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 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
27 more key types including RSA, DSA, ED25519, X25519, ED448 and X448.
28 Previously (in 1.1.1) they would return -2. For key types that do not have
29 parameters then EVP_PKEY_param_check() will always return 1.
31 * The output from numerous "printing" functions such as X509_signature_print(),
32 X509_print_ex(), X509_CRL_print_ex(), and other similar functions has been
33 amended such that there may be cosmetic differences between the output
34 observed in 1.1.1 and 3.0. This also applies to the "-text" output from the
35 x509 and crl applications.
39 * Windows thread synchronization uses read/write primitives (SRWLock) when
40 supported by the OS, otherwise CriticalSection continues to be used.
44 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
45 work on read only BIO source/sinks that do not support these functions.
46 This allows piping or redirection of a file BIO using stdin to be buffered
47 into memory. This is used internally in OSSL_DECODER_from_bio().
51 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
52 this function would return one of the values OSSL_STORE_INFO_NAME,
53 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
54 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
55 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
56 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
57 using this function should be amended to handle the changed return value.
61 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
62 for the TSP implementation.
63 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
64 contains more than one certificate identifier: This means that all
65 certificates referenced there MUST be part of the validation chain.
69 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
70 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
71 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
72 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
73 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
74 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
75 using the EVP APIs to access these ciphers should instead use more modern
76 ciphers. If that is not possible then these applications should ensure that
77 the legacy provider has been loaded. This can be achieved either
78 programmatically or via configuration. See the provider(7) man page for
83 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
84 RIPEMD-160 have been moved to the legacy provider. Applications using the
85 EVP APIs to access these digests should instead use more modern digests. If
86 that is not possible then these applications should ensure that the legacy
87 provider has been loaded. This can be achieved either programmatically or via
88 configuration. See the provider(7) man page for further details.
92 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
97 * The deprecated functions EVP_PKEY_get0_RSA(),
98 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
99 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
100 well as the similarly named "get1" functions behave slightly differently in
101 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
102 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
103 provider. Calling these functions will only return a handle on the internal
104 key where the EVP_PKEY was constructed using this key in the first place, for
105 example using a function or macro such as EVP_PKEY_assign_RSA(),
106 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
107 then these functions now return a cached copy of the key. Changes to
108 the internal provider key that take place after the first time the cached key
109 is accessed will not be reflected back in the cached copy. Similarly any
110 changes made to the cached copy by application code will not be reflected
111 back in the internal provider key.
113 For the above reasons the keys returned from these functions should typically
114 be treated as read-only. To emphasise this the value returned from
115 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
116 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
117 break some existing code. Applications broken by this change should be
118 modified. The preferred solution is to refactor the code to avoid the use of
119 these deprecated functions. Failing this the code should be modified to use a
120 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
121 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
122 non-const pointer to enable them to be "freed". However they should also be
123 treated as read-only.
127 * A number of functions handling low level keys or engines were deprecated
128 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
129 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
130 EVP_PKEY_get0_siphash(). Applications using engines should instead use
131 providers. Applications getting or setting low level keys in an EVP_PKEY
132 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
133 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
137 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
138 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
139 and require returning octet ptr parameters from providers that
140 would like to support them which complicates provider implementations.
144 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
145 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
146 RAND_set_rand_method(). Provider based random number generators should
147 be used instead via EVP_RAND(3).
151 * The SRP APIs have been deprecated. The old APIs do not work via providers,
152 and there is no EVP interface to them. Unfortunately there is no replacement
153 for these APIs at this time.
157 * Add a compile time option to prevent the caching of provider fetched
158 algorithms. This is enabled by including the no-cached-fetch option
159 at configuration time.
163 * The default algorithms for pkcs12 creation with the PKCS12_create() function
164 were changed to more modern PBKDF2 and AES based algorithms. The default
165 MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal
166 with the password-based encryption iteration count. The default digest
167 algorithm for the MAC computation was changed to SHA-256. The pkcs12
168 application now supports -legacy option that restores the previous
169 default algorithms to support interoperability with legacy systems.
171 *Tomáš Mráz and Sahana Prasad*
173 * The openssl speed command does not use low-level API calls anymore. This
174 implies some of the performance numbers might not be fully comparable
175 with the previous releases due to higher overhead. This applies
176 particularly to measuring performance on smaller data chunks.
180 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
181 Typically if OpenSSL has no EC or DH algorithms then it cannot support
182 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
183 through providers. Therefore third party providers may supply group
184 implementations even where there are no built-in ones. Attempting to create
185 TLS connections in such a build without also disabling TLSv1.3 at run time or
186 using third party provider groups may result in handshake failures. TLSv1.3
187 can be disabled at compile time using the "no-tls1_3" Configure option.
191 * The undocumented function X509_certificate_type() has been deprecated;
192 applications can use X509_get0_pubkey() and X509_get0_signature() to
193 get the same information.
