122 that the point list has just one item (besides the optional
12127 * First EC_METHODs for curves over GF(p):
12129 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12130 operations and provides various method functions that can also
12131 operate with faster implementations of modular arithmetic.
12133 EC_GFp_mont_method() reuses most functions that are part of
12134 EC_GFp_simple_method, but uses Montgomery arithmetic.
12136 *Bodo Moeller; point addition and point doubling
12137 implementation directly derived from source code provided by
12138 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12140 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12141 crypto/ec/ec_lib.c):
12143 Curves are EC_GROUP objects (with an optional group generator)
12144 based on EC_METHODs that are built into the library.
12146 Points are EC_POINT objects based on EC_GROUP objects.
12148 Most of the framework would be able to handle curves over arbitrary
12149 finite fields, but as there are no obvious types for fields other
12150 than GF(p), some functions are limited to that for now.
12154 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12155 that the file contains a complete HTTP response.
12159 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12160 change the def and num file printf format specifier from "%-40sXXX"
12161 to "%-39s XXX". The latter will always guarantee a space after the
12162 field while the former will cause them to run together if the field
12163 is 40 of more characters long.
12167 * Constify the cipher and digest 'method' functions and structures
12168 and modify related functions to take constant EVP_MD and EVP_CIPHER
12173 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12174 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12178 * Modify `EVP_Digest*()` routines so they now return values. Although the
12179 internal software routines can never fail additional hardware versions
12184 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12186 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12187 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12190 ERR_R_NESTED_ASN1_ERROR
12192 ERR_R_MISSING_ASN1_EOS
12193 were 4 .. 9, conflicting with
12194 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12196 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12197 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12199 Add new error code 'ERR_R_INTERNAL_ERROR'.
12203 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12208 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12209 sets the subject name for a new request or supersedes the
12210 subject name in a given request. Formats that can be parsed are
12211 'CN=Some Name, OU=myOU, C=IT'
12213 'CN=Some Name/OU=myOU/C=IT'.
12215 Add options '-batch' and '-verbose' to 'openssl req'.
12217 *Massimiliano Pala <madwolf@hackmasters.net>*
12219 * Introduce the possibility to access global variables through
12220 functions on platform were that's the best way to handle exporting
12221 global variables in shared libraries. To enable this functionality,
12222 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12223 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12224 is normally done by Configure or something similar).
12226 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12227 in the source file (foo.c) like this:
12229 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12230 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12232 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12233 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12235 OPENSSL_DECLARE_GLOBAL(int,foo);
12236 #define foo OPENSSL_GLOBAL_REF(foo)
12237 OPENSSL_DECLARE_GLOBAL(double,bar);
12238 #define bar OPENSSL_GLOBAL_REF(bar)
12240 The #defines are very important, and therefore so is including the
12241 header file everywhere where the defined globals are used.
12243 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12244 of ASN.1 items, but that structure is a bit different.
12246 The largest change is in util/mkdef.pl which has been enhanced with
12247 better and easier to understand logic to choose which symbols should
12248 go into the Windows .def files as well as a number of fixes and code
12249 cleanup (among others, algorithm keywords are now sorted
12250 lexicographically to avoid constant rewrites).
12254 * In BN_div() keep a copy of the sign of 'num' before writing the
12255 result to 'rm' because if rm==num the value will be overwritten
12256 and produce the wrong result if 'num' is negative: this caused
12257 problems with BN_mod() and BN_nnmod().
12261 * Function OCSP_request_verify(). This checks the signature on an
12262 OCSP request and verifies the signer certificate. The signer
12263 certificate is just checked for a generic purpose and OCSP request
12268 * Add OCSP_check_validity() function to check the validity of OCSP
12269 responses. OCSP responses are prepared in real time and may only
12270 be a few seconds old. Simply checking that the current time lies
12271 between thisUpdate and nextUpdate max reject otherwise valid responses
12272 caused by either OCSP responder or client clock inaccuracy. Instead
12273 we allow thisUpdate and nextUpdate to fall within a certain period of
12274 the current time. The age of the response can also optionally be
12275 checked. Two new options -validity_period and -status_age added to
12280 * If signature or public key algorithm is unrecognized print out its
12281 OID rather that just UNKNOWN.
12285 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12286 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12287 ID to be generated from the issuer certificate alone which can then be
12288 passed to OCSP_id_issuer_cmp().
12292 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12293 ASN1 modules to export functions returning ASN1_ITEM pointers
12294 instead of the ASN1_ITEM structures themselves. This adds several
12295 new macros which allow the underlying ASN1 function/structure to
12296 be accessed transparently. As a result code should not use ASN1_ITEM
12297 references directly (such as &X509_it) but instead use the relevant
12298 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12299 use of the new ASN1 code on platforms where exporting structures
12300 is problematical (for example in shared libraries) but exporting
12301 functions returning pointers to structures is not.
12305 * Add support for overriding the generation of SSL/TLS session IDs.
12306 These callbacks can be registered either in an SSL_CTX or per SSL.
12307 The purpose of this is to allow applications to control, if they wish,
12308 the arbitrary values chosen for use as session IDs, particularly as it
12309 can be useful for session caching in multiple-server environments. A
12310 command-line switch for testing this (and any client code that wishes
12311 to use such a feature) has been added to "s_server".
12313 *Geoff Thorpe, Lutz Jaenicke*
12315 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12316 of the form '#if defined(...) || defined(...) || ...' and
12317 '#if !defined(...) && !defined(...) && ...'. This also avoids
12318 the growing number of special cases it was previously handling.
12322 * Make all configuration macros available for application by making
12323 sure they are available in opensslconf.h, by giving them names starting
12324 with `OPENSSL_` to avoid conflicts with other packages and by making
12325 sure e_os2.h will cover all platform-specific cases together with
12327 Additionally, it is now possible to define configuration/platform-
12328 specific names (called "system identities"). In the C code, these
12329 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12330 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12331 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12336 * New option -set_serial to 'req' and 'x509' this allows the serial
12337 number to use to be specified on the command line. Previously self
12338 signed certificates were hard coded with serial number 0 and the
12339 CA options of 'x509' had to use a serial number in a file which was
12344 * New options to 'ca' utility to support V2 CRL entry extensions.
12345 Currently CRL reason, invalidity date and hold instruction are
12346 supported. Add new CRL extensions to V3 code and some new objects.
12350 * New function EVP_CIPHER_CTX_set_padding() this is used to
12351 disable standard block padding (aka PKCS#5 padding) in the EVP
12352 API, which was previously mandatory. This means that the data is
12353 not padded in any way and so the total length much be a multiple
12354 of the block size, otherwise an error occurs.
12358 * Initial (incomplete) OCSP SSL support.
12362 * New function OCSP_parse_url(). This splits up a URL into its host,
12363 port and path components: primarily to parse OCSP URLs. New -url
12364 option to ocsp utility.
12368 * New nonce behavior. The return value of OCSP_check_nonce() now
12369 reflects the various checks performed. Applications can decide
12370 whether to tolerate certain situations such as an absent nonce
12371 in a response when one was present in a request: the ocsp application
12372 just prints out a warning. New function OCSP_add1_basic_nonce()
12373 this is to allow responders to include a nonce in a response even if
12374 the request is nonce-less.
12378 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12379 skipped when using openssl x509 multiple times on a single input file,
12380 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12384 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12385 set string type: to handle setting ASN1_TIME structures. Fix ca
12386 utility to correctly initialize revocation date of CRLs.
12390 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12391 the clients preferred ciphersuites and rather use its own preferences.
12392 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12393 Internet Explorer by ensuring unchanged hash method during stepup.
12394 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12398 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12399 to aes and add a new 'exist' option to print out symbols that don't
12404 * Additional options to ocsp utility to allow flags to be set and
12405 additional certificates supplied.
12409 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12410 OCSP client a number of certificate to only verify the response
12415 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12416 handle the new API. Currently only ECB, CBC modes supported. Add new
12419 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12420 Encryption Standard (AES) Ciphersuites for Transport Layer
12421 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12422 not enabled by default and were not part of the "ALL" ciphersuite
12423 alias because they were not yet official; they could be
12424 explicitly requested by specifying the "AESdraft" ciphersuite
12425 group alias. In the final release of OpenSSL 0.9.7, the group
12426 alias is called "AES" and is part of "ALL".)
12428 *Ben Laurie, Steve Henson, Bodo Moeller*
12430 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12431 request to response.
12435 * Functions for OCSP responders. OCSP_request_onereq_count(),
12436 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12437 extract information from a certificate request. OCSP_response_create()
12438 creates a response and optionally adds a basic response structure.
12439 OCSP_basic_add1_status() adds a complete single response to a basic
12440 response and returns the OCSP_SINGLERESP structure just added (to allow
12441 extensions to be included for example). OCSP_basic_add1_cert() adds a
12442 certificate to a basic response and OCSP_basic_sign() signs a basic
12443 response with various flags. New helper functions ASN1_TIME_check()
12444 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12445 (converts ASN1_TIME to GeneralizedTime).
12449 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12450 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12451 structure from a certificate. X509_pubkey_digest() digests the public_key
12452 contents: this is used in various key identifiers.
12456 * Make sk_sort() tolerate a NULL argument.
12458 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12460 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12461 passed by the function are trusted implicitly. If any of them signed the
12462 response then it is assumed to be valid and is not verified.
12466 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12467 to data. This was previously part of the PKCS7 ASN1 code. This
12468 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12469 *Steve Henson, reported by Kenneth R. Robinette
12470 <support@securenetterm.com>*
12472 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12473 routines: without these tracing memory leaks is very painful.
12474 Fix leaks in PKCS12 and PKCS7 routines.
12478 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12479 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12480 effectively meant GeneralizedTime would never be used. Now it
12481 is initialised to -1 but X509_time_adj() now has to check the value
12482 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12483 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12484 *Steve Henson, reported by Kenneth R. Robinette
12485 <support@securenetterm.com>*
12487 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12488 result in a zero length in the ASN1_INTEGER structure which was
12489 not consistent with the structure when d2i_ASN1_INTEGER() was used
12490 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12491 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12492 where it did not print out a minus for negative ASN1_INTEGER.
12496 * Add summary printout to ocsp utility. The various functions which
12497 convert status values to strings have been renamed to:
12498 OCSP_response_status_str(), OCSP_cert_status_str() and
12499 OCSP_crl_reason_str() and are no longer static. New options
12500 to verify nonce values and to disable verification. OCSP response
12501 printout format cleaned up.
12505 * Add additional OCSP certificate checks. These are those specified
12506 in RFC2560. This consists of two separate checks: the CA of the
12507 certificate being checked must either be the OCSP signer certificate
12508 or the issuer of the OCSP signer certificate. In the latter case the
12509 OCSP signer certificate must contain the OCSP signing extended key
12510 usage. This check is performed by attempting to match the OCSP
12511 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12512 in the OCSP_CERTID structures of the response.
12516 * Initial OCSP certificate verification added to OCSP_basic_verify()
12517 and related routines. This uses the standard OpenSSL certificate
12518 verify routines to perform initial checks (just CA validity) and
12519 to obtain the certificate chain. Then additional checks will be
12520 performed on the chain. Currently the root CA is checked to see
12521 if it is explicitly trusted for OCSP signing. This is used to set
12522 a root CA as a global signing root: that is any certificate that
12523 chains to that CA is an acceptable OCSP signing certificate.
12527 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12528 extensions from a separate configuration file.
12529 As when reading extensions from the main configuration file,
12530 the '-extensions ...' option may be used for specifying the
12533 *Massimiliano Pala <madwolf@comune.modena.it>*
12535 * New OCSP utility. Allows OCSP requests to be generated or
12536 read. The request can be sent to a responder and the output
12537 parsed, outputted or printed in text form. Not complete yet:
12538 still needs to check the OCSP response validity.
12542 * New subcommands for 'openssl ca':
12543 `openssl ca -status <serial>` prints the status of the cert with
12544 the given serial number (according to the index file).
12545 `openssl ca -updatedb` updates the expiry status of certificates
12548 *Massimiliano Pala <madwolf@comune.modena.it>*
12550 * New '-newreq-nodes' command option to CA.pl. This is like
12551 '-newreq', but calls 'openssl req' with the '-nodes' option
12552 so that the resulting key is not encrypted.
12554 *Damien Miller <djm@mindrot.org>*
12556 * New configuration for the GNU Hurd.
12558 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12560 * Initial code to implement OCSP basic response verify. This
12561 is currently incomplete. Currently just finds the signer's
12562 certificate and verifies the signature on the response.
12566 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12567 value of OPENSSLDIR. This is available via the new '-d' option
12568 to 'openssl version', and is also included in 'openssl version -a'.
12572 * Allowing defining memory allocation callbacks that will be given
12573 file name and line number information in additional arguments
12574 (a `const char*` and an int). The basic functionality remains, as
12575 well as the original possibility to just replace malloc(),
12576 realloc() and free() by functions that do not know about these
12577 additional arguments. To register and find out the current
12578 settings for extended allocation functions, the following
12579 functions are provided:
12581 CRYPTO_set_mem_ex_functions
12582 CRYPTO_set_locked_mem_ex_functions
12583 CRYPTO_get_mem_ex_functions
12584 CRYPTO_get_locked_mem_ex_functions
12586 These work the same way as CRYPTO_set_mem_functions and friends.
12587 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12588 extended allocation function is enabled.
12589 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12590 a conventional allocation function is enabled.
12592 *Richard Levitte, Bodo Moeller*
12594 * Finish off removing the remaining LHASH function pointer casts.
12595 There should no longer be any prototype-casting required when using
12596 the LHASH abstraction, and any casts that remain are "bugs". See
12597 the callback types and macros at the head of lhash.h for details
12598 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12602 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12603 If /dev/[u]random devices are not available or do not return enough
12604 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12606 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12607 /etc/entropy will be queried once each in this sequence, querying stops
12608 when enough entropy was collected without querying more sockets.
12612 * Change the Unix RAND_poll() variant to be able to poll several
12613 random devices, as specified by DEVRANDOM, until a sufficient amount
12614 of data has been collected. We spend at most 10 ms on each file
12615 (select timeout) and read in non-blocking mode. DEVRANDOM now
12616 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12617 (previously it was just the string "/dev/urandom"), so on typical
12618 platforms the 10 ms delay will never occur.
12619 Also separate out the Unix variant to its own file, rand_unix.c.
12620 For VMS, there's a currently-empty rand_vms.c.
12624 * Move OCSP client related routines to ocsp_cl.c. These
12625 provide utility functions which an application needing
12626 to issue a request to an OCSP responder and analyse the
12627 response will typically need: as opposed to those which an
12628 OCSP responder itself would need which will be added later.
12630 OCSP_request_sign() signs an OCSP request with an API similar
12631 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12632 response. OCSP_response_get1_basic() extracts basic response
12633 from response. OCSP_resp_find_status(): finds and extracts status
12634 information from an OCSP_CERTID structure (which will be created
12635 when the request structure is built). These are built from lower
12636 level functions which work on OCSP_SINGLERESP structures but
12637 won't normally be used unless the application wishes to examine
12638 extensions in the OCSP response for example.
12640 Replace nonce routines with a pair of functions.
12641 OCSP_request_add1_nonce() adds a nonce value and optionally
12642 generates a random value. OCSP_check_nonce() checks the
12643 validity of the nonce in an OCSP response.
12647 * Change function OCSP_request_add() to OCSP_request_add0_id().
12648 This doesn't copy the supplied OCSP_CERTID and avoids the
12649 need to free up the newly created id. Change return type
12650 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12651 This can then be used to add extensions to the request.
12652 Deleted OCSP_request_new(), since most of its functionality
12653 is now in OCSP_REQUEST_new() (and the case insensitive name
12654 clash) apart from the ability to set the request name which
12655 will be added elsewhere.
12659 * Update OCSP API. Remove obsolete extensions argument from
12660 various functions. Extensions are now handled using the new
12661 OCSP extension code. New simple OCSP HTTP function which
12662 can be used to send requests and parse the response.
12666 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12667 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12668 uses the special reorder version of SET OF to sort the attributes
12669 and reorder them to match the encoded order. This resolves a long
12670 standing problem: a verify on a PKCS7 structure just after signing
12671 it used to fail because the attribute order did not match the
12672 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12673 it uses the received order. This is necessary to tolerate some broken
12674 software that does not order SET OF. This is handled by encoding
12675 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12676 to produce the required SET OF.
12680 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12681 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12682 files to get correct declarations of the ASN.1 item variables.
12686 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12687 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12688 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12689 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12690 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12691 ASN1_ITEM and no wrapper functions.
12695 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12696 replace the old function pointer based I/O routines. Change most of
12697 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12701 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12702 lines, recognize more "algorithms" that can be deselected, and make
12703 it complain about algorithm deselection that isn't recognised.
12707 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12708 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12709 to use new functions. Add NO_ASN1_OLD which can be set to remove
12710 some old style ASN1 functions: this can be used to determine if old
12711 code will still work when these eventually go away.
12715 * New extension functions for OCSP structures, these follow the
12716 same conventions as certificates and CRLs.
12720 * New function X509V3_add1_i2d(). This automatically encodes and
12721 adds an extension. Its behaviour can be customised with various
12722 flags to append, replace or delete. Various wrappers added for
12723 certificates and CRLs.
12727 * Fix to avoid calling the underlying ASN1 print routine when
12728 an extension cannot be parsed. Correct a typo in the
12729 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12733 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12734 entries for variables.
12738 * Add functionality to apps/openssl.c for detecting locking
12739 problems: As the program is single-threaded, all we have
12740 to do is register a locking callback using an array for
12741 storing which locks are currently held by the program.
12745 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12746 SSL_get_ex_data_X509_STORE_idx(), which is used in
12747 ssl_verify_cert_chain() and thus can be called at any time
12748 during TLS/SSL handshakes so that thread-safety is essential.
12749 Unfortunately, the ex_data design is not at all suited
12750 for multi-threaded use, so it probably should be abolished.
12754 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12756 *Broadcom, tweaked and integrated by Geoff Thorpe*
12758 * Move common extension printing code to new function
12759 X509V3_print_extensions(). Reorganise OCSP print routines and
12760 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12764 * New function X509_signature_print() to remove duplication in some
12769 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12770 set (this was treated exactly the same as SET OF previously). This
12771 is used to reorder the STACK representing the structure to match the
12772 encoding. This will be used to get round a problem where a PKCS7
12773 structure which was signed could not be verified because the STACK
12774 order did not reflect the encoded order.
12778 * Reimplement the OCSP ASN1 module using the new code.
12782 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12783 for its ASN1 operations. The old style function pointers still exist
12784 for now but they will eventually go away.
12788 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12789 completely replaces the old ASN1 functionality with a table driven
12790 encoder and decoder which interprets an ASN1_ITEM structure describing
12791 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12792 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12793 has also been converted to the new form.
12797 * Change BN_mod_exp_recp so that negative moduli are tolerated
12798 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12799 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12800 for negative moduli.
12804 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12805 of not touching the result's sign bit.
12809 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12814 * Changed the LHASH code to use prototypes for callbacks, and created
12815 macros to declare and implement thin (optionally static) functions
12816 that provide type-safety and avoid function pointer casting for the
12817 type-specific callbacks.