197 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
198 functions. They are identical to BN_rand() and BN_rand_range()
203 * Removed RSA padding mode for SSLv23 (which was only used for
204 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
205 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
210 * Deprecated the obsolete X9.31 RSA key generation related functions
211 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
212 BN_X931_generate_prime_ex().
216 * The default key generation method for the regular 2-prime RSA keys was
217 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
218 Conditions Based on Auxiliary Probable Primes). This method is slower
219 than the original method.
223 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
224 They are replaced with the BN_check_prime() function that avoids possible
225 misuse and always uses at least 64 rounds of the Miller-Rabin
226 primality test. At least 64 rounds of the Miller-Rabin test are now also
227 used for all prime generation, including RSA key generation.
228 This increases key generation time, especially for larger keys.
232 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
233 as they are not useful with non-deprecated functions.
237 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
238 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
239 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
240 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
241 were used to collect all necessary data to form a HTTP request, and to
242 perform the HTTP transfer with that request. With OpenSSL 3.0, the
243 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
244 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
245 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
246 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
247 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
248 OSSL_HTTP_REQ_CTX_set_max_response_length().
250 *Rich Salz and Richard Levitte*
252 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
256 * Validation of SM2 keys has been separated from the validation of regular EC
257 keys, allowing to improve the SM2 validation process to reject loaded private
258 keys that are not conforming to the SM2 ISO standard.
259 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
264 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
265 switches: a validation failure triggers an early exit, returning a failure
266 exit status to the parent process.
270 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
271 to ignore unknown ciphers.
275 * The `-cipher-commands` and `-digest-commands` options
276 of the command line utility `list` have been deprecated.
277 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
281 * All of the low level EC_KEY functions have been deprecated including:
283 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
284 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
285 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
286 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
287 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
288 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
289 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
290 EC_KEY_METHOD_get_verify,
291 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
292 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
293 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
294 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
295 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
296 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
297 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
298 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
299 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
300 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
301 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
302 Applications that need to implement an EC_KEY_METHOD need to consider
303 implementation of the functionality in a special provider.
304 For replacement of the functions manipulating the EC_KEY objects
305 see the EVP_PKEY-EC(7) manual page.
307 Additionally functions that read and write EC_KEY objects such as
308 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
309 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
310 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
311 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
312 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
313 have also been deprecated. Applications should instead use the
314 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
316 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
317 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
318 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
319 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
320 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
322 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
324 * Deprecated all the libcrypto and libssl error string loading
325 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
326 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
327 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
328 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
329 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
330 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
331 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
332 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
333 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
334 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
336 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
337 now loads error strings automatically.
341 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
342 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
343 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
344 are to be used by servers requiring ephemeral DH keys. Instead applications
345 should consider using the built-in DH parameters that are available by
346 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
347 necessary then applications can use the alternative functions
348 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
349 replacement for the "callback" functions. The callback was originally useful
350 in order to have different parameters for export and non-export ciphersuites.
351 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
352 functions should be replaced by one of the other methods described above.
356 * The `-crypt` option to the `passwd` command line tool has been removed.
360 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
365 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
367 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
368 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
369 The inverse ciphers use AES decryption for wrapping, and
370 AES encryption for unwrapping.
374 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
375 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
376 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
377 OpenSSL 3.0 these are replaced by the more generic functions
378 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
379 The old versions have been converted to deprecated macros that just call the
384 * The security callback, which can be customised by application code, supports
385 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
386 in the "other" parameter. In most places this is what is passed. All these
387 places occur server side. However there was one client side call of this
388 security operation and it passed a DH object instead. This is incorrect
389 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
390 of the other locations. Therefore this client side call has been changed to
391 pass an EVP_PKEY instead.
395 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
396 interface. Their functionality remains unchanged.
400 * Added new option for 'openssl list', '-providers', which will display the
401 list of loaded providers, their names, version and status. It optionally
402 displays their gettable parameters.
406 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
407 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
408 type is internally recognised so the workaround is no longer needed.
410 Functionality is still retained as it is, but will only work with
411 EVP_PKEYs with a legacy internal key.
415 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
416 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
420 * Changed all "STACK" functions to be macros instead of inline functions. Macro
421 parameters are still checked for type safety at compile time via helper
426 * Remove the RAND_DRBG API
428 The RAND_DRBG API did not fit well into the new provider concept as
429 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
430 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
431 and some of its API calls are rather low-level. This holds in particular
432 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
434 Adding a compatibility layer to continue supporting the RAND_DRBG API as
435 a legacy API for a regular deprecation period turned out to come at the
436 price of complicating the new provider API unnecessarily. Since the
437 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
440 *Paul Dale and Matthias St. Pierre*
442 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
443 as well as actual hostnames.
447 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
448 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
449 conversely, silently ignore DTLS protocol version bounds when configuring
450 TLS-based contexts. The commands can be repeated to set bounds of both
451 types. The same applies with the corresponding "min_protocol" and
452 "max_protocol" command-line switches, in case some application uses both TLS
455 SSL_CTX instances that are created for a fixed protocol version (e.g.
456 `TLSv1_server_method()`) also silently ignore version bounds. Previously
457 attempts to apply bounds to these protocol versions would result in an
458 error. Now only the "version-flexible" SSL_CTX instances are subject to
459 limits in configuration files in command-line options.
463 * Deprecated the `ENGINE` API. Engines should be replaced with providers
468 * Reworked the recorded ERR codes to make better space for system errors.
469 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
470 given code is a system error (true) or an OpenSSL error (false).
474 * Reworked the test perl framework to better allow parallel testing.
476 *Nicola Tuveri and David von Oheimb*
478 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
479 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
483 * 'Configure' has been changed to figure out the configuration target if
484 none is given on the command line. Consequently, the 'config' script is
485 now only a mere wrapper. All documentation is changed to only mention
488 *Rich Salz and Richard Levitte*
490 * Added a library context `OSSL_LIB_CTX` that applications as well as
491 other libraries can use to form a separate context within which
492 libcrypto operations are performed.
494 There are two ways this can be used:
496 - Directly, by passing a library context to functions that take
497 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
499 - Indirectly, by creating a new library context and then assigning
500 it as the new default, with `OSSL_LIB_CTX_set0_default`.
502 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
503 apart from the functions directly related to `OSSL_LIB_CTX`, accept
504 NULL to indicate that the default library context should be used.
506 Library code that changes the default library context using
507 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
508 second call before returning to the caller.
510 _(Note: the library context was initially called `OPENSSL_CTX` and
511 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
515 * Handshake now fails if Extended Master Secret extension is dropped
520 * Dropped interactive mode from the `openssl` program. From now on,
521 running it without arguments is equivalent to `openssl help`.
525 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
526 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
527 While the old function names have been retained for backward compatibility
528 they should not be used in new developments
529 because their return values are confusing: Unlike other `_cmp()` functions
530 they do not return 0 in case their arguments are equal.
534 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
535 `EC_GROUP_get_field_type()`.
539 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
540 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
541 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
542 Applications should rely on the library automatically assigning a suitable
543 EC_METHOD internally upon EC_GROUP construction.
547 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
548 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
549 assigned internally without application intervention.
550 Users of EC_GROUP_new() should switch to a different suitable constructor.
554 * Add CAdES-BES signature verification support, mostly derived
555 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
557 *Filipe Raimundo da Silva*
559 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
563 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
564 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose
565 is to support encryption and decryption of a digital envelope that is both
566 authenticated and encrypted using AES GCM mode.
570 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
571 functions are not widely used and now OpenSSL automatically perform this
572 conversion when needed.
576 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
577 EC_KEY_precompute_mult(). These functions are not widely used and
578 applications should instead switch to named curves which OpenSSL has
579 hardcoded lookup tables for.
583 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
584 should instead use the L<EC_POINT_mul(3)> function.
588 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
589 that are not applicable to the new provider model. Applications should
590 instead use EVP_default_properties_is_fips_enabled() and
591 EVP_default_properties_enable_fips().
595 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
596 is set, an unexpected EOF is ignored, it pretends a close notify was received
597 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
601 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
602 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
603 used and applications should instead use the
604 L<EC_POINT_set_affine_coordinates(3)> and
605 L<EC_POINT_get_affine_coordinates(3)> functions.
609 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
610 arrays to be more easily constructed via a series of utility functions.
611 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
612 the various push functions and finally convert to a passable OSSL_PARAM
613 array using OSSL_PARAM_BLD_to_param().
617 * The security strength of SHA1 and MD5 based signatures in TLS has been
618 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
619 working at the default security level of 1 and instead requires security
620 level 0. The security level can be changed either using the cipher string
621 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means
622 that where the signature algorithms extension is missing from a ClientHello
623 then the handshake will fail in TLS 1.2 at security level 1. This is because,
624 although this extension is optional, failing to provide one means that
625 OpenSSL will fallback to a default set of signature algorithms. This default
626 set requires the availability of SHA1.
630 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
631 contain a provider side internal key.
635 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
636 They are old functions that we don't use, and that you could disable with
637 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
641 * Project text documents not yet having a proper file name extension
642 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
643 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
647 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
648 have been converted to Markdown with the goal to produce documents
649 which not only look pretty when viewed online in the browser, but
650 remain well readable inside a plain text editor.
652 To achieve this goal, a 'minimalistic' Markdown style has been applied
653 which avoids formatting elements that interfere too much with the
654 reading flow in the text file. For example, it
656 * avoids [ATX headings][] and uses [setext headings][] instead
657 (which works for `<h1>` and `<h2>` headings only).
658 * avoids [inline links][] and uses [reference links][] instead.
659 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
661 [ATX headings]: https://github.github.com/gfm/#atx-headings
662 [setext headings]: https://github.github.com/gfm/#setext-headings
663 [inline links]: https://github.github.com/gfm/#inline-link
664 [reference links]: https://github.github.com/gfm/#reference-link
665 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
666 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
668 *Matthias St. Pierre*
670 * The test suite is changed to preserve results of each test recipe.