12821 * Added Kerberos Cipher Suites to be used with TLS, as written in
12823 *Veers Staats <staatsvr@asc.hpc.mil>,
12824 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12826 * Reformat the FAQ so the different questions and answers can be divided
12827 in sections depending on the subject.
12831 * Have the zlib compression code load ZLIB.DLL dynamically under
12836 * New function BN_mod_sqrt for computing square roots modulo a prime
12837 (using the probabilistic Tonelli-Shanks algorithm unless
12838 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12839 be handled deterministically).
12841 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12843 * Make BN_mod_inverse faster by explicitly handling small quotients
12844 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12845 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12849 * New function BN_kronecker.
12853 * Fix BN_gcd so that it works on negative inputs; the result is
12854 positive unless both parameters are zero.
12855 Previously something reasonably close to an infinite loop was
12856 possible because numbers could be growing instead of shrinking
12857 in the implementation of Euclid's algorithm.
12861 * Fix BN_is_word() and BN_is_one() macros to take into account the
12862 sign of the number in question.
12864 Fix BN_is_word(a,w) to work correctly for w == 0.
12866 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12867 because its test if the absolute value of 'a' equals 'w'.
12868 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12869 it exists mostly for use in the implementations of BN_is_zero(),
12870 BN_is_one(), and BN_is_word().
12874 * New function BN_swap.
12878 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12879 the exponentiation functions are more likely to produce reasonable
12880 results on negative inputs.
12884 * Change BN_mod_mul so that the result is always non-negative.
12885 Previously, it could be negative if one of the factors was negative;
12886 I don't think anyone really wanted that behaviour.
12890 * Move `BN_mod_...` functions into new file crypto/bn/bn_mod.c
12891 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12892 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12893 and add new functions:
12902 BN_mod_lshift1_quick
12904 BN_mod_lshift_quick
12906 These functions always generate non-negative results.
12908 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12909 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12911 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12912 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12913 be reduced modulo m.
12915 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12918 The following entry accidentally appeared in the CHANGES file
12919 distributed with OpenSSL 0.9.7. The modifications described in
12920 it do *not* apply to OpenSSL 0.9.7.
12922 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12923 was actually never needed) and in BN_mul(). The removal in BN_mul()
12924 required a small change in bn_mul_part_recursive() and the addition
12925 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12926 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12927 bn_sub_words() and bn_add_words() except they take arrays with
12933 * In 'openssl passwd', verify passwords read from the terminal
12934 unless the '-salt' option is used (which usually means that
12935 verification would just waste user's time since the resulting
12936 hash is going to be compared with some given password hash)
12937 or the new '-noverify' option is used.
12939 This is an incompatible change, but it does not affect
12940 non-interactive use of 'openssl passwd' (passwords on the command
12941 line, '-stdin' option, '-in ...' option) and thus should not
12942 cause any problems.
12946 * Remove all references to RSAref, since there's no more need for it.
12950 * Make DSO load along a path given through an environment variable
12951 (SHLIB_PATH) with shl_load().
12955 * Constify the ENGINE code as a result of BIGNUM constification.
12956 Also constify the RSA code and most things related to it. In a
12957 few places, most notable in the depth of the ASN.1 code, ugly
12958 casts back to non-const were required (to be solved at a later
12963 * Make it so the openssl application has all engines loaded by default.
12967 * Constify the BIGNUM routines a little more.
12971 * Add the following functions:
12973 ENGINE_load_cswift()
12975 ENGINE_load_atalla()
12976 ENGINE_load_nuron()
12977 ENGINE_load_builtin_engines()
12979 That way, an application can itself choose if external engines that
12980 are built-in in OpenSSL shall ever be used or not. The benefit is
12981 that applications won't have to be linked with libdl or other dso
12982 libraries unless it's really needed.
12984 Changed 'openssl engine' to load all engines on demand.
12985 Changed the engine header files to avoid the duplication of some
12986 declarations (they differed!).
12990 * 'openssl engine' can now list capabilities.
12994 * Better error reporting in 'openssl engine'.
12998 * Never call load_dh_param(NULL) in s_server.
13002 * Add engine application. It can currently list engines by name and
13003 identity, and test if they are actually available.
13007 * Improve RPM specification file by forcing symbolic linking and making
13008 sure the installed documentation is also owned by root.root.
13010 *Damien Miller <djm@mindrot.org>*
13012 * Give the OpenSSL applications more possibilities to make use of
13013 keys (public as well as private) handled by engines.
13017 * Add OCSP code that comes from CertCo.
13021 * Add VMS support for the Rijndael code.
13025 * Added untested support for Nuron crypto accelerator.
13029 * Add support for external cryptographic devices. This code was
13030 previously distributed separately as the "engine" branch.
13032 *Geoff Thorpe, Richard Levitte*
13034 * Rework the filename-translation in the DSO code. It is now possible to
13035 have far greater control over how a "name" is turned into a filename
13036 depending on the operating environment and any oddities about the
13037 different shared library filenames on each system.
13041 * Support threads on FreeBSD-elf in Configure.
13045 * Fix for SHA1 assembly problem with MASM: it produces
13046 warnings about corrupt line number information when assembling
13047 with debugging information. This is caused by the overlapping
13050 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13053 NCONF_get_number() has no error checking at all. As a replacement,
13054 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13055 promoted strongly. The old NCONF_get_number is kept around for
13056 binary backward compatibility.
13057 Make it possible for methods to load from something other than a BIO,
13058 by providing a function pointer that is given a name instead of a BIO.
13059 For example, this could be used to load configuration data from an
13064 * Fix for non blocking accept BIOs. Added new I/O special reason
13065 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13066 with non blocking I/O was not possible because no retry code was
13067 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13072 * Added the beginnings of Rijndael support.
13076 * Fix for bug in DirectoryString mask setting. Add support for
13077 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13078 to allow certificate printing to more controllable, additional
13079 'certopt' option to 'x509' to allow new printing options to be
13084 * Clean old EAY MD5 hack from e_os.h.
13088 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13090 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13091 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13093 *Joe Orton, Steve Henson*
13095 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13097 * Fix additional bug revealed by the NISCC test suite:
13099 Stop bug triggering large recursion when presented with
13100 certain ASN.1 tags [CVE-2003-0851][]
13104 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13106 * Fix various bugs revealed by running the NISCC test suite:
13108 Stop out of bounds reads in the ASN1 code when presented with
13109 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13111 If verify callback ignores invalid public key errors don't try to check
13112 certificate signature with the NULL public key.
13116 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13117 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13122 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13123 extra data after the compression methods not only for TLS 1.0
13124 but also for SSL 3.0 (as required by the specification).
13126 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13128 * Change X509_certificate_type() to mark the key as exported/exportable
13129 when it's 512 *bits* long, not 512 bytes.
13133 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13135 * Countermeasure against the Klima-Pokorny-Rosa extension of
13136 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13137 a protocol version number mismatch like a decryption error
13138 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13142 * Turn on RSA blinding by default in the default implementation
13143 to avoid a timing attack. Applications that don't want it can call
13144 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13145 They would be ill-advised to do so in most cases.
13147 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13149 * Change RSA blinding code so that it works when the PRNG is not
13150 seeded (in this case, the secret RSA exponent is abused as
13151 an unpredictable seed -- if it is not unpredictable, there
13152 is no point in blinding anyway). Make RSA blinding thread-safe
13153 by remembering the creator's thread ID in rsa->blinding and
13154 having all other threads use local one-time blinding factors
13155 (this requires more computation than sharing rsa->blinding, but
13156 avoids excessive locking; and if an RSA object is not shared
13157 between threads, blinding will still be very fast).
13161 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13163 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13164 via timing by performing a MAC computation even if incorrect
13165 block cipher padding has been found. This is a countermeasure
13166 against active attacks where the attacker has to distinguish
13167 between bad padding and a MAC verification error. [CVE-2003-0078][]
13169 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13170 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13171 Martin Vuagnoux (EPFL, Ilion)*
13173 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13175 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13176 memory from its contents. This is done with a counter that will
13177 place alternating values in each byte. This can be used to solve
13178 two issues: 1) the removal of calls to memset() by highly optimizing
13179 compilers, and 2) cleansing with other values than 0, since those can
13180 be read through on certain media, for example a swap space on disk.
13184 * Bugfix: client side session caching did not work with external caching,
13185 because the session->cipher setting was not restored when reloading
13186 from the external cache. This problem was masked, when
13187 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13188 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13192 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13193 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13195 *Zeev Lieber <zeev-l@yahoo.com>*
13197 * Undo an undocumented change introduced in 0.9.6e which caused
13198 repeated calls to OpenSSL_add_all_ciphers() and
13199 OpenSSL_add_all_digests() to be ignored, even after calling
13204 * Change the default configuration reader to deal with last line not
13205 being properly terminated.
13209 * Change X509_NAME_cmp() so it applies the special rules on handling
13210 DN values that are of type PrintableString, as well as RDNs of type
13211 emailAddress where the value has the type ia5String.
13213 *stefank@valicert.com via Richard Levitte*
13215 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13216 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13217 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13218 the bitwise-OR of the two for use by the majority of applications
13219 wanting this behaviour, and update the docs. The documented
13220 behaviour and actual behaviour were inconsistent and had been
13221 changing anyway, so this is more a bug-fix than a behavioural
13224 *Geoff Thorpe, diagnosed by Nadav Har'El*
13226 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13227 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13231 * Fix initialization code race conditions in
13232 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13233 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13234 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13235 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13236 ssl2_get_cipher_by_char(),
13237 ssl3_get_cipher_by_char().
13239 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13241 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13242 the cached sessions are flushed, as the remove_cb() might use ex_data
13243 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13244 (see [openssl.org #212]).
13246 *Geoff Thorpe, Lutz Jaenicke*
13248 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13249 length, instead of the encoding length to d2i_ASN1_OBJECT.
13253 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13255 * [In 0.9.6g-engine release:]
13256 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13258 *Lynn Gazis <lgazis@rainbow.com>*
13260 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13262 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13263 and get fix the header length calculation.
13264 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13265 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13267 * Use proper error handling instead of 'assertions' in buffer
13268 overflow checks added in 0.9.6e. This prevents DoS (the
13269 assertions could call abort()).
13271 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13273 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13275 * Add various sanity checks to asn1_get_length() to reject
13276 the ASN1 length bytes if they exceed sizeof(long), will appear
13277 negative or the content length exceeds the length of the
13280 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13282 * Fix cipher selection routines: ciphers without encryption had no flags
13283 for the cipher strength set and where therefore not handled correctly
13284 by the selection routines (PR #130).
13288 * Fix EVP_dsa_sha macro.
13293 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13294 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13295 that was added in OpenSSL 0.9.6d.
13297 As the countermeasure turned out to be incompatible with some
13298 broken SSL implementations, the new option is part of SSL_OP_ALL.
13299 SSL_OP_ALL is usually employed when compatibility with weird SSL
13300 implementations is desired (e.g. '-bugs' option to 's_client' and
13301 's_server'), so the new option is automatically set in many
13306 * Changes in security patch:
13308 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13309 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13310 Air Force Materiel Command, USAF, under agreement number
13313 * Add various sanity checks to asn1_get_length() to reject
13314 the ASN1 length bytes if they exceed sizeof(long), will appear
13315 negative or the content length exceeds the length of the
13316 supplied buffer. [CVE-2002-0659][]
13318 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13320 * Assertions for various potential buffer overflows, not known to
13321 happen in practice.
13323 *Ben Laurie (CHATS)*
13325 * Various temporary buffers to hold ASCII versions of integers were
13326 too small for 64 bit platforms. [CVE-2002-0655][]
13327 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13329 * Remote buffer overflow in SSL3 protocol - an attacker could
13330 supply an oversized session ID to a client. [CVE-2002-0656][]
13332 *Ben Laurie (CHATS)*
13334 * Remote buffer overflow in SSL2 protocol - an attacker could
13335 supply an oversized client master key. [CVE-2002-0656][]
13337 *Ben Laurie (CHATS)*
13339 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13341 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13342 encoded as NULL) with id-dsa-with-sha1.
13344 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13346 * Check various `X509_...()` return values in apps/req.c.
13348 *Nils Larsch <nla@trustcenter.de>*
13350 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13351 an end-of-file condition would erroneously be flagged, when the CRLF
13352 was just at the end of a processed block. The bug was discovered when
13353 processing data through a buffering memory BIO handing the data to a
13354 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13355 <ptsekov@syntrex.com> and Nedelcho Stanev.
13359 * Implement a countermeasure against a vulnerability recently found
13360 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13361 before application data chunks to avoid the use of known IVs
13362 with data potentially chosen by the attacker.
13366 * Fix length checks in ssl3_get_client_hello().
13370 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13371 to prevent ssl3_read_internal() from incorrectly assuming that
13372 ssl3_read_bytes() found application data while handshake
13373 processing was enabled when in fact s->s3->in_read_app_data was
13374 merely automatically cleared during the initial handshake.
13376 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13378 * Fix object definitions for Private and Enterprise: they were not
13379 recognized in their shortname (=lowercase) representation. Extend
13380 obj_dat.pl to issue an error when using undefined keywords instead
13381 of silently ignoring the problem (Svenning Sorensen
13382 <sss@sss.dnsalias.net>).
13386 * Fix DH_generate_parameters() so that it works for 'non-standard'
13387 generators, i.e. generators other than 2 and 5. (Previously, the
13388 code did not properly initialise the 'add' and 'rem' values to
13389 BN_generate_prime().)
13391 In the new general case, we do not insist that 'generator' is
13392 actually a primitive root: This requirement is rather pointless;
13393 a generator of the order-q subgroup is just as good, if not
13398 * Map new X509 verification errors to alerts. Discovered and submitted by
13399 Tom Wu <tom@arcot.com>.
13403 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13404 returning non-zero before the data has been completely received
13405 when using non-blocking I/O.
13407 *Bodo Moeller; problem pointed out by John Hughes*
13409 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13411 *Ben Laurie, Lutz Jaenicke*
13413 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13414 Yoram Zahavi <YoramZ@gilian.com>).
13418 * Add information about CygWin 1.3 and on, and preserve proper
13419 configuration for the versions before that.
13421 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13423 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13424 check whether we deal with a copy of a session and do not delete from
13425 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13426 <izhar@checkpoint.com>.
13430 * Do not store session data into the internal session cache, if it
13431 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13432 flag is set). Proposed by Aslam <aslam@funk.com>.
13436 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13441 * [In 0.9.6d-engine release:]
13442 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13444 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13446 * Add the configuration target linux-s390x.
13448 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13450 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13451 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13452 variable as an indication that a ClientHello message has been
13453 received. As the flag value will be lost between multiple
13454 invocations of ssl3_accept when using non-blocking I/O, the
13455 function may not be aware that a handshake has actually taken
13456 place, thus preventing a new session from being added to the
13459 To avoid this problem, we now set s->new_session to 2 instead of
13460 using a local variable.
13462 *Lutz Jaenicke, Bodo Moeller*
13464 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13465 if the SSL_R_LENGTH_MISMATCH error is detected.
13467 *Geoff Thorpe, Bodo Moeller*
13469 * New 'shared_ldflag' column in Configure platform table.
13473 * Fix EVP_CIPHER_mode macro.
13475 *"Dan S. Camper" <dan@bti.net>*
13477 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13478 type, we must throw them away by setting rr->length to 0.
13480 *D P Chang <dpc@qualys.com>*
13482 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13484 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13485 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13486 worked incorrectly for those cases where range = `10..._2` and
13487 `3*range` is two bits longer than range.)
13491 * Only add signing time to PKCS7 structures if it is not already
13496 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13497 OBJ_ld_ce should be OBJ_id_ce.
13498 Also some ip-pda OIDs in crypto/objects/objects.txt were
13499 incorrect (cf. RFC 3039).
13501 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13503 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13504 returns early because it has nothing to do.
13506 *Andy Schneider <andy.schneider@bjss.co.uk>*
13508 * [In 0.9.6c-engine release:]
13509 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13511 *Andy Schneider <andy.schneider@bjss.co.uk>*
13513 * [In 0.9.6c-engine release:]
13514 Add support for Cryptographic Appliance's keyserver technology.
13515 (Use engine 'keyclient')
13517 *Cryptographic Appliances and Geoff Thorpe*
13519 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13520 is called via tools/c89.sh because arguments have to be
13521 rearranged (all '-L' options must appear before the first object
13524 *Richard Shapiro <rshapiro@abinitio.com>*
13526 * [In 0.9.6c-engine release:]
13527 Add support for Broadcom crypto accelerator cards, backported
13530 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13532 * [In 0.9.6c-engine release:]
13533 Add support for SureWare crypto accelerator cards from
13534 Baltimore Technologies. (Use engine 'sureware')
13536 *Baltimore Technologies and Mark Cox*
13538 * [In 0.9.6c-engine release:]
13539 Add support for crypto accelerator cards from Accelerated
13540 Encryption Processing, www.aep.ie. (Use engine 'aep')
13542 *AEP Inc. and Mark Cox*
13544 * Add a configuration entry for gcc on UnixWare.
13546 *Gary Benson <gbenson@redhat.com>*
13548 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13549 messages are stored in a single piece (fixed-length part and
13550 variable-length part combined) and fix various bugs found on the way.
13554 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13555 instead. BIO_gethostbyname() does not know what timeouts are
13556 appropriate, so entries would stay in cache even when they have
13558 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13560 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13561 faced with a pathologically small ClientHello fragment that does
13562 not contain client_version: Instead of aborting with an error,
13563 simply choose the highest available protocol version (i.e.,
13564 TLS 1.0 unless it is disabled). In practice, ClientHello
13565 messages are never sent like this, but this change gives us
13566 strictly correct behaviour at least for TLS.
13570 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13571 never resets s->method to s->ctx->method when called from within
13572 one of the SSL handshake functions.
13574 *Bodo Moeller; problem pointed out by Niko Baric*
13576 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13577 (sent using the client's version number) if client_version is
13578 smaller than the protocol version in use. Also change
13579 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13580 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13581 the client will at least see that alert.
13585 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13590 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13591 client receives HelloRequest while in a handshake.
13593 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13595 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13596 should end in 'break', not 'goto end' which circumvents various
13597 cleanups done in state SSL_ST_OK. But session related stuff
13598 must be disabled for SSL_ST_OK in the case that we just sent a
13601 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13602 before just sending a HelloRequest.
13604 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13606 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13607 reveal whether illegal block cipher padding was found or a MAC
13608 verification error occurred. (Neither SSLerr() codes nor alerts
13609 are directly visible to potential attackers, but the information
13610 may leak via logfiles.)
13612 Similar changes are not required for the SSL 2.0 implementation
13613 because the number of padding bytes is sent in clear for SSL 2.0,
13614 and the extra bytes are just ignored. However ssl/s2_pkt.c
13615 failed to verify that the purported number of padding bytes is in
13620 * Add OpenUNIX-8 support including shared libraries
13621 (Boyd Lynn Gerber <gerberb@zenez.com>).
13625 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13626 'wristwatch attack' using huge encoding parameters (cf.
13627 James H. Manger's CRYPTO 2001 paper). Note that the
13628 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13629 encoding parameters and hence was not vulnerable.
13633 * BN_sqr() bug fix.
13635 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13637 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13638 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13639 followed by modular reduction.
13641 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13643 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13644 equivalent based on BN_pseudo_rand() instead of BN_rand().
13648 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13649 This function was broken, as the check for a new client hello message
13650 to handle SGC did not allow these large messages.
13651 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13655 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13659 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13660 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13664 * Rework the configuration and shared library support for Tru64 Unix.
13665 The configuration part makes use of modern compiler features and
13666 still retains old compiler behavior for those that run older versions
13667 of the OS. The shared library support part includes a variant that
13668 uses the RPATH feature, and is available through the special
13669 configuration target "alpha-cc-rpath", which will never be selected
13672 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13674 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13675 with the same message size as in ssl3_get_certificate_request().
13676 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13677 messages might inadvertently be reject as too long.
13679 *Petr Lampa <lampa@fee.vutbr.cz>*
13681 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13685 * Modified SSL library such that the verify_callback that has been set
13686 specifically for an SSL object with SSL_set_verify() is actually being
13687 used. Before the change, a verify_callback set with this function was
13688 ignored and the verify_callback() set in the SSL_CTX at the time of
13689 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13690 to allow the necessary settings.
13694 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13695 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13696 done automatically (in contradiction to the requirements of the C
13697 standard). This made problems when used from OpenSSH.
13701 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13702 dh->length and always used
13704 BN_rand_range(priv_key, dh->p).
13706 BN_rand_range() is not necessary for Diffie-Hellman, and this
13707 specific range makes Diffie-Hellman unnecessarily inefficient if
13708 dh->length (recommended exponent length) is much smaller than the
13709 length of dh->p. We could use BN_rand_range() if the order of
13710 the subgroup was stored in the DH structure, but we only have
13715 BN_rand(priv_key, l, ...)
13717 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13724 RSA_eay_public_encrypt
13725 RSA_eay_private_decrypt
13726 RSA_eay_private_encrypt (signing)
13727 RSA_eay_public_decrypt (signature verification)
13729 (default implementations for RSA_public_encrypt,
13730 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13731 always reject numbers >= n.
13735 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13736 to synchronize access to 'locking_thread'. This is necessary on
13737 systems where access to 'locking_thread' (an 'unsigned long'
13738 variable) is not atomic.
13742 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13743 *before* setting the 'crypto_lock_rand' flag. The previous code had
13744 a race condition if 0 is a valid thread ID.
13746 *Travis Vitek <vitek@roguewave.com>*
13748 * Add support for shared libraries under Irix.
13750 *Albert Chin-A-Young <china@thewrittenword.com>*
13752 * Add configuration option to build on Linux on both big-endian and
13753 little-endian MIPS.
13755 *Ralf Baechle <ralf@uni-koblenz.de>*
13757 * Add the possibility to create shared libraries on HP-UX.
13761 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13763 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13764 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13765 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13766 PRNG state recovery was possible based on the output of
13767 one PRNG request appropriately sized to gain knowledge on
13768 'md' followed by enough consecutive 1-byte PRNG requests
13769 to traverse all of 'state'.
13771 1. When updating 'md_local' (the current thread's copy of 'md')
13772 during PRNG output generation, hash all of the previous
13773 'md_local' value, not just the half used for PRNG output.
13775 2. Make the number of bytes from 'state' included into the hash
13776 independent from the number of PRNG bytes requested.
13778 The first measure alone would be sufficient to avoid
13779 Markku-Juhani's attack. (Actually it had never occurred
13780 to me that the half of 'md_local' used for chaining was the
13781 half from which PRNG output bytes were taken -- I had always
13782 assumed that the secret half would be used.) The second
13783 measure makes sure that additional data from 'state' is never
13784 mixed into 'md_local' in small portions; this heuristically
13785 further strengthens the PRNG.
13789 * Fix crypto/bn/asm/mips3.s.
13793 * When only the key is given to "enc", the IV is undefined. Print out
13794 an error message in this case.
13798 * Handle special case when X509_NAME is empty in X509 printing routines.
13802 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13803 positive and less than q.
13807 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13808 used: it isn't thread safe and the add_lock_callback should handle
13811 *Paul Rose <Paul.Rose@bridge.com>*
13813 * Verify that incoming data obeys the block size in
13814 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13820 *Ulf Möller, Bodo Möller*
13822 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13823 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13824 when fixing the server behaviour for backwards-compatible 'client
13825 hello' messages. (Note that the attack is impractical against
13826 SSL 3.0 and TLS 1.0 anyway because length and version checking
13827 means that the probability of guessing a valid ciphertext is
13828 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13831 Before 0.9.5, the countermeasure (hide the error by generating a
13832 random 'decryption result') did not work properly because
13833 ERR_clear_error() was missing, meaning that SSL_get_error() would
13834 detect the supposedly ignored error.
13836 Both problems are now fixed.
13840 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13841 (previously it was 1024).
13845 * Fix for compatibility mode trust settings: ignore trust settings
13846 unless some valid trust or reject settings are present.
13850 * Fix for blowfish EVP: its a variable length cipher.
13854 * Fix various bugs related to DSA S/MIME verification. Handle missing
13855 parameters in DSA public key structures and return an error in the
13856 DSA routines if parameters are absent.
13860 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13861 in the current directory if neither $RANDFILE nor $HOME was set.
13862 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13863 caused some confusion to Windows users who haven't defined $HOME.
13864 Thus RAND_file_name() is changed again: e_os.h can define a
13865 DEFAULT_HOME, which will be used if $HOME is not set.
13866 For Windows, we use "C:"; on other platforms, we still require
13867 environment variables.
13869 * Move 'if (!initialized) RAND_poll()' into regions protected by
13870 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13871 having multiple threads call RAND_poll() concurrently.
13875 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13876 combination of a flag and a thread ID variable.
13877 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13878 flag), *other* threads can enter ssleay_add_bytes without obeying
13879 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13880 that they do not hold after the first thread unsets add_do_not_lock).
13884 * Change bctest again: '-x' expressions are not available in all
13885 versions of 'test'.
13889 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13891 * Fix a couple of memory leaks in PKCS7_dataDecode()
13893 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13895 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13896 the default extension for executables, if any. Also, make the perl
13897 scripts that use symlink() to test if it really exists and use "cp"
13898 if it doesn't. All this made OpenSSL compilable and installable in
13903 * Fix for asn1_GetSequence() for indefinite length constructed data.
13904 If SEQUENCE is length is indefinite just set c->slen to the total
13905 amount of data available.
13907 *Steve Henson, reported by shige@FreeBSD.org*
13909 *This change does not apply to 0.9.7.*
13911 * Change bctest to avoid here-documents inside command substitution
13912 (workaround for FreeBSD /bin/sh bug).
13913 For compatibility with Ultrix, avoid shell functions (introduced
13914 in the bctest version that searches along $PATH).
13918 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13919 with des_encrypt() defined on some operating systems, like Solaris
13924 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13925 On the Importance of Eliminating Errors in Cryptographic
13926 Computations, J. Cryptology 14 (2001) 2, 101-119,
13927 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13931 * MIPS assembler BIGNUM division bug fix.
13935 * Disabled incorrect Alpha assembler code.
13939 * Fix PKCS#7 decode routines so they correctly update the length
13940 after reading an EOC for the EXPLICIT tag.
13944 *This change does not apply to 0.9.7.*
13946 * Fix bug in PKCS#12 key generation routines. This was triggered
13947 if a 3DES key was generated with a 0 initial byte. Include
13948 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13949 (but broken) behaviour.
13953 * Enhance bctest to search for a working bc along $PATH and print
13956 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13958 * Fix memory leaks in err.c: free err_data string if necessary;
13959 don't write to the wrong index in ERR_set_error_data.
13963 * Implement ssl23_peek (analogous to ssl23_read), which previously
13968 * Replace rdtsc with `_emit` statements for VC++ version 5.
13970 *Jeremy Cooper <jeremy@baymoo.org>*
13972 * Make it possible to reuse SSLv2 sessions.
13976 * In copy_email() check for >= 0 as a return value for
13977 X509_NAME_get_index_by_NID() since 0 is a valid index.
13979 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13981 * Avoid coredump with unsupported or invalid public keys by checking if
13982 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13983 PKCS7_verify() fails with non detached data.
13987 * Don't use getenv in library functions when run as setuid/setgid.
13988 New function OPENSSL_issetugid().
13992 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13993 due to incorrect handling of multi-threading:
13995 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13997 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13999 3. Count how many times MemCheck_off() has been called so that
14000 nested use can be treated correctly. This also avoids
14001 inband-signalling in the previous code (which relied on the
14002 assumption that thread ID 0 is impossible).
14006 * Add "-rand" option also to s_client and s_server.
14010 * Fix CPU detection on Irix 6.x.
14011 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14012 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14014 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14019 *This change does not apply to 0.9.7.*
14021 * Use the cached encoding of an X509_NAME structure rather than
14022 copying it. This is apparently the reason for the libsafe "errors"
14023 but the code is actually correct.
14027 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14028 Bleichenbacher's DSA attack.
14029 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14030 to be set and top=0 forces the highest bit to be set; top=-1 is new
14031 and leaves the highest bit random.
14033 *Ulf Moeller, Bodo Moeller*
14035 * In the `NCONF_...`-based implementations for `CONF_...` queries
14036 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14037 a temporary CONF structure with the data component set to NULL
14038 (which gives segmentation faults in lh_retrieve).
14039 Instead, use NULL for the CONF pointer in CONF_get_string and
14040 CONF_get_number (which may use environment variables) and directly
14041 return NULL from CONF_get_section.
14045 * Fix potential buffer overrun for EBCDIC.
14049 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14050 keyUsage if basicConstraints absent for a CA.
14054 * Make SMIME_write_PKCS7() write mail header values with a format that
14055 is more generally accepted (no spaces before the semicolon), since
14056 some programs can't parse those values properly otherwise. Also make
14057 sure BIO's that break lines after each write do not create invalid
14062 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14063 macros previously used would not encode an empty SEQUENCE OF
14064 and break the signature.
14068 *This change does not apply to 0.9.7.*
14070 * Zero the premaster secret after deriving the master secret in
14075 * Add some EVP_add_digest_alias registrations (as found in
14076 OpenSSL_add_all_digests()) to SSL_library_init()
14077 aka OpenSSL_add_ssl_algorithms(). This provides improved
14078 compatibility with peers using X.509 certificates
14079 with unconventional AlgorithmIdentifier OIDs.
14083 * Fix for Irix with NO_ASM.
14085 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14087 * ./config script fixes.
14089 *Ulf Moeller, Richard Levitte*
14091 * Fix 'openssl passwd -1'.
14095 * Change PKCS12_key_gen_asc() so it can cope with non null
14096 terminated strings whose length is passed in the passlen
14097 parameter, for example from PEM callbacks. This was done
14098 by adding an extra length parameter to asc2uni().
14100 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14102 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14103 call failed, free the DSA structure.
14107 * Fix to uni2asc() to cope with zero length Unicode strings.
14108 These are present in some PKCS#12 files.
14112 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14113 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14114 when writing a 32767 byte record.
14116 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14118 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14119 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14121 (RSA objects have a reference count access to which is protected
14122 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14123 so they are meant to be shared between threads.)
14124 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14125 "Reddie, Steven" <Steven.Reddie@ca.com>*
14127 * Fix a deadlock in CRYPTO_mem_leaks().
14131 * Use better test patterns in bntest.
14135 * rand_win.c fix for Borland C.
14139 * BN_rshift bugfix for n == 0.
14143 * Add a 'bctest' script that checks for some known 'bc' bugs
14144 so that 'make test' does not abort just because 'bc' is broken.
14148 * Store verify_result within SSL_SESSION also for client side to
14149 avoid potential security hole. (Re-used sessions on the client side
14150 always resulted in verify_result==X509_V_OK, not using the original
14151 result of the server certificate verification.)
14155 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14156 SSL3_RT_APPLICATION_DATA, return 0.
14157 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14162 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14163 releases, have been re-implemented by renaming the previous
14164 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14165 and ssl3_read_internal, respectively, and adding 'peek' parameters
14166 to them. The new ssl[23]_{read,peek} functions are calls to
14167 ssl[23]_read_internal with the 'peek' flag set appropriately.
14168 A 'peek' parameter has also been added to ssl3_read_bytes, which
14169 does the actual work for ssl3_read_internal.
14173 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14174 the method-specific "init()" handler. Also clean up ex_data after
14175 calling the method-specific "finish()" handler. Previously, this was
14176 happening the other way round.
14180 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14181 The previous value, 12, was not always sufficient for BN_mod_exp().
14185 * Make sure that shared libraries get the internal name engine with
14186 the full version number and not just 0. This should mark the
14187 shared libraries as not backward compatible. Of course, this should
14188 be changed again when we can guarantee backward binary compatibility.
14192 * Fix typo in get_cert_by_subject() in by_dir.c
14194 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14196 * Rework the system to generate shared libraries:
14198 - Make note of the expected extension for the shared libraries and
14199 if there is a need for symbolic links from for example libcrypto.so.0
14200 to libcrypto.so.0.9.7. There is extended info in Configure for
14203 - Make as few rebuilds of the shared libraries as possible.
14205 - Still avoid linking the OpenSSL programs with the shared libraries.
14207 - When installing, install the shared libraries separately from the
14212 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14214 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14215 and not in SSL_clear because the latter is also used by the
14216 accept/connect functions; previously, the settings made by
14217 SSL_set_read_ahead would be lost during the handshake.
14219 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14221 * Correct util/mkdef.pl to be selective about disabled algorithms.
14222 Previously, it would create entries for disabled algorithms no
14227 * Added several new manual pages for SSL_* function.
14231 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14233 * In ssl23_get_client_hello, generate an error message when faced
14234 with an initial SSL 3.0/TLS record that is too small to contain the
14235 first two bytes of the ClientHello message, i.e. client_version.
14236 (Note that this is a pathologic case that probably has never happened
14237 in real life.) The previous approach was to use the version number
14238 from the record header as a substitute; but our protocol choice
14239 should not depend on that one because it is not authenticated
14240 by the Finished messages.
14244 * More robust randomness gathering functions for Windows.
14246 *Jeffrey Altman <jaltman@columbia.edu>*
14248 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14249 not set then we don't setup the error code for issuer check errors
14250 to avoid possibly overwriting other errors which the callback does
14251 handle. If an application does set the flag then we assume it knows
14252 what it is doing and can handle the new informational codes
14257 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14258 a general "ANY" type, as such it should be able to decode anything
14259 including tagged types. However it didn't check the class so it would
14260 wrongly interpret tagged types in the same way as their universal
14261 counterpart and unknown types were just rejected. Changed so that the
14262 tagged and unknown types are handled in the same way as a SEQUENCE:
14263 that is the encoding is stored intact. There is also a new type
14264 "V_ASN1_OTHER" which is used when the class is not universal, in this
14265 case we have no idea what the actual type is so we just lump them all
14270 * On VMS, stdout may very well lead to a file that is written to
14271 in a record-oriented fashion. That means that every write() will
14272 write a separate record, which will be read separately by the
14273 programs trying to read from it. This can be very confusing.
14275 The solution is to put a BIO filter in the way that will buffer
14276 text until a linefeed is reached, and then write everything a
14277 line at a time, so every record written will be an actual line,
14278 not chunks of lines and not (usually doesn't happen, but I've
14279 seen it once) several lines in one record. BIO_f_linebuffer() is
14282 Currently, it's a VMS-only method, because that's where it has
14283 been tested well enough.
14287 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14288 it can return incorrect results.
14289 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14290 but it was in 0.9.6-beta[12].)
14294 * Disable the check for content being present when verifying detached
14295 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14296 include zero length content when signing messages.
14300 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14301 BIO_ctrl (for BIO pairs).
14305 * Add DSO method for VMS.
14309 * Bug fix: Montgomery multiplication could produce results with the
14314 * Add RPM specification openssl.spec and modify it to build three
14315 packages. The default package contains applications, application
14316 documentation and run-time libraries. The devel package contains
14317 include files, static libraries and function documentation. The
14318 doc package contains the contents of the doc directory. The original
14319 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14323 * Add a large number of documentation files for many SSL routines.
14325 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14327 * Add a configuration entry for Sony News 4.
14329 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14331 * Don't set the two most significant bits to one when generating a
14332 random number < q in the DSA library.
14336 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14337 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14338 the underlying transport is blocking) if a handshake took place.
14339 (The default behaviour is needed by applications such as s_client
14340 and s_server that use select() to determine when to use SSL_read;
14341 but for applications that know in advance when to expect data, it
14342 just makes things more complicated.)
14346 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14351 * Add a few more EBCDIC conditionals that make `req` and `x509`
14352 work better on such systems.
14354 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14356 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14357 Update PKCS12_parse() so it copies the friendlyName and the
14358 keyid to the certificates aux info.
14362 * Fix bug in PKCS7_verify() which caused an infinite loop
14363 if there was more than one signature.
14365 *Sven Uszpelkat <su@celocom.de>*
14367 * Major change in util/mkdef.pl to include extra information
14368 about each symbol, as well as presenting variables as well
14369 as functions. This change means that there's n more need
14370 to rebuild the .num files when some algorithms are excluded.
14374 * Allow the verify time to be set by an application,
14375 rather than always using the current time.
14379 * Phase 2 verify code reorganisation. The certificate
14380 verify code now looks up an issuer certificate by a
14381 number of criteria: subject name, authority key id
14382 and key usage. It also verifies self signed certificates
14383 by the same criteria. The main comparison function is
14384 X509_check_issued() which performs these checks.
14386 Lot of changes were necessary in order to support this
14387 without completely rewriting the lookup code.
14389 Authority and subject key identifier are now cached.
14391 The LHASH 'certs' is X509_STORE has now been replaced
14392 by a STACK_OF(X509_OBJECT). This is mainly because an
14393 LHASH can't store or retrieve multiple objects with
14394 the same hash value.
14396 As a result various functions (which were all internal
14397 use only) have changed to handle the new X509_STORE
14398 structure. This will break anything that messed round
14399 with X509_STORE internally.
14401 The functions X509_STORE_add_cert() now checks for an
14402 exact match, rather than just subject name.
14404 The X509_STORE API doesn't directly support the retrieval
14405 of multiple certificates matching a given criteria, however
14406 this can be worked round by performing a lookup first
14407 (which will fill the cache with candidate certificates)
14408 and then examining the cache for matches. This is probably
14409 the best we can do without throwing out X509_LOOKUP
14410 entirely (maybe later...).
14412 The X509_VERIFY_CTX structure has been enhanced considerably.
14414 All certificate lookup operations now go via a get_issuer()
14415 callback. Although this currently uses an X509_STORE it
14416 can be replaced by custom lookups. This is a simple way
14417 to bypass the X509_STORE hackery necessary to make this
14418 work and makes it possible to use more efficient techniques
14419 in future. A very simple version which uses a simple
14420 STACK for its trusted certificate store is also provided
14421 using X509_STORE_CTX_trusted_stack().
14423 The verify_cb() and verify() callbacks now have equivalents
14424 in the X509_STORE_CTX structure.
14426 X509_STORE_CTX also has a 'flags' field which can be used
14427 to customise the verify behaviour.
14431 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14432 excludes S/MIME capabilities.