671 A new directory test-runs/ with subdirectories named like the
672 test recipes are created in the build tree for this purpose.
676 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
677 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
678 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
680 *David von Oheimb, Martin Peylo*
682 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
683 The legacy OCSP-focused and only partly documented API is retained for
684 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
688 * Added `util/check-format.pl`, a tool for checking adherence to the
689 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
690 The checks performed are incomplete and yield some false positives.
691 Still the tool should be useful for detecting most typical glitches.
695 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
696 If domain name resolution yields multiple IP addresses all of them are tried
697 after `connect()` failures.
701 * All of the low level RSA functions have been deprecated including:
703 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
704 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
705 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
706 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
707 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
708 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
709 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
710 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
711 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
712 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
713 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
714 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
715 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
716 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
717 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
718 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
719 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
720 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
721 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
722 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
723 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
724 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
725 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
726 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
727 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
728 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
729 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
730 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
732 Use of these low level functions has been informally discouraged for a long
733 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
734 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
735 L<EVP_PKEY_decrypt(3)>.
739 * X509 certificates signed using SHA1 are no longer allowed at security
741 In TLS/SSL the default security level is 1. It can be set either
742 using the cipher string with `@SECLEVEL`, or calling
743 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
744 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
746 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
747 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
748 options of the commands.
752 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
753 modified to use PKEY APIs. These commands are now in maintenance mode
754 and no new features will be added to them.
758 * The command line utility rsautl has been deprecated.
759 Instead use the pkeyutl program.
763 * The command line utilities genrsa and rsa have been modified to use PKEY
764 APIs. They now write PKCS#8 keys by default. These commands are now in
765 maintenance mode and no new features will be added to them.
769 * All of the low level DH functions have been deprecated including:
771 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
772 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
773 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
774 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
775 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
776 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
777 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
778 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
779 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
780 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
781 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
782 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
783 DH_meth_get_generate_params and DH_meth_set_generate_params.
785 Use of these low level functions has been informally discouraged for a long
786 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
787 and L<EVP_PKEY_derive(3)>.
789 Additionally functions that read and write DH objects such as d2i_DHparams,
790 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
791 functions have also been deprecated. Applications should instead use the
792 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
794 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
795 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
796 `EVP_PKEY_set1_DH()` are also deprecated.
797 Applications should instead either read or write an
798 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
799 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
801 *Paul Dale and Matt Caswell*
803 * All of the low level DSA functions have been deprecated including:
805 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
806 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
807 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
808 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
809 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
810 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
811 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
812 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
813 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
814 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
815 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
816 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
817 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
818 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
819 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
820 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
821 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
823 Use of these low level functions has been informally discouraged for a long
824 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
825 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
827 Finaly functions that assign or obtain DSA objects from an EVP_PKEY such as
828 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
829 `EVP_PKEY_set1_DSA()` are also deprecated.
830 Applications should instead either read or write an
831 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
832 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
836 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
837 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
838 This means that applications don't have to look at the curve NID and
839 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
840 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
841 a no-op when the EVP_PKEY is already of the given type.
843 Parameter and key generation is also reworked to make it possible
844 to generate EVP_PKEY_SM2 parameters and keys without having to go
845 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
846 However, code that does the latter will still work as before.
850 * Deprecated low level ECDH and ECDSA functions. These include:
852 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
853 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
856 Use of these low level functions has been informally discouraged for a long
857 time. Instead applications should use the EVP_PKEY_derive(3),
858 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
862 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
863 and EVP_PKEY_decrypt() instead.
864 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
865 and EVP_PKEY_encrypt() instead.
869 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
870 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
871 a new formulation to include all the things it can be used for,
872 as well as words of caution.
876 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
877 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
881 * All of the low level HMAC functions have been deprecated including:
883 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
884 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
887 Use of these low level functions has been informally discouraged for a long
888 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
889 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
890 and L<EVP_MAC_final(3)>.
894 * Over two thousand fixes were made to the documentation, including:
895 - Common options (such as -rand/-writerand, TLS version control, etc)
896 were refactored and point to newly-enhanced descriptions in openssl.pod.
897 - Added style conformance for all options (with help from Richard Levitte),
898 documented all reported missing options, added a CI build to check
899 that all options are documented and that no unimplemented options
901 - Documented some internals, such as all use of environment variables.
902 - Addressed all internal broken L<> references.
906 * All of the low level CMAC functions have been deprecated including:
908 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
909 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
911 Use of these low level functions has been informally discouraged for a long
912 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
913 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
914 and L<EVP_MAC_final(3)>.
918 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
919 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
922 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
923 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
924 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
925 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
926 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
927 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
928 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
929 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
930 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
931 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
933 Use of these low level functions has been informally discouraged
934 for a long time. Applications should use the EVP_DigestInit_ex(3),
935 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
939 * Corrected the documentation of the return values from the `EVP_DigestSign*`
940 set of functions. The documentation mentioned negative values for some
941 errors, but this was never the case, so the mention of negative values
944 Code that followed the documentation and thereby check with something
945 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
949 * All of the low level cipher functions have been deprecated including:
951 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
952 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
953 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
954 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
955 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
956 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
957 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
958 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
959 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
960 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
961 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
962 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
963 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
964 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
965 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
966 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
967 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
968 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
969 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
970 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
971 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
972 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
973 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
974 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
975 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
976 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
977 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
978 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
979 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
981 Use of these low level functions has been informally discouraged for
982 a long time. Applications should use the high level EVP APIs, e.g.