14436 * When a certificate request is read in keep a copy of the
14437 original encoding of the signed data and use it when outputting
14438 again. Signatures then use the original encoding rather than
14439 a decoded, encoded version which may cause problems if the
14440 request is improperly encoded.
14444 * For consistency with other BIO_puts implementations, call
14445 buffer_write(b, ...) directly in buffer_puts instead of calling
14448 In BIO_puts, increment b->num_write as in BIO_write.
14450 *Peter.Sylvester@EdelWeb.fr*
14452 * Fix BN_mul_word for the case where the word is 0. (We have to use
14453 BN_zero, we may not return a BIGNUM with an array consisting of
14454 words set to zero.)
14458 * Avoid calling abort() from within the library when problems are
14459 detected, except if preprocessor symbols have been defined
14460 (such as REF_CHECK, BN_DEBUG etc.).
14464 * New openssl application 'rsautl'. This utility can be
14465 used for low level RSA operations. DER public key
14466 BIO/fp routines also added.
14470 * New Configure entry and patches for compiling on QNX 4.
14472 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14474 * A demo state-machine implementation was sponsored by
14475 Nuron (<http://www.nuron.com/>) and is now available in
14476 demos/state_machine.
14480 * New options added to the 'dgst' utility for signature
14481 generation and verification.
14485 * Unrecognized PKCS#7 content types are now handled via a
14486 catch all ASN1_TYPE structure. This allows unsupported
14487 types to be stored as a "blob" and an application can
14488 encode and decode it manually.
14492 * Fix various signed/unsigned issues to make a_strex.c
14493 compile under VC++.
14495 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14497 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14498 length if passed a buffer. ASN1_INTEGER_to_BN failed
14499 if passed a NULL BN and its argument was negative.
14501 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14503 * Modification to PKCS#7 encoding routines to output definite
14504 length encoding. Since currently the whole structures are in
14505 memory there's not real point in using indefinite length
14506 constructed encoding. However if OpenSSL is compiled with
14507 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14511 * Added BIO_vprintf() and BIO_vsnprintf().
14515 * Added more prefixes to parse for in the strings written
14516 through a logging bio, to cover all the levels that are available
14517 through syslog. The prefixes are now:
14519 PANIC, EMERG, EMR => LOG_EMERG
14520 ALERT, ALR => LOG_ALERT
14521 CRIT, CRI => LOG_CRIT
14522 ERROR, ERR => LOG_ERR
14523 WARNING, WARN, WAR => LOG_WARNING
14524 NOTICE, NOTE, NOT => LOG_NOTICE
14525 INFO, INF => LOG_INFO
14526 DEBUG, DBG => LOG_DEBUG
14528 and as before, if none of those prefixes are present at the
14529 beginning of the string, LOG_ERR is chosen.
14531 On Win32, the `LOG_*` levels are mapped according to this:
14533 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14534 LOG_WARNING => EVENTLOG_WARNING_TYPE
14535 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14539 * Made it possible to reconfigure with just the configuration
14540 argument "reconf" or "reconfigure". The command line arguments
14541 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14542 and are retrieved from there when reconfiguring.
14548 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14550 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14554 * The obj_dat.pl script was messing up the sorting of object
14555 names. The reason was that it compared the quoted version
14556 of strings as a result "OCSP" > "OCSP Signing" because
14557 " > SPACE. Changed script to store unquoted versions of
14558 names and add quotes on output. It was also omitting some
14559 names from the lookup table if they were given a default
14560 value (that is if SN is missing it is given the same
14561 value as LN and vice versa), these are now added on the
14562 grounds that if an object has a name we should be able to
14563 look it up. Finally added warning output when duplicate
14564 short or long names are found.
14568 * Changes needed for Tandem NSK.
14570 *Scott Uroff <scott@xypro.com>*
14572 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14573 RSA_padding_check_SSLv23(), special padding was never detected
14574 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14575 version rollback attacks was not effective.
14577 In s23_clnt.c, don't use special rollback-attack detection padding
14578 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14579 client; similarly, in s23_srvr.c, don't do the rollback check if
14580 SSL 2.0 is the only protocol enabled in the server.
14584 * Make it possible to get hexdumps of unprintable data with 'openssl
14585 asn1parse'. By implication, the functions ASN1_parse_dump() and
14586 BIO_dump_indent() are added.
14590 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14591 these print out strings and name structures based on various
14592 flags including RFC2253 support and proper handling of
14593 multibyte characters. Added options to the 'x509' utility
14594 to allow the various flags to be set.
14598 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14599 Also change the functions X509_cmp_current_time() and
14600 X509_gmtime_adj() work with an ASN1_TIME structure,
14601 this will enable certificates using GeneralizedTime in validity
14602 dates to be checked.
14606 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14607 negative public key encodings) on by default,
14608 NO_NEG_PUBKEY_BUG can be set to disable it.
14612 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14613 content octets. An i2c_ASN1_OBJECT is unnecessary because
14614 the encoding can be trivially obtained from the structure.
14618 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14619 not read locks (`CRYPTO_r_[un]lock`).
14623 * A first attempt at creating official support for shared
14624 libraries through configuration. I've kept it so the
14625 default is static libraries only, and the OpenSSL programs
14626 are always statically linked for now, but there are
14627 preparations for dynamic linking in place.
14628 This has been tested on Linux and Tru64.
14632 * Randomness polling function for Win9x, as described in:
14633 Peter Gutmann, Software Generation of Practically Strong
14638 * Fix so PRNG is seeded in req if using an already existing
14643 * New options to smime application. -inform and -outform
14644 allow alternative formats for the S/MIME message including
14645 PEM and DER. The -content option allows the content to be
14646 specified separately. This should allow things like Netscape
14647 form signing output easier to verify.
14651 * Fix the ASN1 encoding of tags using the 'long form'.
14655 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14656 STRING types. These convert content octets to and from the
14657 underlying type. The actual tag and length octets are
14658 already assumed to have been read in and checked. These
14659 are needed because all other string types have virtually
14660 identical handling apart from the tag. By having versions
14661 of the ASN1 functions that just operate on content octets
14662 IMPLICIT tagging can be handled properly. It also allows
14663 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14664 and ASN1_INTEGER are identical apart from the tag.
14668 * Change the handling of OID objects as follows:
14670 - New object identifiers are inserted in objects.txt, following
14671 the syntax given in objects.README.
14672 - objects.pl is used to process obj_mac.num and create a new
14674 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14677 This is currently kind of a hack, and the perl code in objects.pl
14678 isn't very elegant, but it works as I intended. The simplest way
14679 to check that it worked correctly is to look in obj_dat.h and
14680 check the array nid_objs and make sure the objects haven't moved
14681 around (this is important!). Additions are OK, as well as
14682 consistent name changes.
14686 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14690 * Addition of the command line parameter '-rand file' to 'openssl req'.
14691 The given file adds to whatever has already been seeded into the
14692 random pool through the RANDFILE configuration file option or
14693 environment variable, or the default random state file.
14697 * mkstack.pl now sorts each macro group into lexical order.
14698 Previously the output order depended on the order the files
14699 appeared in the directory, resulting in needless rewriting
14704 * Patches to make OpenSSL compile under Win32 again. Mostly
14705 work arounds for the VC++ problem that it treats func() as
14706 func(void). Also stripped out the parts of mkdef.pl that
14707 added extra typesafe functions: these no longer exist.
14711 * Reorganisation of the stack code. The macros are now all
14712 collected in safestack.h . Each macro is defined in terms of
14713 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14714 DEBUG_SAFESTACK is now handled in terms of function casts,
14715 this has the advantage of retaining type safety without the
14716 use of additional functions. If DEBUG_SAFESTACK is not defined
14717 then the non typesafe macros are used instead. Also modified the
14718 mkstack.pl script to handle the new form. Needs testing to see
14719 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14720 the default if no major problems. Similar behaviour for ASN1_SET_OF
14721 and PKCS12_STACK_OF.
14725 * When some versions of IIS use the 'NET' form of private key the
14726 key derivation algorithm is different. Normally MD5(password) is
14727 used as a 128 bit RC4 key. In the modified case
14728 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14729 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14730 as the old Netscape_RSA functions except they have an additional
14731 'sgckey' parameter which uses the modified algorithm. Also added
14732 an -sgckey command line option to the rsa utility. Thanks to
14733 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14734 algorithm to openssl-dev.
14738 * The evp_local.h macros were using 'c.##kname' which resulted in
14739 invalid expansion on some systems (SCO 5.0.5 for example).
14740 Corrected to 'c.kname'.
14742 *Phillip Porch <root@theporch.com>*
14744 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14745 a STACK of email addresses from a certificate or request, these look
14746 in the subject name and the subject alternative name extensions and
14747 omit any duplicate addresses.
14751 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14752 This makes DSA verification about 2 % faster.
14756 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14757 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14758 plus overhead for 1024 bit moduli).
14759 This makes exponentiations about 0.5 % faster for 1024 bit
14760 exponents (as measured by "openssl speed rsa2048").
14764 * Rename memory handling macros to avoid conflicts with other
14766 Malloc => OPENSSL_malloc
14767 Malloc_locked => OPENSSL_malloc_locked
14768 Realloc => OPENSSL_realloc
14769 Free => OPENSSL_free
14773 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14774 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14778 * CygWin32 support.
14780 *John Jarvie <jjarvie@newsguy.com>*
14782 * The type-safe stack code has been rejigged. It is now only compiled
14783 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14784 by default all type-specific stack functions are "#define"d back to
14785 standard stack functions. This results in more streamlined output
14786 but retains the type-safety checking possibilities of the original
14791 * The STACK code has been cleaned up, and certain type declarations
14792 that didn't make a lot of sense have been brought in line. This has
14793 also involved a cleanup of sorts in safestack.h to more correctly
14794 map type-safe stack functions onto their plain stack counterparts.
14795 This work has also resulted in a variety of "const"ifications of
14796 lots of the code, especially `_cmp` operations which should normally
14797 be prototyped with "const" parameters anyway.
14801 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14802 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14803 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14804 where all of 'md' is used each time the PRNG is used, but 'state'
14805 is used only indexed by a cyclic counter. As entropy may not be
14806 well distributed from the beginning, 'md' is important as a
14807 chaining variable. However, the output function chains only half
14808 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14809 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14810 in all of 'state' being rewritten, with the new values depending
14811 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14815 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14816 the handshake is continued after ssl_verify_cert_chain();
14817 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14818 can lead to 'unexplainable' connection aborts later.
14820 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14822 * Major EVP API cipher revision.
14823 Add hooks for extra EVP features. This allows various cipher
14824 parameters to be set in the EVP interface. Support added for variable
14825 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14826 setting of RC2 and RC5 parameters.
14828 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14831 Remove lots of duplicated code from the EVP library. For example *every*
14832 cipher init() function handles the 'iv' in the same way according to the
14833 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14834 for CFB and OFB modes they zero ctx->num.
14836 New functionality allows removal of S/MIME code RC2 hack.
14838 Most of the routines have the same form and so can be declared in terms
14841 By shifting this to the top level EVP_CipherInit() it can be removed from
14842 all individual ciphers. If the cipher wants to handle IVs or keys
14843 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14846 Change lots of functions like EVP_EncryptUpdate() to now return a
14847 value: although software versions of the algorithms cannot fail
14848 any installed hardware versions can.
14852 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14853 this option is set, tolerate broken clients that send the negotiated
14854 protocol version number instead of the requested protocol version
14859 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14860 i.e. non-zero for export ciphersuites, zero otherwise.
14861 Previous versions had this flag inverted, inconsistent with
14862 rsa_tmp_cb (..._TMP_RSA_CB).
14864 *Bodo Moeller; problem reported by Amit Chopra*
14866 * Add missing DSA library text string. Work around for some IIS
14867 key files with invalid SEQUENCE encoding.
14871 * Add a document (doc/standards.txt) that list all kinds of standards
14872 and so on that are implemented in OpenSSL.
14876 * Enhance c_rehash script. Old version would mishandle certificates
14877 with the same subject name hash and wouldn't handle CRLs at all.
14878 Added -fingerprint option to crl utility, to support new c_rehash
14883 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14887 * Fix for SSL server purpose checking. Server checking was
14888 rejecting certificates which had extended key usage present
14889 but no ssl client purpose.
14891 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14893 * Make PKCS#12 code work with no password. The PKCS#12 spec
14894 is a little unclear about how a blank password is handled.
14895 Since the password in encoded as a BMPString with terminating
14896 double NULL a zero length password would end up as just the
14897 double NULL. However no password at all is different and is
14898 handled differently in the PKCS#12 key generation code. NS
14899 treats a blank password as zero length. MSIE treats it as no
14900 password on export: but it will try both on import. We now do
14901 the same: PKCS12_parse() tries zero length and no password if
14902 the password is set to "" or NULL (NULL is now a valid password:
14903 it wasn't before) as does the pkcs12 application.
14907 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14908 perror when PEM_read_bio_X509_REQ fails, the error message must
14909 be obtained from the error queue.
14913 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14914 it in ERR_remove_state if appropriate, and change ERR_get_state
14915 accordingly to avoid race conditions (this is necessary because
14916 thread_hash is no longer constant once set).
14920 * Bugfix for linux-elf makefile.one.
14924 * RSA_get_default_method() will now cause a default
14925 RSA_METHOD to be chosen if one doesn't exist already.
14926 Previously this was only set during a call to RSA_new()
14927 or RSA_new_method(NULL) meaning it was possible for
14928 RSA_get_default_method() to return NULL.
14932 * Added native name translation to the existing DSO code
14933 that will convert (if the flag to do so is set) filenames
14934 that are sufficiently small and have no path information
14935 into a canonical native form. Eg. "blah" converted to
14936 "libblah.so" or "blah.dll" etc.
14940 * New function ERR_error_string_n(e, buf, len) which is like
14941 ERR_error_string(e, buf), but writes at most 'len' bytes
14942 including the 0 terminator. For ERR_error_string_n, 'buf'
14945 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14947 * CONF library reworked to become more general. A new CONF
14948 configuration file reader "class" is implemented as well as a
14949 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14950 old `CONF_*` functions are still there, but are reimplemented to
14951 work in terms of the new functions. Also, a set of functions
14952 to handle the internal storage of the configuration data is
14953 provided to make it easier to write new configuration file
14954 reader "classes" (I can definitely see something reading a
14955 configuration file in XML format, for example), called `_CONF_*`,
14956 or "the configuration storage API"...
14958 The new configuration file reading functions are:
14960 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14961 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14963 NCONF_default, NCONF_WIN32
14965 NCONF_dump_fp, NCONF_dump_bio
14967 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14968 NCONF_new creates a new CONF object. This works in the same way
14969 as other interfaces in OpenSSL, like the BIO interface.
14970 `NCONF_dump_*` dump the internal storage of the configuration file,
14971 which is useful for debugging. All other functions take the same
14972 arguments as the old `CONF_*` functions with the exception of the
14973 first that must be a `CONF *` instead of a `LHASH *`.
14975 To make it easier to use the new classes with the old `CONF_*` functions,
14976 the function CONF_set_default_method is provided.
14980 * Add '-tls1' option to 'openssl ciphers', which was already
14981 mentioned in the documentation but had not been implemented.
14982 (This option is not yet really useful because even the additional
14983 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14987 * Initial DSO code added into libcrypto for letting OpenSSL (and
14988 OpenSSL-based applications) load shared libraries and bind to
14989 them in a portable way.
14991 *Geoff Thorpe, with contributions from Richard Levitte*
14993 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
14995 * Make sure _lrotl and _lrotr are only used with MSVC.
14997 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14998 (the default implementation of RAND_status).
15000 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15001 to '-clrext' (= clear extensions), as intended and documented.
15002 *Bodo Moeller; inconsistency pointed out by Michael Attili
15003 <attili@amaxo.com>*
15005 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15006 was larger than the MD block size.
15008 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15010 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15011 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15012 using the passed key: if the passed key was a private key the result
15013 of X509_print(), for example, would be to print out all the private key
15018 * des_quad_cksum() byte order bug fix.
15019 *Ulf Möller, using the problem description in krb4-0.9.7, where
15020 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15022 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15025 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15027 * For easily testing in shell scripts whether some command
15028 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15029 returns with exit code 0 iff no command of the given name is available.
15030 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15031 the output goes to stdout and nothing is printed to stderr.
15032 Additional arguments are always ignored.
15034 Since for each cipher there is a command of the same name,
15035 the 'no-cipher' compilation switches can be tested this way.
15037 ('openssl no-XXX' is not able to detect pseudo-commands such
15038 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15042 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15046 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15047 is set; it will be thrown away anyway because each handshake creates
15049 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15050 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15051 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15052 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15056 * New s_client option -ign_eof: EOF at stdin is ignored, and
15057 'Q' and 'R' lose their special meanings (quit/renegotiate).
15058 This is part of what -quiet does; unlike -quiet, -ign_eof
15059 does not suppress any output.
15063 * Add compatibility options to the purpose and trust code. The
15064 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15065 accepts a certificate or CA, this was the previous behaviour,
15066 with all the associated security issues.
15068 X509_TRUST_COMPAT is the old trust behaviour: only and
15069 automatically trust self signed roots in certificate store. A
15070 new trust setting X509_TRUST_DEFAULT is used to specify that
15071 a purpose has no associated trust setting and it should instead
15072 use the value in the default purpose.
15076 * Fix the PKCS#8 DSA private key code so it decodes keys again
15077 and fix a memory leak.
15081 * In util/mkerr.pl (which implements 'make errors'), preserve
15082 reason strings from the previous version of the .c file, as
15083 the default to have only downcase letters (and digits) in
15084 automatically generated reasons codes is not always appropriate.
15088 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15089 using strerror. Previously, ERR_reason_error_string() returned
15090 library names as reason strings for SYSerr; but SYSerr is a special
15091 case where small numbers are errno values, not library numbers.
15095 * Add '-dsaparam' option to 'openssl dhparam' application. This
15096 converts DSA parameters into DH parameters. (When creating parameters,
15097 DSA_generate_parameters is used.)
15101 * Include 'length' (recommended exponent length) in C code generated
15102 by 'openssl dhparam -C'.
15106 * The second argument to set_label in perlasm was already being used
15107 so couldn't be used as a "file scope" flag. Moved to third argument
15112 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15113 instead of RAND_bytes for encryption IVs and salts.
15117 * Include RAND_status() into RAND_METHOD instead of implementing
15118 it only for md_rand.c Otherwise replacing the PRNG by calling
15119 RAND_set_rand_method would be impossible.
15123 * Don't let DSA_generate_key() enter an infinite loop if the random
15124 number generation fails.
15128 * New 'rand' application for creating pseudo-random output.
15132 * Added configuration support for Linux/IA64
15134 *Rolf Haberrecker <rolf@suse.de>*
15136 * Assembler module support for Mingw32.
15140 * Shared library support for HPUX (in shlib/).
15142 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15144 * Shared library support for Solaris gcc.
15146 *Lutz Behnke <behnke@trustcenter.de>*
15148 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15150 * PKCS7_encrypt() was adding text MIME headers twice because they
15151 were added manually and by SMIME_crlf_copy().
15155 * In bntest.c don't call BN_rand with zero bits argument.
15157 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15159 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15160 case was implemented. This caused BN_div_recp() to fail occasionally.
15164 * Add an optional second argument to the set_label() in the perl
15165 assembly language builder. If this argument exists and is set
15166 to 1 it signals that the assembler should use a symbol whose
15167 scope is the entire file, not just the current function. This
15168 is needed with MASM which uses the format label:: for this scope.
15170 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15172 * Change the ASN1 types so they are typedefs by default. Before
15173 almost all types were #define'd to ASN1_STRING which was causing
15174 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15179 * Change names of new functions to the new get1/get0 naming
15180 convention: After 'get1', the caller owns a reference count
15181 and has to call `..._free`; 'get0' returns a pointer to some
15182 data structure without incrementing reference counters.
15183 (Some of the existing 'get' functions increment a reference
15184 counter, some don't.)
15185 Similarly, 'set1' and 'add1' functions increase reference
15186 counters or duplicate objects.
15190 * Allow for the possibility of temp RSA key generation failure:
15191 the code used to assume it always worked and crashed on failure.
15195 * Fix potential buffer overrun problem in BIO_printf().
15196 *Ulf Möller, using public domain code by Patrick Powell; problem
15197 pointed out by David Sacerdote <das33@cornell.edu>*
15199 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15200 RAND_egd() and RAND_status(). In the command line application,
15201 the EGD socket can be specified like a seed file using RANDFILE
15206 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15207 Some CAs (e.g. Verisign) distribute certificates in this form.
15211 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15212 list to exclude them. This means that no special compilation option
15213 is needed to use anonymous DH: it just needs to be included in the
15218 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15219 EVP_MD_type. The old functionality is available in a new macro called
15220 EVP_MD_md(). Change code that uses it and update docs.
15224 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15225 where the `void *` argument is replaced by a function pointer argument.
15226 Previously `void *` was abused to point to functions, which works on
15227 many platforms, but is not correct. As these functions are usually
15228 called by macros defined in OpenSSL header files, most source code
15229 should work without changes.
15233 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15234 sections with information on -D... compiler switches used for
15235 compiling the library so that applications can see them. To enable
15236 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15237 must be defined. E.g.,
15238 #define OPENSSL_ALGORITHM_DEFINES
15239 #include <openssl/opensslconf.h>
15240 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15242 *Richard Levitte, Ulf and Bodo Möller*
15244 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15249 * Change the 'other' type in certificate aux info to a STACK_OF
15250 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15251 the required ASN1 format: arbitrary types determined by an OID.
15255 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15256 argument to 'req'. This is not because the function is newer or
15257 better than others it just uses the work 'NEW' in the certificate
15258 request header lines. Some software needs this.
15262 * Reorganise password command line arguments: now passwords can be
15263 obtained from various sources. Delete the PEM_cb function and make
15264 it the default behaviour: i.e. if the callback is NULL and the
15265 usrdata argument is not NULL interpret it as a null terminated pass
15266 phrase. If usrdata and the callback are NULL then the pass phrase
15267 is prompted for as usual.
15271 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15272 the support is automatically enabled. The resulting binaries will
15273 autodetect the card and use it if present.
15275 *Ben Laurie and Compaq Inc.*
15277 * Work around for Netscape hang bug. This sends certificate request
15278 and server done in one record. Since this is perfectly legal in the
15279 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15280 the bugs/SSLv3 entry for more info.
15284 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15288 * Add -rand argument to smime and pkcs12 applications and read/write
15293 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15297 * Add command line password options to the remaining applications.
15301 * Bug fix for BN_div_recp() for numerators with an even number of
15306 * More tests in bntest.c, and changed test_bn output.
15310 * ./config recognizes MacOS X now.
15314 * Bug fix for BN_div() when the first words of num and divisor are
15315 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15319 * Add support for various broken PKCS#8 formats, and command line
15320 options to produce them.
15324 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15325 get temporary BIGNUMs from a BN_CTX.
15329 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15334 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15335 include a #define from the old name to the new. The original intent
15336 was that statically linked binaries could for example just call
15337 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15338 link with digests. This never worked because SSLeay_add_all_digests()
15339 and SSLeay_add_all_ciphers() were in the same source file so calling
15340 one would link with the other. They are now in separate source files.
15344 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15348 * Use a less unusual form of the Miller-Rabin primality test (it used
15349 a binary algorithm for exponentiation integrated into the Miller-Rabin
15350 loop, our standard modexp algorithms are faster).
15354 * Support for the EBCDIC character set completed.
15356 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15358 * Source code cleanups: use const where appropriate, eliminate casts,
15359 use `void *` instead of `char *` in lhash.
15363 * Bugfix: ssl3_send_server_key_exchange was not restartable
15364 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15365 this the server could overwrite ephemeral keys that the client
15370 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15371 using 50 iterations of the Rabin-Miller test.
15373 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15374 iterations of the Rabin-Miller test as required by the appendix
15375 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15376 As BN_is_prime_fasttest includes trial division, DSA parameter
15377 generation becomes much faster.
15379 This implies a change for the callback functions in DSA_is_prime
15380 and DSA_generate_parameters: The callback function is called once
15381 for each positive witness in the Rabin-Miller test, not just
15382 occasionally in the inner loop; and the parameters to the
15383 callback function now provide an iteration count for the outer
15384 loop rather than for the current invocation of the inner loop.
15385 DSA_generate_parameters additionally can call the callback
15386 function with an 'iteration count' of -1, meaning that a
15387 candidate has passed the trial division test (when q is generated
15388 from an application-provided seed, trial division is skipped).
15392 * New function BN_is_prime_fasttest that optionally does trial
15393 division before starting the Rabin-Miller test and has
15394 an additional BN_CTX * argument (whereas BN_is_prime always
15395 has to allocate at least one BN_CTX).
15396 'callback(1, -1, cb_arg)' is called when a number has passed the
15397 trial division stage.
15401 * Fix for bug in CRL encoding. The validity dates weren't being handled
15406 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15410 * New function BN_pseudo_rand().
15414 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15415 bignum version of BN_from_montgomery() with the working code from
15416 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15421 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15422 made it impossible to use the same SSL_SESSION data structure in
15423 SSL2 clients in multiple threads.
15427 * The return value of RAND_load_file() no longer counts bytes obtained
15428 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15429 to seed the PRNG (previously an explicit byte count was required).
15431 *Ulf Möller, Bodo Möller*
15433 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15434 used `char *` instead of `void *` and had casts all over the place.
15438 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15442 * Retain source code compatibility for BN_prime_checks macro:
15443 BN_is_prime(..., BN_prime_checks, ...) now uses
15444 BN_prime_checks_for_size to determine the appropriate number of
15445 Rabin-Miller iterations.
15449 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15450 DH_CHECK_P_NOT_SAFE_PRIME.
15451 (Check if this is true? OpenPGP calls them "strong".)
15455 * Merge the functionality of "dh" and "gendh" programs into a new program
15456 "dhparam". The old programs are retained for now but will handle DH keys
15457 (instead of parameters) in future.
15461 * Make the ciphers, s_server and s_client programs check the return values
15462 when a new cipher list is set.
15466 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15467 ciphers. Before when the 56bit ciphers were enabled the sorting was
15470 The syntax for the cipher sorting has been extended to support sorting by
15471 cipher-strength (using the strength_bits hard coded in the tables).
15472 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15474 Fix a bug in the cipher-command parser: when supplying a cipher command
15475 string with an "undefined" symbol (neither command nor alphanumeric
15476 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15477 an error is flagged.
15479 Due to the strength-sorting extension, the code of the
15480 ssl_create_cipher_list() function was completely rearranged. I hope that
15481 the readability was also increased :-)
15483 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15485 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15486 for the first serial number and places 2 in the serial number file. This
15487 avoids problems when the root CA is created with serial number zero and
15488 the first user certificate has the same issuer name and serial number
15493 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15494 the new code. Add documentation for this stuff.
15498 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15499 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15500 structures and behave in an analogous way to the X509v3 functions:
15501 they shouldn't be called directly but wrapper functions should be used
15504 So we also now have some wrapper functions that call the X509at functions
15505 when passed certificate requests. (TO DO: similar things can be done with
15506 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15507 things. Some of these need some d2i or i2d and print functionality
15508 because they handle more complex structures.)
15512 * Add missing #ifndefs that caused missing symbols when building libssl
15513 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15514 NO_RSA in `ssl/s2*.c`.
15516 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15518 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15519 has a return value which indicates the quality of the random data
15520 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15521 error queue. New function RAND_pseudo_bytes() generates output that is
15522 guaranteed to be unique but not unpredictable. RAND_add is like
15523 RAND_seed, but takes an extra argument for an entropy estimate
15524 (RAND_seed always assumes full entropy).
15528 * Do more iterations of Rabin-Miller probable prime test (specifically,
15529 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15530 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15531 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15532 false-positive rate of at most 2^-80 for random input.
15536 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15540 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15541 in the 0.9.5 release), this returns the chain
15542 from an X509_CTX structure with a dup of the stack and all
15543 the X509 reference counts upped: so the stack will exist
15544 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15547 Also make SSL_SESSION_print() print out the verify return
15552 * Add manpage for the pkcs12 command. Also change the default
15553 behaviour so MAC iteration counts are used unless the new
15554 -nomaciter option is used. This improves file security and
15555 only older versions of MSIE (4.0 for example) need it.
15559 * Honor the no-xxx Configure options when creating .DEF files.
15563 * Add PKCS#10 attributes to field table: challengePassword,
15564 unstructuredName and unstructuredAddress. These are taken from
15565 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15566 international characters are used.
15568 More changes to X509_ATTRIBUTE code: allow the setting of types
15569 based on strings. Remove the 'loc' parameter when adding
15570 attributes because these will be a SET OF encoding which is sorted
15575 * Initial changes to the 'req' utility to allow request generation
15576 automation. This will allow an application to just generate a template
15577 file containing all the field values and have req construct the
15580 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15581 used all over the place including certificate requests and PKCS#7
15582 structures. They are currently handled manually where necessary with
15583 some primitive wrappers for PKCS#7. The new functions behave in a
15584 manner analogous to the X509 extension functions: they allow
15585 attributes to be looked up by NID and added.
15587 Later something similar to the X509V3 code would be desirable to
15588 automatically handle the encoding, decoding and printing of the
15589 more complex types. The string types like challengePassword can
15590 be handled by the string table functions.
15592 Also modified the multi byte string table handling. Now there is
15593 a 'global mask' which masks out certain types. The table itself
15594 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15595 is useful when for example there is only one permissible type
15596 (as in countryName) and using the mask might result in no valid
15601 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15602 SSL_get_peer_finished to allow applications to obtain the latest
15603 Finished messages sent to the peer or expected from the peer,
15604 respectively. (SSL_get_peer_finished is usually the Finished message
15605 actually received from the peer, otherwise the protocol will be aborted.)
15607 As the Finished message are message digests of the complete handshake
15608 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15609 be used for external authentication procedures when the authentication
15610 provided by SSL/TLS is not desired or is not enough.
15614 * Enhanced support for Alpha Linux is added. Now ./config checks if
15615 the host supports BWX extension and if Compaq C is present on the
15616 $PATH. Just exploiting of the BWX extension results in 20-30%
15617 performance kick for some algorithms, e.g. DES and RC4 to mention
15618 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15623 * Add support for MS "fast SGC". This is arguably a violation of the
15624 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15625 weak crypto and after checking the certificate is SGC a second one
15626 with strong crypto. MS SGC stops the first handshake after receiving
15627 the server certificate message and sends a second client hello. Since
15628 a server will typically do all the time consuming operations before
15629 expecting any further messages from the client (server key exchange
15630 is the most expensive) there is little difference between the two.
15632 To get OpenSSL to support MS SGC we have to permit a second client
15633 hello message after we have sent server done. In addition we have to
15634 reset the MAC if we do get this second client hello.
15638 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15639 if a DER encoded private key is RSA or DSA traditional format. Changed
15640 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15641 format DER encoded private key. Newer code should use PKCS#8 format which
15642 has the key type encoded in the ASN1 structure. Added DER private key
15643 support to pkcs8 application.
15647 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15648 ciphersuites has been selected (as required by the SSL 3/TLS 1
15649 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15650 is set, we interpret this as a request to violate the specification
15651 (the worst that can happen is a handshake failure, and 'correct'
15652 behaviour would result in a handshake failure anyway).
15656 * In SSL_CTX_add_session, take into account that there might be multiple
15657 SSL_SESSION structures with the same session ID (e.g. when two threads
15658 concurrently obtain them from an external cache).
15659 The internal cache can handle only one SSL_SESSION with a given ID,
15660 so if there's a conflict, we now throw out the old one to achieve
15665 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15666 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15667 some routines that use cipher OIDs: some ciphers do not have OIDs
15668 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15673 * Simplify the trust setting structure and code. Now we just have
15674 two sequences of OIDs for trusted and rejected settings. These will
15675 typically have values the same as the extended key usage extension
15676 and any application specific purposes.
15678 The trust checking code now has a default behaviour: it will just
15679 check for an object with the same NID as the passed id. Functions can
15680 be provided to override either the default behaviour or the behaviour
15681 for a given id. SSL client, server and email already have functions
15682 in place for compatibility: they check the NID and also return "trusted"
15683 if the certificate is self signed.
15687 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15688 traditional format into an EVP_PKEY structure.
15692 * Add a password callback function PEM_cb() which either prompts for
15693 a password if usr_data is NULL or otherwise assumes it is a null
15694 terminated password. Allow passwords to be passed on command line
15695 environment or config files in a few more utilities.
15699 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15700 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15701 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15702 Update documentation.
15706 * Support for ASN1 "NULL" type. This could be handled before by using
15707 ASN1_TYPE but there wasn't any function that would try to read a NULL
15708 and produce an error if it couldn't. For compatibility we also have
15709 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15710 don't allocate anything because they don't need to.
15714 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15717 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15719 * Rebuild of the memory allocation routines used by OpenSSL code and
15720 possibly others as well. The purpose is to make an interface that
15721 provide hooks so anyone can build a separate set of allocation and
15722 deallocation routines to be used by OpenSSL, for example memory
15723 pool implementations, or something else, which was previously hard
15724 since Malloc(), Realloc() and Free() were defined as macros having
15725 the values malloc, realloc and free, respectively (except for Win32
15726 compilations). The same is provided for memory debugging code.
15727 OpenSSL already comes with functionality to find memory leaks, but
15728 this gives people a chance to debug other memory problems.
15730 With these changes, a new set of functions and macros have appeared:
15732 CRYPTO_set_mem_debug_functions() [F]
15733 CRYPTO_get_mem_debug_functions() [F]
15734 CRYPTO_dbg_set_options() [F]
15735 CRYPTO_dbg_get_options() [F]
15736 CRYPTO_malloc_debug_init() [M]
15738 The memory debug functions are NULL by default, unless the library
15739 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15740 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15741 gives the standard debugging functions that come with OpenSSL) or
15742 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15743 provided by the library user) must be used. When the standard
15744 debugging functions are used, CRYPTO_dbg_set_options can be used to
15745 request additional information:
15746 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15747 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15749 Also, things like CRYPTO_set_mem_functions will always give the
15750 expected result (the new set of functions is used for allocation
15751 and deallocation) at all times, regardless of platform and compiler
15754 To finish it up, some functions that were never use in any other
15755 way than through macros have a new API and new semantic:
15757 CRYPTO_dbg_malloc()
15758 CRYPTO_dbg_realloc()
15761 All macros of value have retained their old syntax.
15763 *Richard Levitte and Bodo Moeller*
15765 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15766 ordering of SMIMECapabilities wasn't in "strength order" and there
15767 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15772 * Some ASN1 types with illegal zero length encoding (INTEGER,
15773 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15775 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15777 * Merge in my S/MIME library for OpenSSL. This provides a simple
15778 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15779 functionality to handle multipart/signed properly) and a utility
15780 called 'smime' to call all this stuff. This is based on code I
15781 originally wrote for Celo who have kindly allowed it to be
15782 included in OpenSSL.
15786 * Add variants des_set_key_checked and des_set_key_unchecked of
15787 des_set_key (aka des_key_sched). Global variable des_check_key
15788 decides which of these is called by des_set_key; this way
15789 des_check_key behaves as it always did, but applications and
15790 the library itself, which was buggy for des_check_key == 1,
15791 have a cleaner way to pick the version they need.
15795 * New function PKCS12_newpass() which changes the password of a
15800 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15801 dynamic mix. In both cases the ids can be used as an index into the
15802 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15803 functions so they accept a list of the field values and the
15804 application doesn't need to directly manipulate the X509_TRUST
15809 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15814 * Modify the way the V3 extension code looks up extensions. This now
15815 works in a similar way to the object code: we have some "standard"
15816 extensions in a static table which is searched with OBJ_bsearch()
15817 and the application can add dynamic ones if needed. The file
15818 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15819 updated whenever a new extension is added to the core code and kept
15820 in ext_nid order. There is a simple program 'tabtest.c' which checks
15821 this. New extensions are not added too often so this file can readily
15822 be maintained manually.
15824 There are two big advantages in doing things this way. The extensions
15825 can be looked up immediately and no longer need to be "added" using
15826 X509V3_add_standard_extensions(): this function now does nothing.
15827 Side note: I get *lots* of email saying the extension code doesn't
15828 work because people forget to call this function.
15829 Also no dynamic allocation is done unless new extensions are added:
15830 so if we don't add custom extensions there is no need to call
15831 X509V3_EXT_cleanup().
15835 * Modify enc utility's salting as follows: make salting the default. Add a
15836 magic header, so unsalted files fail gracefully instead of just decrypting
15837 to garbage. This is because not salting is a big security hole, so people
15838 should be discouraged from doing it.
15842 * Fixes and enhancements to the 'x509' utility. It allowed a message
15843 digest to be passed on the command line but it only used this
15844 parameter when signing a certificate. Modified so all relevant
15845 operations are affected by the digest parameter including the
15846 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15847 DSA key was used because it didn't fix the digest.
15851 * Initial certificate chain verify code. Currently tests the untrusted
15852 certificates for consistency with the verify purpose (which is set
15853 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15855 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15856 this is because it will reject chains with invalid extensions whereas
15857 every previous version of OpenSSL and SSLeay made no checks at all.
15859 Trust code: checks the root CA for the relevant trust settings. Trust
15860 settings have an initial value consistent with the verify purpose: e.g.
15861 if the verify purpose is for SSL client use it expects the CA to be
15862 trusted for SSL client use. However the default value can be changed to
15863 permit custom trust settings: one example of this would be to only trust
15864 certificates from a specific "secure" set of CAs.
15866 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15867 which should be used for version portability: especially since the
15868 verify structure is likely to change more often now.
15870 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15871 to set them. If not set then assume SSL clients will verify SSL servers
15874 Two new options to the verify program: -untrusted allows a set of
15875 untrusted certificates to be passed in and -purpose which sets the
15876 intended purpose of the certificate. If a purpose is set then the
15877 new chain verify code is used to check extension consistency.
15881 * Support for the authority information access extension.
15885 * Modify RSA and DSA PEM read routines to transparently handle
15886 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15887 public keys in a format compatible with certificate
15888 SubjectPublicKeyInfo structures. Unfortunately there were already
15889 functions called *_PublicKey_* which used various odd formats so
15890 these are retained for compatibility: however the DSA variants were
15891 never in a public release so they have been deleted. Changed dsa/rsa
15892 utilities to handle the new format: note no releases ever handled public
15893 keys so we should be OK.