983 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
984 equivalently named decrypt functions instead.
986 *Matt Caswell and Paul Dale*
988 * Removed include/openssl/opensslconf.h.in and replaced it with
989 include/openssl/configuration.h.in, which differs in not including
990 <openssl/macros.h>. A short header include/openssl/opensslconf.h
991 was added to include both.
993 This allows internal hacks where one might need to modify the set
994 of configured macros, for example this if deprecated symbols are
995 still supposed to be available internally:
997 #include <openssl/configuration.h>
999 #undef OPENSSL_NO_DEPRECATED
1000 #define OPENSSL_SUPPRESS_DEPRECATED
1002 #include <openssl/macros.h>
1004 This should not be used by applications that use the exported
1005 symbols, as that will lead to linking errors.
1009 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1010 used in exponentiation with 512-bit moduli. No EC algorithms are
1011 affected. Analysis suggests that attacks against 2-prime RSA1024,
1012 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1013 difficult to perform and are not believed likely. Attacks against DH512
1014 are considered just feasible. However, for an attack the target would
1015 have to re-use the DH512 private key, which is not recommended anyway.
1016 Also applications directly using the low level API BN_mod_exp may be
1017 affected if they use BN_FLG_CONSTTIME.
1022 * Most memory-debug features have been deprecated, and the functionality
1023 replaced with no-ops.
1027 * Added documentation for the STACK API.
1031 * Introduced a new method type and API, OSSL_ENCODER, to
1032 represent generic encoders. An implementation is expected to
1033 be able to encode an object associated with a given name (such
1034 as an algorithm name for an asymmetric key) into forms given by
1035 implementation properties.
1037 Encoders are primarily used from inside libcrypto, through
1038 calls to functions like EVP_PKEY_print_private(),
1039 PEM_write_bio_PrivateKey() and similar.
1041 Encoders are specified in such a way that they can be made to
1042 directly handle the provider side portion of an object, if this
1043 provider side part comes from the same provider as the encoder
1044 itself, but can also be made to handle objects in parametrized
1045 form (as an OSSL_PARAM array of data). This allows a provider to
1046 offer generic encoders as a service for any other provider.
1050 * Added a .pragma directive to the syntax of configuration files, to
1051 allow varying behavior in a supported and predictable manner.
1052 Currently added pragma:
1056 This allows dollar signs to be a keyword character unless it's
1057 followed by a opening brace or parenthesis. This is useful for
1058 platforms where dollar signs are commonly used in names, such as
1059 volume names and system directory names on VMS.
1063 * Added functionality to create an EVP_PKEY from user data. This
1064 is effectively the same as creating a RSA, DH or DSA object and
1065 then assigning them to an EVP_PKEY, but directly using algorithm
1066 agnostic EVP functions. A benefit is that this should be future
1067 proof for public key algorithms to come.
1071 * Change the interpretation of the '--api' configuration option to
1072 mean that this is a desired API compatibility level with no
1073 further meaning. The previous interpretation, that this would
1074 also mean to remove all deprecated symbols up to and including
1075 the given version, no requires that 'no-deprecated' is also used
1076 in the configuration.
1078 When building applications, the desired API compatibility level
1079 can be set with the OPENSSL_API_COMPAT macro like before. For
1080 API compatibility version below 3.0, the old style numerical
1081 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1082 For version 3.0 and on, the value is expected to be the decimal
1083 value calculated from the major and minor version like this:
1085 MAJOR * 10000 + MINOR * 100
1089 -DOPENSSL_API_COMPAT=30000 For 3.0
1090 -DOPENSSL_API_COMPAT=30200 For 3.2
1092 To hide declarations that are deprecated up to and including the
1093 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1094 given when building the application as well.
1098 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1099 access to certificate and CRL stores via URIs and OSSL_STORE
1102 This adds the following functions:
1104 - X509_LOOKUP_store()
1105 - X509_STORE_load_file()
1106 - X509_STORE_load_path()
1107 - X509_STORE_load_store()
1108 - SSL_add_store_cert_subjects_to_stack()
1109 - SSL_CTX_set_default_verify_store()
1110 - SSL_CTX_load_verify_file()
1111 - SSL_CTX_load_verify_dir()
1112 - SSL_CTX_load_verify_store()
1116 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1117 The presence of this system service is determined at run-time.
1121 * Added functionality to create an EVP_PKEY context based on data
1122 for methods from providers. This takes an algorithm name and a
1123 property query string and simply stores them, with the intent
1124 that any operation that uses this context will use those strings
1125 to fetch the needed methods implicitly, thereby making the port
1126 of application written for pre-3.0 OpenSSL easier.