15895 The primary motivation for this change is to avoid the same fiasco
15896 that dogs private keys: there are several incompatible private key
15897 formats some of which are standard and some OpenSSL specific and
15898 require various evil hacks to allow partial transparent handling and
15899 even then it doesn't work with DER formats. Given the option anything
15900 other than PKCS#8 should be dumped: but the other formats have to
15901 stay in the name of compatibility.
15903 With public keys and the benefit of hindsight one standard format
15904 is used which works with EVP_PKEY, RSA or DSA structures: though
15905 it clearly returns an error if you try to read the wrong kind of key.
15907 Added a -pubkey option to the 'x509' utility to output the public key.
15908 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15909 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15910 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15911 that do the same as the `EVP_PKEY_assign_*()` except they up the
15912 reference count of the added key (they don't "swallow" the
15917 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15918 CRLs would fail if the file contained no certificates or no CRLs:
15919 added a new function to read in both types and return the number
15920 read: this means that if none are read it will be an error. The
15921 DER versions of the certificate and CRL reader would always fail
15922 because it isn't possible to mix certificates and CRLs in DER format
15923 without choking one or the other routine. Changed this to just read
15924 a certificate: this is the best we can do. Also modified the code
15925 in apps/verify.c to take notice of return codes: it was previously
15926 attempting to read in certificates from NULL pointers and ignoring
15927 any errors: this is one reason why the cert and CRL reader seemed
15928 to work. It doesn't check return codes from the default certificate
15929 routines: these may well fail if the certificates aren't installed.
15933 * Code to support otherName option in GeneralName.
15937 * First update to verify code. Change the verify utility
15938 so it warns if it is passed a self signed certificate:
15939 for consistency with the normal behaviour. X509_verify
15940 has been modified to it will now verify a self signed
15941 certificate if *exactly* the same certificate appears
15942 in the store: it was previously impossible to trust a
15943 single self signed certificate. This means that:
15944 openssl verify ss.pem
15945 now gives a warning about a self signed certificate but
15946 openssl verify -CAfile ss.pem ss.pem
15951 * For servers, store verify_result in SSL_SESSION data structure
15952 (and add it to external session representation).
15953 This is needed when client certificate verifications fails,
15954 but an application-provided verification callback (set by
15955 SSL_CTX_set_cert_verify_callback) allows accepting the session
15956 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15957 but returns 1): When the session is reused, we have to set
15958 ssl->verify_result to the appropriate error code to avoid
15961 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15963 * Fix a bug in the new PKCS#7 code: it didn't consider the
15964 case in PKCS7_dataInit() where the signed PKCS7 structure
15965 didn't contain any existing data because it was being created.
15967 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15969 * Add a salt to the key derivation routines in enc.c. This
15970 forms the first 8 bytes of the encrypted file. Also add a
15971 -S option to allow a salt to be input on the command line.
15975 * New function X509_cmp(). Oddly enough there wasn't a function
15976 to compare two certificates. We do this by working out the SHA1
15977 hash and comparing that. X509_cmp() will be needed by the trust
15982 * SSL_get1_session() is like SSL_get_session(), but increments
15983 the reference count in the SSL_SESSION returned.
15985 *Geoff Thorpe <geoff@eu.c2.net>*
15987 * Fix for 'req': it was adding a null to request attributes.
15988 Also change the X509_LOOKUP and X509_INFO code to handle
15989 certificate auxiliary information.
15993 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15998 * Add the possibility to add extra information to the memory leak
15999 detecting output, to form tracebacks, showing from where each
16000 allocation was originated: CRYPTO_push_info("constant string") adds
16001 the string plus current file name and line number to a per-thread
16002 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16003 is like calling CYRPTO_pop_info() until the stack is empty.
16004 Also updated memory leak detection code to be multi-thread-safe.
16008 * Add options -text and -noout to pkcs7 utility and delete the
16009 encryption options which never did anything. Update docs.
16013 * Add options to some of the utilities to allow the pass phrase
16014 to be included on either the command line (not recommended on
16015 OSes like Unix) or read from the environment. Update the
16016 manpages and fix a few bugs.
16020 * Add a few manpages for some of the openssl commands.
16024 * Fix the -revoke option in ca. It was freeing up memory twice,
16025 leaking and not finding already revoked certificates.
16029 * Extensive changes to support certificate auxiliary information.
16030 This involves the use of X509_CERT_AUX structure and X509_AUX
16031 functions. An X509_AUX function such as PEM_read_X509_AUX()
16032 can still read in a certificate file in the usual way but it
16033 will also read in any additional "auxiliary information". By
16034 doing things this way a fair degree of compatibility can be
16035 retained: existing certificates can have this information added
16036 using the new 'x509' options.
16038 Current auxiliary information includes an "alias" and some trust
16039 settings. The trust settings will ultimately be used in enhanced
16040 certificate chain verification routines: currently a certificate
16041 can only be trusted if it is self signed and then it is trusted
16046 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16047 The problem was that one of the replacement routines had not been working
16048 since SSLeay releases. For now the offending routine has been replaced
16049 with non-optimised assembler. Even so, this now gives around 95%
16050 performance improvement for 1024 bit RSA signs.
16054 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16055 handling. Most clients have the effective key size in bits equal to
16056 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16057 A few however don't do this and instead use the size of the decrypted key
16058 to determine the RC2 key length and the AlgorithmIdentifier to determine
16059 the effective key length. In this case the effective key length can still
16060 be 40 bits but the key length can be 168 bits for example. This is fixed
16061 by manually forcing an RC2 key into the EVP_PKEY structure because the
16062 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16063 the key length and effective key length are equal.
16067 * Add a bunch of functions that should simplify the creation of
16068 X509_NAME structures. Now you should be able to do:
16069 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16070 and have it automatically work out the correct field type and fill in
16071 the structures. The more adventurous can try:
16072 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16073 and it will (hopefully) work out the correct multibyte encoding.
16077 * Change the 'req' utility to use the new field handling and multibyte
16078 copy routines. Before the DN field creation was handled in an ad hoc
16079 way in req, ca, and x509 which was rather broken and didn't support
16080 BMPStrings or UTF8Strings. Since some software doesn't implement
16081 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16082 using the dirstring_type option. See the new comment in the default
16083 openssl.cnf for more info.
16087 * Make crypto/rand/md_rand.c more robust:
16088 - Assure unique random numbers after fork().
16089 - Make sure that concurrent threads access the global counter and
16090 md serializably so that we never lose entropy in them
16091 or use exactly the same state in multiple threads.
16092 Access to the large state is not always serializable because
16093 the additional locking could be a performance killer, and
16094 md should be large enough anyway.
16098 * New file apps/app_rand.c with commonly needed functionality
16099 for handling the random seed file.
16101 Use the random seed file in some applications that previously did not:
16103 dsaparam -genkey (which also ignored its '-rand' option),
16106 x509 (when signing).
16107 Except on systems with /dev/urandom, it is crucial to have a random
16108 seed file at least for key creation, DSA signing, and for DH exchanges;
16109 for RSA signatures we could do without one.
16111 gendh and gendsa (unlike genrsa) used to read only the first byte
16112 of each file listed in the '-rand' option. The function as previously
16113 found in genrsa is now in app_rand.c and is used by all programs
16114 that support '-rand'.
16118 * In RAND_write_file, use mode 0600 for creating files;
16119 don't just chmod when it may be too late.
16123 * Report an error from X509_STORE_load_locations
16124 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16128 * New function ASN1_mbstring_copy() this copies a string in either
16129 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16130 into an ASN1_STRING type. A mask of permissible types is passed
16131 and it chooses the "minimal" type to use or an error if not type
16136 * Add function equivalents to the various macros in asn1.h. The old
16137 macros are retained with an `M_` prefix. Code inside the library can
16138 use the `M_` macros. External code (including the openssl utility)
16139 should *NOT* in order to be "shared library friendly".
16143 * Add various functions that can check a certificate's extensions
16144 to see if it usable for various purposes such as SSL client,
16145 server or S/MIME and CAs of these types. This is currently
16146 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16147 verification. Also added a -purpose flag to x509 utility to
16148 print out all the purposes.
16152 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16157 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16158 for, obtain and decode and extension and obtain its critical flag.
16159 This allows all the necessary extension code to be handled in a
16160 single function call.
16164 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16165 platforms. See crypto/rc4/rc4_enc.c for further details.
16169 * New -noout option to asn1parse. This causes no output to be produced
16170 its main use is when combined with -strparse and -out to extract data
16171 from a file (which may not be in ASN.1 format).
16175 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16176 when producing the local key id.
16178 *Richard Levitte <levitte@stacken.kth.se>*
16180 * New option -dhparam in s_server. This allows a DH parameter file to be
16181 stated explicitly. If it is not stated then it tries the first server
16182 certificate file. The previous behaviour hard coded the filename
16187 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16188 a public key to be input or output. For example:
16189 openssl rsa -in key.pem -pubout -out pubkey.pem
16190 Also added necessary DSA public key functions to handle this.
16194 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16195 in the message. This was handled by allowing
16196 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16198 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16200 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16201 to the end of the strings whereas this didn't. This would cause problems
16202 if strings read with d2i_ASN1_bytes() were later modified.
16204 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16206 * Fix for base64 decode bug. When a base64 bio reads only one line of
16207 data and it contains EOF it will end up returning an error. This is
16208 caused by input 46 bytes long. The cause is due to the way base64
16209 BIOs find the start of base64 encoded data. They do this by trying a
16210 trial decode on each line until they find one that works. When they
16211 do a flag is set and it starts again knowing it can pass all the
16212 data directly through the decoder. Unfortunately it doesn't reset
16213 the context it uses. This means that if EOF is reached an attempt
16214 is made to pass two EOFs through the context and this causes the
16215 resulting error. This can also cause other problems as well. As is
16216 usual with these problems it takes *ages* to find and the fix is
16217 trivial: move one line.
16219 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16221 * Ugly workaround to get s_client and s_server working under Windows. The
16222 old code wouldn't work because it needed to select() on sockets and the
16223 tty (for keypresses and to see if data could be written). Win32 only
16224 supports select() on sockets so we select() with a 1s timeout on the
16225 sockets and then see if any characters are waiting to be read, if none
16226 are present then we retry, we also assume we can always write data to
16227 the tty. This isn't nice because the code then blocks until we've
16228 received a complete line of data and it is effectively polling the
16229 keyboard at 1s intervals: however it's quite a bit better than not
16230 working at all :-) A dedicated Windows application might handle this
16231 with an event loop for example.
16235 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16236 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16237 will be called when RSA_sign() and RSA_verify() are used. This is useful
16238 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16239 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16240 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16241 This necessitated the support of an extra signature type NID_md5_sha1
16242 for SSL signatures and modifications to the SSL library to use it instead
16243 of calling RSA_public_decrypt() and RSA_private_encrypt().
16247 * Add new -verify -CAfile and -CApath options to the crl program, these
16248 will lookup a CRL issuers certificate and verify the signature in a
16249 similar way to the verify program. Tidy up the crl program so it
16250 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16251 less strict. It will now permit CRL extensions even if it is not
16252 a V2 CRL: this will allow it to tolerate some broken CRLs.
16256 * Initialize all non-automatic variables each time one of the openssl
16257 sub-programs is started (this is necessary as they may be started
16258 multiple times from the "OpenSSL>" prompt).
16260 *Lennart Bang, Bodo Moeller*
16262 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16263 removing all other RSA functionality (this is what NO_RSA does). This
16264 is so (for example) those in the US can disable those operations covered
16265 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16270 * Non-copying interface to BIO pairs.
16271 (still largely untested)
16275 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16276 ASCII string. This was handled independently in various places before.
16280 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16281 UTF8 strings a character at a time.
16285 * Use client_version from client hello to select the protocol
16286 (s23_srvr.c) and for RSA client key exchange verification
16287 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16291 * Add various utility functions to handle SPKACs, these were previously
16292 handled by poking round in the structure internals. Added new function
16293 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16294 print, verify and generate SPKACs. Based on an original idea from
16295 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16299 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16303 * Allow the config file extension section to be overwritten on the
16304 command line. Based on an original idea from Massimiliano Pala
16305 <madwolf@comune.modena.it>. The new option is called -extensions
16306 and can be applied to ca, req and x509. Also -reqexts to override
16307 the request extensions in req and -crlexts to override the crl extensions
16312 * Add new feature to the SPKAC handling in ca. Now you can include
16313 the same field multiple times by preceding it by "XXXX." for example:
16316 this is the same syntax as used in the req config file.
16320 * Allow certificate extensions to be added to certificate requests. These
16321 are specified in a 'req_extensions' option of the req section of the
16322 config file. They can be printed out with the -text option to req but
16323 are otherwise ignored at present.
16327 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16328 data read consists of only the final block it would not decrypted because
16329 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16330 A misplaced 'break' also meant the decrypted final block might not be
16331 copied until the next read.
16335 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16336 a few extra parameters to the DH structure: these will be useful if
16337 for example we want the value of 'q' or implement X9.42 DH.
16341 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16342 provides hooks that allow the default DSA functions or functions on a
16343 "per key" basis to be replaced. This allows hardware acceleration and
16344 hardware key storage to be handled without major modification to the
16345 library. Also added low level modexp hooks and CRYPTO_EX structure and
16346 associated functions.
16350 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16351 as "read only": it can't be written to and the buffer it points to will
16352 not be freed. Reading from a read only BIO is much more efficient than
16353 a normal memory BIO. This was added because there are several times when
16354 an area of memory needs to be read from a BIO. The previous method was
16355 to create a memory BIO and write the data to it, this results in two
16356 copies of the data and an O(n^2) reading algorithm. There is a new
16357 function BIO_new_mem_buf() which creates a read only memory BIO from
16358 an area of memory. Also modified the PKCS#7 routines to use read only
16363 * Bugfix: ssl23_get_client_hello did not work properly when called in
16364 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16365 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16366 but a retry condition occurred while trying to read the rest.
16370 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16371 NID_pkcs7_encrypted by default: this was wrong since this should almost
16372 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16373 the encrypted data type: this is a more sensible place to put it and it
16374 allows the PKCS#12 code to be tidied up that duplicated this
16379 * Changed obj_dat.pl script so it takes its input and output files on
16380 the command line. This should avoid shell escape redirection problems
16385 * Initial support for certificate extension requests, these are included
16386 in things like Xenroll certificate requests. Included functions to allow
16387 extensions to be obtained and added.
16391 * -crlf option to s_client and s_server for sending newlines as
16392 CRLF (as required by many protocols).
16396 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16398 * Install libRSAglue.a when OpenSSL is built with RSAref.
16400 *Ralf S. Engelschall*
16402 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16404 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16406 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16411 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16412 DH parameters/keys (q is lost during that conversion, but the resulting
16413 DH parameters contain its length).
16415 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16416 much faster than DH_generate_parameters (which creates parameters
16417 where `p = 2*q + 1`), and also the smaller q makes DH computations
16418 much more efficient (160-bit exponentiation instead of 1024-bit
16419 exponentiation); so this provides a convenient way to support DHE
16420 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16421 utter importance to use
16422 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16424 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16425 when such DH parameters are used, because otherwise small subgroup
16426 attacks may become possible!
16430 * Avoid memory leak in i2d_DHparams.
16434 * Allow the -k option to be used more than once in the enc program:
16435 this allows the same encrypted message to be read by multiple recipients.
16439 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16440 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16441 it will always use the numerical form of the OID, even if it has a short
16446 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16447 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16448 otherwise bn_mod_exp was called. In the case of hardware keys for example
16449 no private key components need be present and it might store extra data
16450 in the RSA structure, which cannot be accessed from bn_mod_exp.
16451 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16452 private key operations.
16456 * Added support for SPARC Linux.
16460 * pem_password_cb function type incompatibly changed from
16461 typedef int pem_password_cb(char *buf, int size, int rwflag);
16463 ....(char *buf, int size, int rwflag, void *userdata);
16464 so that applications can pass data to their callbacks:
16465 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16466 additional void * argument, which is just handed through whenever
16467 the password callback is called.
16469 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16471 New function SSL_CTX_set_default_passwd_cb_userdata.
16473 Compatibility note: As many C implementations push function arguments
16474 onto the stack in reverse order, the new library version is likely to
16475 interoperate with programs that have been compiled with the old
16476 pem_password_cb definition (PEM_whatever takes some data that
16477 happens to be on the stack as its last argument, and the callback
16478 just ignores this garbage); but there is no guarantee whatsoever that
16481 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16482 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16483 problems not only on Windows, but also on some Unix platforms.
16484 To avoid problematic command lines, these definitions are now in an
16485 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16486 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16490 * MIPS III/IV assembler module is reimplemented.
16494 * More DES library cleanups: remove references to srand/rand and
16495 delete an unused file.
16499 * Add support for the free Netwide assembler (NASM) under Win32,
16500 since not many people have MASM (ml) and it can be hard to obtain.
16501 This is currently experimental but it seems to work OK and pass all
16502 the tests. Check out INSTALL.W32 for info.
16506 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16507 without temporary keys kept an extra copy of the server key,
16508 and connections with temporary keys did not free everything in case
16513 * New function RSA_check_key and new openssl rsa option -check
16514 for verifying the consistency of RSA keys.
16516 *Ulf Moeller, Bodo Moeller*
16518 * Various changes to make Win32 compile work:
16519 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16520 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16521 comparison" warnings.
16522 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16526 * Add a debugging option to PKCS#5 v2 key generation function: when
16527 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16528 derived keys are printed to stderr.
16532 * Copy the flags in ASN1_STRING_dup().
16534 *Roman E. Pavlov <pre@mo.msk.ru>*
16536 * The x509 application mishandled signing requests containing DSA
16537 keys when the signing key was also DSA and the parameters didn't match.
16539 It was supposed to omit the parameters when they matched the signing key:
16540 the verifying software was then supposed to automatically use the CA's
16541 parameters if they were absent from the end user certificate.
16543 Omitting parameters is no longer recommended. The test was also
16544 the wrong way round! This was probably due to unusual behaviour in
16545 EVP_cmp_parameters() which returns 1 if the parameters match.
16546 This meant that parameters were omitted when they *didn't* match and
16547 the certificate was useless. Certificates signed with 'ca' didn't have
16550 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16552 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16553 The interface is as follows:
16554 Applications can use
16555 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16556 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16557 "off" is now the default.
16558 The library internally uses
16559 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16560 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16561 to disable memory-checking temporarily.
16563 Some inconsistent states that previously were possible (and were
16564 even the default) are now avoided.
16566 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16567 with each memory chunk allocated; this is occasionally more helpful
16568 than just having a counter.
16570 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16572 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16577 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16578 which largely parallels "options", but is for changing API behaviour,
16579 whereas "options" are about protocol behaviour.
16580 Initial "mode" flags are:
16582 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16583 a single record has been written.
16584 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16585 retries use the same buffer location.
16586 (But all of the contents must be
16591 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16594 * Fix problems with no-hmac etc.
16596 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16598 * New functions RSA_get_default_method(), RSA_set_method() and
16599 RSA_get_method(). These allows replacement of RSA_METHODs without having
16600 to mess around with the internals of an RSA structure.
16604 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16605 Also really enable memory leak checks in openssl.c and in some
16608 *Chad C. Mulligan, Bodo Moeller*
16610 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16611 up the length of negative integers. This has now been simplified to just
16612 store the length when it is first determined and use it later, rather
16613 than trying to keep track of where data is copied and updating it to
16615 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16617 * Add a new function PKCS7_signatureVerify. This allows the verification
16618 of a PKCS#7 signature but with the signing certificate passed to the
16619 function itself. This contrasts with PKCS7_dataVerify which assumes the
16620 certificate is present in the PKCS#7 structure. This isn't always the
16621 case: certificates can be omitted from a PKCS#7 structure and be
16622 distributed by "out of band" means (such as a certificate database).
16626 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16627 function prototypes in pem.h, also change util/mkdef.pl to add the
16628 necessary function names.
16632 * mk1mf.pl (used by Windows builds) did not properly read the
16633 options set by Configure in the top level Makefile, and Configure
16634 was not even able to write more than one option correctly.
16635 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16639 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16640 file to be loaded from a BIO or FILE pointer. The BIO version will
16641 for example allow memory BIOs to contain config info.
16645 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16646 Whoever hopes to achieve shared-library compatibility across versions
16647 must use this, not the compile-time macro.
16648 (Exercise 0.9.4: Which is the minimum library version required by
16650 Note: All this applies only to multi-threaded programs, others don't
16655 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16656 through a BIO pair triggered the default case, i.e.
16657 SSLerr(...,SSL_R_UNKNOWN_STATE).
16661 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16662 can use the SSL library even if none of the specific BIOs is
16667 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16668 for the encoded length.
16670 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16672 * Add initial documentation of the X509V3 functions.
16676 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16677 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16678 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16679 secure PKCS#8 private key format with a high iteration count.
16683 * Fix determination of Perl interpreter: A perl or perl5
16684 *directory* in $PATH was also accepted as the interpreter.
16686 *Ralf S. Engelschall*
16688 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16689 wrong with it but it was very old and did things like calling
16690 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16691 unusual formatting.
16695 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16696 to use the new extension code.
16700 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16701 with macros. This should make it easier to change their form, add extra
16702 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16707 * Add to configuration table a new entry that can specify an alternative
16708 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16709 according to Mark Crispin <MRC@Panda.COM>.
16713 * DES CBC did not update the IV. Weird.
16717 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16718 Changing the behaviour of the former might break existing programs --
16719 where IV updating is needed, des_ncbc_encrypt can be used.
16722 * When bntest is run from "make test" it drives bc to check its
16723 calculations, as well as internally checking them. If an internal check
16724 fails, it needs to cause bc to give a non-zero result or make test carries
16725 on without noticing the failure. Fixed.
16729 * DES library cleanups.
16733 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16734 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16735 ciphers. NOTE: although the key derivation function has been verified
16736 against some published test vectors it has not been extensively tested
16737 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16742 * Instead of "mkdir -p", which is not fully portable, use new
16743 Perl script "util/mkdir-p.pl".
16747 * Rewrite the way password based encryption (PBE) is handled. It used to
16748 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16749 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16750 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16751 the 'parameter' field of the AlgorithmIdentifier is passed to the
16752 underlying key generation function so it must do its own ASN1 parsing.
16753 This has also changed the EVP_PBE_CipherInit() function which now has a
16754 'parameter' argument instead of literal salt and iteration count values
16755 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16759 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16760 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16761 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16762 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16763 value was just used as a "magic string" and not used directly its
16764 value doesn't matter.
16768 * Introduce some semblance of const correctness to BN. Shame C doesn't
16773 * "linux-sparc64" configuration (ultrapenguin).
16775 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16776 "linux-sparc" configuration.
16778 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16780 * config now generates no-xxx options for missing ciphers.
16784 * Support the EBCDIC character set (work in progress).
16785 File ebcdic.c not yet included because it has a different license.
16787 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16789 * Support BS2000/OSD-POSIX.
16791 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16793 * Make callbacks for key generation use `void *` instead of `char *`.
16797 * Make S/MIME samples compile (not yet tested).
16801 * Additional typesafe stacks.
16805 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16809 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16811 * New configuration variant "sco5-gcc".
16813 * Updated some demos.
16815 *Sean O Riordain, Wade Scholine*
16817 * Add missing BIO_free at exit of pkcs12 application.
16821 * Fix memory leak in conf.c.
16825 * Updates for Win32 to assembler version of MD5.
16829 * Set #! path to perl in apps/der_chop to where we found it
16830 instead of using a fixed path.
16834 * SHA library changes for irix64-mips4-cc.
16838 * Improvements for VMS support.
16842 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16844 * Bignum library bug fix. IRIX 6 passes "make test" now!
16845 This also avoids the problems with SC4.2 and unpatched SC5.
16847 *Andy Polyakov <appro@fy.chalmers.se>*
16849 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16850 These are required because of the typesafe stack would otherwise break
16851 existing code. If old code used a structure member which used to be STACK
16852 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16853 sk_num or sk_value it would produce an error because the num, data members
16854 are not present in STACK_OF. Now it just produces a warning. sk_set
16855 replaces the old method of assigning a value to sk_value
16856 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16857 that does this will no longer work (and should use sk_set instead) but
16858 this could be regarded as a "questionable" behaviour anyway.
16862 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16863 correctly handle encrypted S/MIME data.
16867 * Change type of various DES function arguments from des_cblock
16868 (which means, in function argument declarations, pointer to char)
16869 to des_cblock * (meaning pointer to array with 8 char elements),
16870 which allows the compiler to do more typechecking; it was like
16871 that back in SSLeay, but with lots of ugly casts.
16873 Introduce new type const_des_cblock.
16877 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16878 problems: find RecipientInfo structure that matches recipient certificate
16879 and initialise the ASN1 structures properly based on passed cipher.
16883 * Belatedly make the BN tests actually check the results.
16887 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16888 to and from BNs: it was completely broken. New compilation option
16889 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16890 key elements as negative integers.
16894 * Reorganize and speed up MD5.
16896 *Andy Polyakov <appro@fy.chalmers.se>*
16900 *Richard Levitte <richard@levitte.org>*
16902 * New option -out to asn1parse to allow the parsed structure to be
16903 output to a file. This is most useful when combined with the -strparse
16904 option to examine the output of things like OCTET STRINGS.
16908 * Make SSL library a little more fool-proof by not requiring any longer
16909 that `SSL_set_{accept,connect}_state` be called before
16910 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16911 in many applications because usually everything *appeared* to work as
16912 intended anyway -- now it really works as intended).
16916 * Move openssl.cnf out of lib/.
16920 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16921 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16922 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16924 *Ralf S. Engelschall*
16926 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16927 handle PKCS#7 enveloped data properly.
16929 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16931 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16932 copying pointers. The cert_st handling is changed by this in
16933 various ways (and thus what used to be known as ctx->default_cert
16934 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16935 any longer when s->cert does not give us what we need).
16936 ssl_cert_instantiate becomes obsolete by this change.
16937 As soon as we've got the new code right (possibly it already is?),
16938 we have solved a couple of bugs of the earlier code where s->cert
16939 was used as if it could not have been shared with other SSL structures.
16941 Note that using the SSL API in certain dirty ways now will result
16942 in different behaviour than observed with earlier library versions:
16943 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16944 does not influence s as it used to.
16946 In order to clean up things more thoroughly, inside SSL_SESSION
16947 we don't use CERT any longer, but a new structure SESS_CERT
16948 that holds per-session data (if available); currently, this is
16949 the peer's certificate chain and, for clients, the server's certificate
16950 and temporary key. CERT holds only those values that can have
16951 meaningful defaults in an SSL_CTX.
16955 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16956 from the internal representation. Various PKCS#7 fixes: remove some
16957 evil casts and set the enc_dig_alg field properly based on the signing
16962 * Allow PKCS#12 password to be set from the command line or the
16963 environment. Let 'ca' get its config file name from the environment
16964 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16969 * Allow certificate policies extension to use an IA5STRING for the
16970 organization field. This is contrary to the PKIX definition but
16971 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16976 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16977 without disallowing inline assembler and the like for non-pedantic builds.
16981 * Support Borland C++ builder.
16983 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16989 * SHA-1 cleanups and performance enhancements.
16991 *Andy Polyakov <appro@fy.chalmers.se>*
16993 * Sparc v8plus assembler for the bignum library.
16995 *Andy Polyakov <appro@fy.chalmers.se>*
16997 * Accept any -xxx and +xxx compiler options in Configure.
17001 * Update HPUX configuration.
17005 * Add missing `sk_<type>_unshift()` function to safestack.h
17007 *Ralf S. Engelschall*
17009 * New function SSL_CTX_use_certificate_chain_file that sets the
17010 "extra_cert"s in addition to the certificate. (This makes sense
17011 only for "PEM" format files, as chains as a whole are not
17016 * Support verify_depth from the SSL API.
17017 x509_vfy.c had what can be considered an off-by-one-error:
17018 Its depth (which was not part of the external interface)
17019 was actually counting the number of certificates in a chain;
17020 now it really counts the depth.
17024 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17025 instead of X509err, which often resulted in confusing error
17026 messages since the error codes are not globally unique
17027 (e.g. an alleged error in ssl3_accept when a certificate
17028 didn't match the private key).
17030 * New function SSL_CTX_set_session_id_context that allows to set a default
17031 value (so that you don't need SSL_set_session_id_context for each
17032 connection using the SSL_CTX).
17036 * OAEP decoding bug fix.
17040 * Support INSTALL_PREFIX for package builders, as proposed by
17045 * New Configure options "threads" and "no-threads". For systems
17046 where the proper compiler options are known (currently Solaris
17047 and Linux), "threads" is the default.
17051 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17055 * Install various scripts to $(OPENSSLDIR)/misc, not to
17056 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17057 such as /usr/local/bin.
17061 * "make linux-shared" to build shared libraries.
17063 *Niels Poppe <niels@netbox.org>*
17065 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17069 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17070 extension adding in x509 utility.
17074 * Remove NOPROTO sections and error code comments.
17078 * Partial rewrite of the DEF file generator to now parse the ANSI
17083 * New Configure options --prefix=DIR and --openssldir=DIR.
17087 * Complete rewrite of the error code script(s). It is all now handled
17088 by one script at the top level which handles error code gathering,
17089 header rewriting and C source file generation. It should be much better
17090 than the old method: it now uses a modified version of Ulf's parser to
17091 read the ANSI prototypes in all header files (thus the old K&R definitions
17092 aren't needed for error creation any more) and do a better job of
17093 translating function codes into names. The old 'ASN1 error code embedded
17094 in a comment' is no longer necessary and it doesn't use .err files which
17095 have now been deleted. Also the error code call doesn't have to appear all
17096 on one line (which resulted in some large lines...).
17100 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17104 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17105 0 (which usually indicates a closed connection), but continue reading.
17109 * Fix some race conditions.
17113 * Add support for CRL distribution points extension. Add Certificate
17114 Policies and CRL distribution points documentation.
17118 * Move the autogenerated header file parts to crypto/opensslconf.h.
17122 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17123 8 of keying material. Merlin has also confirmed interop with this fix
17124 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17126 *Merlin Hughes <merlin@baltimore.ie>*
17128 * Fix lots of warnings.
17130 *Richard Levitte <levitte@stacken.kth.se>*
17132 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17133 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17135 *Richard Levitte <levitte@stacken.kth.se>*
17137 * Fix problems with sizeof(long) == 8.
17139 *Andy Polyakov <appro@fy.chalmers.se>*
17141 * Change functions to ANSI C.
17145 * Fix typos in error codes.
17147 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17149 * Remove defunct assembler files from Configure.
17153 * SPARC v8 assembler BIGNUM implementation.
17155 *Andy Polyakov <appro@fy.chalmers.se>*
17157 * Support for Certificate Policies extension: both print and set.
17158 Various additions to support the r2i method this uses.
17162 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17163 return a const string when you are expecting an allocated buffer.
17167 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17168 types DirectoryString and DisplayText.
17172 * Add code to allow r2i extensions to access the configuration database,
17173 add an LHASH database driver and add several ctx helper functions.
17177 * Fix an evil bug in bn_expand2() which caused various BN functions to
17178 fail when they extended the size of a BIGNUM.
17182 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17183 support typesafe stack.
17187 * Fix typo in SSL_[gs]et_options().
17189 *Nils Frostberg <nils@medcom.se>*
17191 * Delete various functions and files that belonged to the (now obsolete)
17192 old X509V3 handling code.
17196 * New Configure option "rsaref".
17200 * Don't auto-generate pem.h.
17204 * Introduce type-safe ASN.1 SETs.
17208 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17210 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17212 * Introduce type-safe STACKs. This will almost certainly break lots of code
17213 that links with OpenSSL (well at least cause lots of warnings), but fear
17214 not: the conversion is trivial, and it eliminates loads of evil casts. A
17215 few STACKed things have been converted already. Feel free to convert more.
17216 In the fullness of time, I'll do away with the STACK type altogether.
17220 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17221 specified in `<certfile>` by updating the entry in the index.txt file.
17222 This way one no longer has to edit the index.txt file manually for
17223 revoking a certificate. The -revoke option does the gory details now.
17225 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17227 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17228 `-text` option at all and this way the `-noout -text` combination was
17229 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17231 *Ralf S. Engelschall*
17233 * Make sure a corresponding plain text error message exists for the
17234 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17235 verify callback function determined that a certificate was revoked.
17237 *Ralf S. Engelschall*
17239 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17240 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17241 all available ciphers including rc5, which was forgotten until now.
17242 In order to let the testing shell script know which algorithms
17243 are available, a new (up to now undocumented) command
17244 `openssl list-cipher-commands` is used.
17248 * Bugfix: s_client occasionally would sleep in select() when
17249 it should have checked SSL_pending() first.
17253 * New functions DSA_do_sign and DSA_do_verify to provide access to
17254 the raw DSA values prior to ASN.1 encoding.
17258 * Tweaks to Configure
17260 *Niels Poppe <niels@netbox.org>*
17262 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17267 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17271 * New config option to avoid instructions that are illegal on the 80386.
17272 The default code is faster, but requires at least a 486.
17276 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17277 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17278 same as SSL2_VERSION anyway.
17282 * New "-showcerts" option for s_client.
17286 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17287 application. Various cleanups and fixes.
17291 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17292 modify error routines to work internally. Add error codes and PBE init
17293 to library startup routines.
17297 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17298 packing functions to asn1 and evp. Changed function names and error
17299 codes along the way.
17303 * PKCS12 integration: and so it begins... First of several patches to
17304 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17305 objects to objects.h
17309 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17310 and display support for Thawte strong extranet extension.
17314 * Add LinuxPPC support.
17316 *Jeff Dubrule <igor@pobox.org>*
17318 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17319 bn_div_words in alpha.s.
17321 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17323 * Make sure the RSA OAEP test is skipped under -DRSAref because
17324 OAEP isn't supported when OpenSSL is built with RSAref.
17326 *Ulf Moeller <ulf@fitug.de>*
17328 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17329 so they no longer are missing under -DNOPROTO.
17331 *Soren S. Jorvang <soren@t.dk>*
17333 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17335 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17336 doesn't work when the session is reused. Coming soon!
17340 * Fix a security hole, that allows sessions to be reused in the wrong
17341 context thus bypassing client cert protection! All software that uses
17342 client certs and session caches in multiple contexts NEEDS PATCHING to
17343 allow session reuse! A fuller solution is in the works.
17345 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17347 * Some more source tree cleanups (removed obsolete files
17348 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17349 permission on "config" script to be executable) and a fix for the INSTALL
17352 *Ulf Moeller <ulf@fitug.de>*
17354 * Remove some legacy and erroneous uses of malloc, free instead of
17357 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17359 * Make rsa_oaep_test return non-zero on error.
17361 *Ulf Moeller <ulf@fitug.de>*
17363 * Add support for native Solaris shared libraries. Configure
17364 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17365 if someone would make that last step automatic.
17367 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17369 * ctx_size was not built with the right compiler during "make links". Fixed.
17373 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17374 except NULL ciphers". This means the default cipher list will no longer
17375 enable NULL ciphers. They need to be specifically enabled e.g. with
17376 the string "DEFAULT:eNULL".
17380 * Fix to RSA private encryption routines: if p < q then it would
17381 occasionally produce an invalid result. This will only happen with
17382 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17386 * Be less restrictive and allow also `perl util/perlpath.pl
17387 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17388 because this way one can also use an interpreter named `perl5' (which is
17389 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17390 installed as `perl').
17392 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17394 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17396 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17398 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17399 advapi32.lib to Win32 build and change the pem test comparison
17400 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17401 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17402 and crypto/des/ede_cbcm_enc.c.
17406 * DES quad checksum was broken on big-endian architectures. Fixed.
17410 * Comment out two functions in bio.h that aren't implemented. Fix up the
17411 Win32 test batch file so it (might) work again. The Win32 test batch file
17412 is horrible: I feel ill....
17416 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17417 in e_os.h. Audit of header files to check ANSI and non ANSI
17418 sections: 10 functions were absent from non ANSI section and not exported
17419 from Windows DLLs. Fixed up libeay.num for new functions.
17423 * Make `openssl version' output lines consistent.
17425 *Ralf S. Engelschall*
17427 * Fix Win32 symbol export lists for BIO functions: Added
17428 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17429 to ms/libeay{16,32}.def.
17431 *Ralf S. Engelschall*
17433 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17434 fine under Unix and passes some trivial tests I've now added. But the
17435 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17436 added to make sure no one expects that this stuff really works in the
17437 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17438 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17441 *Ralf S. Engelschall*
17443 * Fix the generation of two part addresses in perl.
17445 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17447 * Add config entry for Linux on MIPS.
17449 *John Tobey <jtobey@channel1.com>*
17451 * Make links whenever Configure is run, unless we are on Windoze.
17455 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17456 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17461 * Add a useful kludge to allow package maintainers to specify compiler and
17462 other platforms details on the command line without having to patch the
17463 Configure script every time: One now can use
17464 `perl Configure <id>:<details>`,
17465 i.e. platform ids are allowed to have details appended
17466 to them (separated by colons). This is treated as there would be a static
17467 pre-configured entry in Configure's %table under key `<id>` with value
17468 `<details>` and `perl Configure <id>` is called. So, when you want to
17469 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17470 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17471 now, which overrides the FreeBSD-elf entry on-the-fly.
17473 *Ralf S. Engelschall*
17475 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17479 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17480 on the `perl Configure ...' command line. This way one can compile
17481 OpenSSL libraries with Position Independent Code (PIC) which is needed
17482 for linking it into DSOs.
17484 *Ralf S. Engelschall*
17486 * Remarkably, export ciphers were totally broken and no-one had noticed!
17491 * Cleaned up the LICENSE document: The official contact for any license
17492 questions now is the OpenSSL core team under openssl-core@openssl.org.
17493 And add a paragraph about the dual-license situation to make sure people
17494 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17495 to the OpenSSL toolkit.
17497 *Ralf S. Engelschall*
17499 * General source tree makefile cleanups: Made `making xxx in yyy...'