1130 * The undocumented function NCONF_WIN32() has been deprecated; for
1131 conversion details see the HISTORY section of doc/man5/config.pod
1135 * Introduced the new functions EVP_DigestSignInit_ex() and
1136 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1137 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1138 pages for further details.
1142 * Over two thousand fixes were made to the documentation, including:
1143 adding missing command flags, better style conformance, documentation
1146 *Rich Salz, Richard Levitte*
1148 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1149 X25519, X448, Ed25519 and Ed448.
1153 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1158 * Deprecated the public definition of `ERR_STATE` as well as the function
1159 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1164 * Added ERR functionality to give callers access to the stored function
1165 names that have replaced the older function code based functions.
1167 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1168 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1169 ERR_peek_error_all() and ERR_peek_last_error_all().
1171 These functions have become deprecated: ERR_get_error_line(),
1172 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1173 ERR_peek_last_error_line_data() and ERR_func_error_string().
1175 Users are recommended to use ERR_get_error_all(), or to pick information
1176 with ERR_peek functions and finish off with getting the error code by using
1181 * Extended testing to be verbose for failing tests only. The make variables
1182 VERBOSE_FAILURE or VF can be used to enable this:
1184 $ make VF=1 test # Unix
1185 $ mms /macro=(VF=1) test ! OpenVMS
1186 $ nmake VF=1 test # Windows
1190 * Added the `-copy_extensions` option to the `x509` command for use with
1191 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1192 all extensions in the request are copied to the certificate or vice versa.
1194 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1196 * Added the `-copy_extensions` option to the `req` command for use with
1197 `-x509`. When given with the `copy` or `copyall` argument,
1198 all extensions in the certification request are copied to the certificate.
1202 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1203 they generate are by default RFC 5280 compliant in the following sense:
1204 There is a subjectKeyIdentifier extension with a hash value of the public key
1205 and for not self-signed certs there is an authorityKeyIdentifier extension
1206 with a keyIdentifier field or issuer information identifying the signing key.
1207 This is done unless some configuration overrides the new default behavior,
1208 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1212 * Added several checks to `X509_verify_cert()` according to requirements in
1213 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1214 (which may be done by using the CLI option `-x509_strict`):
1215 * The basicConstraints of CA certificates must be marked critical.
1216 * CA certificates must explicitly include the keyUsage extension.
1217 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1218 * The issuer name of any certificate must not be empty.
1219 * The subject name of CA certs, certs with keyUsage crlSign,
1220 and certs without subjectAlternativeName must not be empty.
1221 * If a subjectAlternativeName extension is given it must not be empty.
1222 * The signatureAlgorithm field and the cert signature must be consistent.
1223 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1224 must not be marked critical.
1225 * The authorityKeyIdentifier must be given for X.509v3 certs
1226 unless they are self-signed.
1227 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1231 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1232 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1236 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1237 used even when parsing explicit parameters, when loading a encoded key
1238 or calling `EC_GROUP_new_from_ecpkparameters()`/
1239 `EC_GROUP_new_from_ecparameters()`.
1240 This prevents bypass of security hardening and performance gains,
1241 especially for curves with specialized EC_METHODs.
1242 By default, if a key encoded with explicit parameters is loaded and later
1243 encoded, the output is still encoded with explicit parameters, even if
1244 internally a "named" EC_GROUP is used for computation.
1248 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1249 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1250 NULL. After this change, only the cofactor parameter can be NULL. It also
1251 does some minimal sanity checks on the passed order.
1256 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1257 An attack is simple, if the first CMS_recipientInfo is valid but the
1258 second CMS_recipientInfo is chosen ciphertext. If the second
1259 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1260 encryption key will be replaced by garbage, and the message cannot be
1261 decoded, but if the RSA decryption fails, the correct encryption key is
1262 used and the recipient will not notice the attack.
1263 As a work around for this potential attack the length of the decrypted
1264 key must be equal to the cipher default key length, in case the
1265 certifiate is not given and all recipientInfo are tried out.
1266 The old behaviour can be re-enabled in the CMS code by setting the
1267 CMS_DEBUG_DECRYPT flag.
1271 * Early start up entropy quality from the DEVRANDOM seed source has been
1272 improved for older Linux systems. The RAND subsystem will wait for
1273 /dev/random to be producing output before seeding from /dev/urandom.
1274 The seeded state is stored for future library initialisations using
1275 a system global shared memory segment. The shared memory identifier
1276 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1277 the desired value. The default identifier is 114.
1281 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1282 when primes for RSA keys are computed.
1283 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1284 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1285 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1286 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1287 This avoids possible fingerprinting of newly generated RSA modules.
1291 * Correct the extended master secret constant on EBCDIC systems. Without this
1292 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1293 negotiate EMS will fail. Unfortunately this also means that TLS connections
1294 between EBCDIC systems with this fix, and EBCDIC systems without this
1295 fix will fail if they negotiate EMS.