17500 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17501 Additionally cleaned up the `make links' target: Remove unnecessary
17502 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17503 to speed processing and no longer clutter the display with confusing
17504 stuff. Instead only the actually done links are displayed.
17506 *Ralf S. Engelschall*
17508 * Permit null encryption ciphersuites, used for authentication only. It used
17509 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17510 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17515 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17516 signed attributes when verifying signatures (this would break them),
17517 the detached data encoding was wrong and public keys obtained using
17518 X509_get_pubkey() weren't freed.
17522 * Add text documentation for the BUFFER functions. Also added a work around
17523 to a Win95 console bug. This was triggered by the password read stuff: the
17524 last character typed gets carried over to the next fread(). If you were
17525 generating a new cert request using 'req' for example then the last
17526 character of the passphrase would be CR which would then enter the first
17531 * Added the new 'Includes OpenSSL Cryptography Software' button as
17532 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17533 button and can be used by applications based on OpenSSL to show the
17534 relationship to the OpenSSL project.
17536 *Ralf S. Engelschall*
17538 * Remove confusing variables in function signatures in files
17539 ssl/ssl_lib.c and ssl/ssl.h.
17541 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17543 * Don't install bss_file.c under PREFIX/include/
17545 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17547 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17548 functions that return function pointers and has support for NT specific
17549 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17550 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17551 unsigned to signed types: this was killing the Win32 compile.
17555 * Add new certificate file to stack functions,
17556 SSL_add_dir_cert_subjects_to_stack() and
17557 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17558 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17559 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17560 This means that Apache-SSL and similar packages don't have to mess around
17561 to add as many CAs as they want to the preferred list.
17565 * Experiment with doxygen documentation. Currently only partially applied to
17567 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17568 openssl.doxy as the configuration file.
17572 * Get rid of remaining C++-style comments which strict C compilers hate.
17574 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17576 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17577 compiled in by default: it has problems with large keys.
17581 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17582 DH private keys and/or callback functions which directly correspond to
17583 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17584 is needed for applications which have to configure certificates on a
17585 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17587 For the RSA certificate situation is makes no difference, but
17588 for the DSA certificate situation this fixes the "no shared cipher"
17589 problem where the OpenSSL cipher selection procedure failed because the
17590 temporary keys were not overtaken from the context and the API provided
17591 no way to reconfigure them.
17592 The new functions now let applications reconfigure the stuff and they
17593 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17594 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17595 non-public-API function ssl_cert_instantiate() is used as a helper
17596 function and also to reduce code redundancy inside ssl_rsa.c.
17598 *Ralf S. Engelschall*
17600 * Move s_server -dcert and -dkey options out of the undocumented feature
17601 area because they are useful for the DSA situation and should be
17602 recognized by the users.
17604 *Ralf S. Engelschall*
17606 * Fix the cipher decision scheme for export ciphers: the export bits are
17607 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17608 SSL_EXP_MASK. So, the original variable has to be used instead of the
17609 already masked variable.
17611 *Richard Levitte <levitte@stacken.kth.se>*
17613 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17615 *Richard Levitte <levitte@stacken.kth.se>*
17617 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17618 from `int` to `unsigned int` because it is a length and initialized by
17619 EVP_DigestFinal() which expects an `unsigned int *`.
17621 *Richard Levitte <levitte@stacken.kth.se>*
17623 * Don't hard-code path to Perl interpreter on shebang line of Configure
17624 script. Instead use the usual Shell->Perl transition trick.
17626 *Ralf S. Engelschall*
17628 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17629 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17630 -noout -modulus' as it's already the case for `openssl rsa -noout
17631 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17632 currently the public key is printed (a decision which was already done by
17633 `openssl dsa -modulus' in the past) which serves a similar purpose.
17634 Additionally the NO_RSA no longer completely removes the whole -modulus
17635 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17638 *Ralf S. Engelschall*
17640 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17641 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17643 *Arne Ansper <arne@ats.cyber.ee>*
17645 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17646 to be added. Now both 'req' and 'ca' can use new objects defined in the
17651 * Add cool BIO that does syslog (or event log on NT).
17653 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17655 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17656 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17657 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17658 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17662 * Add preliminary config info for new extension code.
17666 * Make RSA_NO_PADDING really use no padding.
17668 *Ulf Moeller <ulf@fitug.de>*
17670 * Generate errors when private/public key check is done.
17674 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17675 for some CRL extensions and new objects added.
17679 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17680 key usage extension and fuller support for authority key id.
17684 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17685 padding method for RSA, which is recommended for new applications in PKCS
17686 #1 v2.0 (RFC 2437, October 1998).
17687 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17688 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17689 against Bleichbacher's attack on RSA.
17690 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17693 * Updates to the new SSL compression code
17695 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17697 * Fix so that the version number in the master secret, when passed
17698 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17699 (because the server will not accept higher), that the version number
17700 is 0x03,0x01, not 0x03,0x00
17702 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17704 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17705 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17706 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17710 * Support for RAW extensions where an arbitrary extension can be
17711 created by including its DER encoding. See apps/openssl.cnf for
17716 * Make sure latest Perl versions don't interpret some generated C array
17717 code as Perl array code in the crypto/err/err_genc.pl script.
17719 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17721 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17722 not many people have the assembler. Various Win32 compilation fixes and
17723 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17724 build instructions.
17728 * Modify configure script 'Configure' to automatically create crypto/date.h
17729 file under Win32 and also build pem.h from pem.org. New script
17730 util/mkfiles.pl to create the MINFO file on environments that can't do a
17731 'make files': perl util/mkfiles.pl >MINFO should work.
17735 * Major rework of DES function declarations, in the pursuit of correctness
17736 and purity. As a result, many evil casts evaporated, and some weirdness,
17737 too. You may find this causes warnings in your code. Zapping your evil
17738 casts will probably fix them. Mostly.
17742 * Fix for a typo in asn1.h. Bug fix to object creation script
17743 obj_dat.pl. It considered a zero in an object definition to mean
17744 "end of object": none of the objects in objects.h have any zeros
17745 so it wasn't spotted.
17747 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17749 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17750 Masking (CBCM). In the absence of test vectors, the best I have been able
17751 to do is check that the decrypt undoes the encrypt, so far. Send me test
17752 vectors if you have them.
17756 * Correct calculation of key length for export ciphers (too much space was
17757 allocated for null ciphers). This has not been tested!
17761 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17762 message is now correct (it understands "crypto" and "ssl" on its
17763 command line). There is also now an "update" option. This will update
17764 the util/ssleay.num and util/libeay.num files with any new functions.
17766 perl util/mkdef.pl crypto ssl update
17767 it will update them.
17771 * Overhauled the Perl interface:
17772 - ported BN stuff to OpenSSL's different BN library
17773 - made the perl/ source tree CVS-aware
17774 - renamed the package from SSLeay to OpenSSL (the files still contain
17775 their history because I've copied them in the repository)
17776 - removed obsolete files (the test scripts will be replaced
17777 by better Test::Harness variants in the future)
17779 *Ralf S. Engelschall*
17781 * First cut for a very conservative source tree cleanup:
17782 1. merge various obsolete readme texts into doc/ssleay.txt
17783 where we collect the old documents and readme texts.
17784 2. remove the first part of files where I'm already sure that we no
17785 longer need them because of three reasons: either they are just temporary
17786 files which were left by Eric or they are preserved original files where
17787 I've verified that the diff is also available in the CVS via "cvs diff
17788 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17789 the crypto/md/ stuff).
17791 *Ralf S. Engelschall*
17793 * More extension code. Incomplete support for subject and issuer alt
17794 name, issuer and authority key id. Change the i2v function parameters
17795 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17796 what that's for :-) Fix to ASN1 macro which messed up
17797 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17801 * Preliminary support for ENUMERATED type. This is largely copied from the
17806 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17808 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17810 * Make sure `make rehash` target really finds the `openssl` program.
17812 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17814 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17815 like to hear about it if this slows down other processors.
17819 * Add CygWin32 platform information to Configure script.
17821 *Alan Batie <batie@aahz.jf.intel.com>*
17823 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17825 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17827 * New program nseq to manipulate netscape certificate sequences
17831 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17836 * Fixes to BN code. Previously the default was to define BN_RECURSION
17837 but the BN code had some problems that would cause failures when
17838 doing certificate verification and some other functions.
17840 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17842 * Add ASN1 and PEM code to support netscape certificate sequences.
17846 * Add ASN1 and PEM code to support netscape certificate sequences.
17850 * Add several PKIX and private extended key usage OIDs.
17854 * Modify the 'ca' program to handle the new extension code. Modify
17855 openssl.cnf for new extension format, add comments.
17859 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17860 and add a sample to openssl.cnf so req -x509 now adds appropriate
17865 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17866 error code, add initial support to X509_print() and x509 application.
17870 * Takes a deep breath and start adding X509 V3 extension support code. Add
17871 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17872 stuff is currently isolated and isn't even compiled yet.
17876 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17877 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17878 Removed the versions check from X509 routines when loading extensions:
17879 this allows certain broken certificates that don't set the version
17880 properly to be processed.
17884 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17885 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17886 can still be regenerated with "make depend".
17890 * Spelling mistake in C version of CAST-128.
17892 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17894 * Changes to the error generation code. The perl script err-code.pl
17895 now reads in the old error codes and retains the old numbers, only
17896 adding new ones if necessary. It also only changes the .err files if new
17897 codes are added. The makefiles have been modified to only insert errors
17898 when needed (to avoid needlessly modifying header files). This is done
17899 by only inserting errors if the .err file is newer than the auto generated
17900 C file. To rebuild all the error codes from scratch (the old behaviour)
17901 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17902 or delete all the .err files.
17906 * CAST-128 was incorrectly implemented for short keys. The C version has
17907 been fixed, but is untested. The assembler versions are also fixed, but
17908 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17909 to regenerate it if needed.
17910 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17911 Hagino <itojun@kame.net>*
17913 * File was opened incorrectly in randfile.c.
17915 *Ulf Möller <ulf@fitug.de>*
17917 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17918 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17919 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17920 al: it's just almost always a UTCTime. Note this patch adds new error
17921 codes so do a "make errors" if there are problems.
17925 * Correct Linux 1 recognition in config.
17927 *Ulf Möller <ulf@fitug.de>*
17929 * Remove pointless MD5 hash when using DSA keys in ca.
17931 *Anonymous <nobody@replay.com>*
17933 * Generate an error if given an empty string as a cert directory. Also
17934 generate an error if handed NULL (previously returned 0 to indicate an
17935 error, but didn't set one).
17937 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17939 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17943 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17944 parameters. This was causing a warning which killed off the Win32 compile.
17948 * Remove C++ style comments from crypto/bn/bn_local.h.
17950 *Neil Costigan <neil.costigan@celocom.com>*
17952 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17953 based on a text string, looking up short and long names and finally
17954 "dot" format. The "dot" format stuff didn't work. Added new function
17955 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17956 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17957 OID is not part of the table.
17961 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17962 X509_LOOKUP_by_alias().
17966 * Sort openssl functions by name.
17970 * Get the gendsa program working (hopefully) and add it to app list. Remove
17971 encryption from sample DSA keys (in case anyone is interested the password
17976 * Make *all* `*_free` functions accept a NULL pointer.
17978 *Frans Heymans <fheymans@isaserver.be>*
17980 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17983 *Anonymous <nobody@replay.com>*
17985 * s_server should send the CAfile as acceptable CAs, not its own cert.
17987 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17989 * Don't blow it for numeric -newkey arguments to apps/req.
17991 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17993 * Temp key "for export" tests were wrong in s3_srvr.c.
17995 *Anonymous <nobody@replay.com>*
17997 * Add prototype for temp key callback functions
17998 SSL_CTX_set_tmp_{rsa,dh}_callback().
18002 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18003 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18007 * X509_name_add_entry() freed the wrong thing after an error.
18009 *Arne Ansper <arne@ats.cyber.ee>*
18011 * rsa_eay.c would attempt to free a NULL context.
18013 *Arne Ansper <arne@ats.cyber.ee>*
18015 * BIO_s_socket() had a broken should_retry() on Windoze.
18017 *Arne Ansper <arne@ats.cyber.ee>*
18019 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18021 *Arne Ansper <arne@ats.cyber.ee>*
18023 * Make sure the already existing X509_STORE->depth variable is initialized
18024 in X509_STORE_new(), but document the fact that this variable is still
18025 unused in the certificate verification process.
18027 *Ralf S. Engelschall*
18029 * Fix the various library and apps files to free up pkeys obtained from
18030 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18034 * Fix reference counting in X509_PUBKEY_get(). This makes
18035 demos/maurice/example2.c work, amongst others, probably.
18037 *Steve Henson and Ben Laurie*
18039 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18040 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18041 are no longer created. This way we have a single and consistent command
18042 line interface `openssl <command>`, similar to `cvs <command>`.
18044 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18046 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18047 BIT STRING wrapper always have zero unused bits.
18051 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18055 * Make the top-level INSTALL documentation easier to understand.
18059 * Makefiles updated to exit if an error occurs in a sub-directory
18060 make (including if user presses ^C) [Paul Sutton]
18062 * Make Montgomery context stuff explicit in RSA data structure.
18066 * Fix build order of pem and err to allow for generated pem.h.
18070 * Fix renumbering bug in X509_NAME_delete_entry().
18074 * Enhanced the err-ins.pl script so it makes the error library number
18075 global and can add a library name. This is needed for external ASN1 and
18076 other error libraries.
18080 * Fixed sk_insert which never worked properly.
18084 * Fix ASN1 macros so they can handle indefinite length constructed
18085 EXPLICIT tags. Some non standard certificates use these: they can now
18090 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18091 into a single doc/ssleay.txt bundle. This way the information is still
18092 preserved but no longer messes up this directory. Now it's new room for
18093 the new set of documentation files.
18095 *Ralf S. Engelschall*
18097 * SETs were incorrectly DER encoded. This was a major pain, because they
18098 shared code with SEQUENCEs, which aren't coded the same. This means that
18099 almost everything to do with SETs or SEQUENCEs has either changed name or
18100 number of arguments.
18102 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18104 * Fix test data to work with the above.
18108 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18109 was already fixed by Eric for 0.9.1 it seems.
18111 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18113 * Autodetect FreeBSD3.
18117 * Fix various bugs in Configure. This affects the following platforms:
18121 unixware-2.0-pentium
18126 * Eliminate generated files from CVS. Reorder tests to regenerate files
18127 before they are needed.
18131 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18135 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18137 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18138 changed SSLeay to OpenSSL in version strings.
18140 *Ralf S. Engelschall*
18142 * Some fixups to the top-level documents.
18146 * Fixed the nasty bug where rsaref.h was not found under compile-time
18147 because the symlink to include/ was missing.
18149 *Ralf S. Engelschall*
18151 * Incorporated the popular no-RSA/DSA-only patches
18152 which allow to compile a RSA-free SSLeay.
18154 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18156 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18157 when "ssleay" is still not found.
18159 *Ralf S. Engelschall*
18161 * Added more platforms to Configure: Cray T3E, HPUX 11,
18163 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18165 * Updated the README file.
18167 *Ralf S. Engelschall*
18169 * Added various .cvsignore files in the CVS repository subdirs
18170 to make a "cvs update" really silent.
18172 *Ralf S. Engelschall*
18174 * Recompiled the error-definition header files and added
18175 missing symbols to the Win32 linker tables.
18177 *Ralf S. Engelschall*
18179 * Cleaned up the top-level documents;
18180 o new files: CHANGES and LICENSE
18181 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18182 o merged COPYRIGHT into LICENSE
18183 o removed obsolete TODO file
18184 o renamed MICROSOFT to INSTALL.W32
18186 *Ralf S. Engelschall*
18188 * Removed dummy files from the 0.9.1b source tree:
18189 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18190 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18191 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18192 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18193 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18195 *Ralf S. Engelschall*
18197 * Added various platform portability fixes.
18201 * The Genesis of the OpenSSL rpject:
18202 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18203 Young and Tim J. Hudson created while they were working for C2Net until
18206 *The OpenSSL Project*
18208 ### Changes between 0.9.0b and 0.9.1b [not released]
18210 * Updated a few CA certificates under certs/
18214 * Changed some BIGNUM api stuff.
18218 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18219 DGUX x86, Linux Alpha, etc.
18223 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18224 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18229 * Add -strparse option to asn1pars program which parses nested
18232 *Dr Stephen Henson <shenson@bigfoot.com>*
18234 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18238 * DSA fix for "ca" program.
18242 * Added "-genkey" option to "dsaparam" program.
18246 * Added RIPE MD160 (rmd160) message digest.
18250 * Added -a (all) option to "ssleay version" command.
18254 * Added PLATFORM define which is the id given to Configure.
18258 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18262 * Extended the ASN.1 parser routines.
18266 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18270 * Added a BN_CTX to the BN library.
18274 * Fixed the weak key values in DES library
18278 * Changed API in EVP library for cipher aliases.
18282 * Added support for RC2/64bit cipher.
18286 * Converted the lhash library to the crypto/mem.c functions.
18290 * Added more recognized ASN.1 object ids.
18294 * Added more RSA padding checks for SSL/TLS.
18298 * Added BIO proxy/filter functionality.
18302 * Added extra_certs to SSL_CTX which can be used
18303 send extra CA certificates to the client in the CA cert chain sending
18304 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18308 * Now Fortezza is denied in the authentication phase because
18309 this is key exchange mechanism is not supported by SSLeay at all.
18313 * Additional PKCS1 checks.
18317 * Support the string "TLSv1" for all TLS v1 ciphers.
18321 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18322 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18326 * Fixed a few memory leaks.
18330 * Fixed various code and comment typos.
18334 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18335 bytes sent in the client random.
18337 *Edward Bishop <ebishop@spyglass.com>*
18341 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18342 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18343 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18344 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18345 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18346 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18347 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18348 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18349 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18350 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18351 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18352 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18353 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18354 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18355 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18356 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18357 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18358 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18359 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18360 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18361 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18362 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18363 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18364 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18365 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18366 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18367 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18368 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18369 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18370 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18371 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18372 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18373 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18374 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18375 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18376 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18377 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18378 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18379 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18380 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18381 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18382 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18383 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18384 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18385 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18386 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18387 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18388 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18389 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18390 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18391 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18392 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18393 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18394 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18395 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18396 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18397 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18398 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18399 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18400 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18401 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18402 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18403 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18404 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18405 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18406 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18407 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18408 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18409 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18410 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18411 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18412 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18413 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18414 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18415 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18416 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18417 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18418 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18419 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18420 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18421 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18422 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18423 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18424 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18425 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18426 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18427 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18428 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18429 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18430 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18431 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18432 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18433 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18434 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18435 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18436 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18437 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18438 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18439 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18440 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18441 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18442 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18443 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18444 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18445 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18446 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18447 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18448 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18449 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18450 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18451 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18452 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18453 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18454 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18455 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18456 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18457 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18458 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18459 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18460 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18461 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18462 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18463 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18464 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18465 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18466 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18467 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18468 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18469 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18470 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18471 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18472 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18473 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18474 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18475 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18476 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18477 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18478 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18479 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18480 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18481 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18482 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18483 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18484 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18485 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18486 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18487 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18488 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18489 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18490 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18491 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18492 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18493 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18494 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18495 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18496 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18497 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18498 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18499 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18500 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655