1299 * Changed the library initialisation so that the config file is now loaded
1300 by default. This was already the case for libssl. It now occurs for both
1301 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1302 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1306 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1307 where the former acts as a replacement for `ERR_put_error()`, and the
1308 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1309 `ERR_raise_data()` adds more flexibility by taking a format string and
1310 an arbitrary number of arguments following it, to be processed with
1315 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1316 to check if a named provider is loaded and available. When called, it
1317 will also activate all fallback providers if such are still present.
1321 * Enforce a minimum DH modulus size of 512 bits.
1325 * Changed DH parameters to generate the order q subgroup instead of 2q.
1326 Previously generated DH parameters are still accepted by DH_check
1327 but DH_generate_key works around that by clearing bit 0 of the
1328 private key for those. This avoids leaking bit 0 of the private key.
1332 * Significantly reduce secure memory usage by the randomness pools.
1336 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1341 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1342 algorithms. An implementation of a key exchange algorithm can be obtained
1343 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1344 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1345 the older EVP_PKEY_derive_init() function. See the man pages for the new
1346 functions for further details.
1350 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1354 * Removed the function names from error messages and deprecated the
1357 * Removed NextStep support and the macro OPENSSL_UNISTD
1361 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1362 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1363 Also removed "export var as function" capability; we do not export
1364 variables, only functions.
1368 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1369 an error and 1 indicating success. In previous versions of OpenSSL this
1370 was a void type. If a key was set longer than the maximum possible this
1375 * Support SM2 signing and verification schemes with X509 certificate.
1379 * Use SHA256 as the default digest for TS query in the `ts` app.
1383 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1384 This checks that the salt length is at least 128 bits, the derived key
1385 length is at least 112 bits, and that the iteration count is at least 1000.
1386 For backwards compatibility these checks are disabled by default in the
1387 default provider, but are enabled by default in the fips provider.
1388 To enable or disable these checks use the control
1389 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1393 * Default cipher lists/suites are now available via a function, the
1394 #defines are deprecated.
1398 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1399 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1400 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1404 * Join the directories crypto/x509 and crypto/x509v3
1408 * Added command 'openssl kdf' that uses the EVP_KDF API.
1412 * Added command 'openssl mac' that uses the EVP_MAC API.
1416 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1417 as default directories. Also added the command 'openssl info'
1418 for scripting purposes.
1422 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1423 deprecated. These undocumented functions were never integrated into the EVP
1424 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1425 Bi-directional IGE mode. These modes were never formally standardised and
1426 usage of these functions is believed to be very small. In particular
1427 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1428 is ever used. The security implications are believed to be minimal, but
1429 this issue was never fixed for backwards compatibility reasons. New code
1430 should not use these modes.
1434 * Add prediction resistance to the DRBG reseeding process.
1438 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1439 mandated by IEEE Std 1619-2018.
1443 * Added newline escaping functionality to a filename when using openssl dgst.
1444 This output format is to replicate the output format found in the `*sum`
1445 checksum programs. This aims to preserve backward compatibility.
1447 *Matt Eaton, Richard Levitte, and Paul Dale*
1449 * Removed the heartbeat message in DTLS feature, as it has very
1450 little usage and doesn't seem to fulfill a valuable purpose.
1451 The configuration option is now deprecated.
1455 * Changed the output of 'openssl {digestname} < file' to display the
1456 digest name in its output.
1460 * Added a new generic trace API which provides support for enabling
1461 instrumentation through trace output. This feature is mainly intended
1462 as an aid for developers and is disabled by default. To utilize it,
1463 OpenSSL needs to be configured with the `enable-trace` option.
1465 If the tracing API is enabled, the application can activate trace output
1466 by registering BIOs as trace channels for a number of tracing and debugging
1469 The `openssl` program has been expanded to enable any of the types
1470 available via environment variables defined by the user, and serves as
1471 one possible example on how to use this functionality.
1473 *Richard Levitte & Matthias St. Pierre*
1475 * Added build tests for C++. These are generated files that only do one
1476 thing, to include one public OpenSSL head file each. This tests that
1477 the public header files can be usefully included in a C++ application.
1479 This test isn't enabled by default. It can be enabled with the option
1480 'enable-buildtest-c++'.
1484 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1488 * Add KMAC to EVP_MAC.
1492 * Added property based algorithm implementation selection framework to
1497 * Added SCA hardening for modular field inversion in EC_GROUP through
1498 a new dedicated field_inv() pointer in EC_METHOD.
1499 This also addresses a leakage affecting conversions from projective
1500 to affine coordinates.
1502 *Billy Bob Brumley, Nicola Tuveri*
1504 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1505 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1506 those algorithms that were already supported through the EVP_PKEY API
1507 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1508 and scrypt are now wrappers that call EVP_KDF.
1512 * Build devcrypto engine as a dynamic engine.
1514 *Eneas U de Queiroz*
1516 * Add keyed BLAKE2 to EVP_MAC.
1520 * Fix a bug in the computation of the endpoint-pair shared secret used
1521 by DTLS over SCTP. This breaks interoperability with older versions
1522 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1523 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1524 interoperability with such broken implementations. However, enabling
1525 this switch breaks interoperability with correct implementations.
1527 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1528 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1532 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1536 * Changed the license to the Apache License v2.0.
1540 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1542 - Major releases (indicated by incrementing the MAJOR release number)
1543 may introduce incompatible API/ABI changes.
1544 - Minor releases (indicated by incrementing the MINOR release number)
1545 may introduce new features but retain API/ABI compatibility.
1546 - Patch releases (indicated by incrementing the PATCH number)
1547 are intended for bug fixes and other improvements of existing
1548 features only (like improving performance or adding documentation)
1549 and retain API/ABI compatibility.
1553 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1557 * Remove the 'dist' target and add a tarball building script. The
1558 'dist' target has fallen out of use, and it shouldn't be
1559 necessary to configure just to create a source distribution.
1563 * Recreate the OS390-Unix config target. It no longer relies on a
1564 special script like it did for OpenSSL pre-1.1.0.
1568 * Instead of having the source directories listed in Configure, add
1569 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1574 * Add GMAC to EVP_MAC.
1578 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1582 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1583 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1584 to facilitate the continued use of MACs through raw private keys in
1585 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1589 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1590 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1594 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1595 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1596 are retained for backwards compatibility.
1600 * AES-XTS mode now enforces that its two keys are different to mitigate
1601 the attacked described in "Efficient Instantiations of Tweakable
1602 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1603 Details of this attack can be obtained from:
1604 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1608 * Rename the object files, i.e. give them other names than in previous
1609 versions. Their names now include the name of the final product, as
1610 well as its type mnemonic (bin, lib, shlib).
1614 * Added new option for 'openssl list', '-objects', which will display the
1615 list of built in objects, i.e. OIDs with names.
1619 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1620 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1625 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1626 improves application performance by removing data copies and providing
1627 applications with zero-copy system calls such as sendfile and splice.
1631 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1632 option is set, openssl cleanses (zeroize) plaintext bytes from
1633 internal buffers after delivering them to the application. Note,
1634 the application is still responsible for cleansing other copies
1635 (e.g.: data received by SSL_read(3)).
1639 * `PKCS12_parse` now maintains the order of the parsed certificates
1640 when outputting them via `*ca` (rather than reversing it).
1644 * Deprecated pthread fork support methods. These were unused so no
1645 replacement is required.
1647 - OPENSSL_fork_prepare()
1648 - OPENSSL_fork_parent()
1649 - OPENSSL_fork_child()
1656 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1658 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1660 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1661 create a unique hash value based on the issuer and serial number data
1662 contained within an X509 certificate. However it was failing to correctly
1663 handle any errors that may occur while parsing the issuer field (which might
1664 occur if the issuer field is maliciously constructed). This may subsequently
1665 result in a NULL pointer deref and a crash leading to a potential denial of
1671 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1672 padding mode to correctly check for rollback attacks. This is considered a
1673 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1678 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1679 functions. Previously they could overflow the output length argument in some
1680 cases where the input length is close to the maximum permissable length for
1681 an integer on the platform. In such cases the return value from the function
1682 call would be 1 (indicating success), but the output length value would be
1683 negative. This could cause applications to behave incorrectly or crash.
1688 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1689 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1690 could be exploited in a side channel attack to recover the password. Since
1691 the attack is local host only this is outside of the current OpenSSL
1692 threat model and therefore no CVE is assigned.
1694 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1699 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1701 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1702 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1703 If an attacker can control both items being compared then this could lead
1704 to a possible denial of service attack. OpenSSL itself uses the
1705 GENERAL_NAME_cmp function for two purposes:
1706 1) Comparing CRL distribution point names between an available CRL and a
1707 CRL distribution point embedded in an X509 certificate
1708 2) When verifying that a timestamp response token signer matches the
1709 timestamp authority name (exposed via the API functions
1710 TS_RESP_verify_response and TS_RESP_verify_token)
1715 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1717 * Certificates with explicit curve parameters are now disallowed in
1718 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1722 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1723 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1724 conversely, silently ignore DTLS protocol version bounds when configuring
1725 TLS-based contexts. The commands can be repeated to set bounds of both
1726 types. The same applies with the corresponding "min_protocol" and
1727 "max_protocol" command-line switches, in case some application uses both TLS
1730 SSL_CTX instances that are created for a fixed protocol version (e.g.
1731 TLSv1_server_method()) also silently ignore version bounds. Previously
1732 attempts to apply bounds to these protocol versions would result in an
1733 error. Now only the "version-flexible" SSL_CTX instances are subject to
1734 limits in configuration files in command-line options.
1738 * Handshake now fails if Extended Master Secret extension is dropped
1743 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1745 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1747 * Fixed segmentation fault in SSL_check_chain()
1748 Server or client applications that call the SSL_check_chain() function
1749 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1750 dereference as a result of incorrect handling of the
1751 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1752 or unrecognised signature algorithm is received from the peer. This could
1753 be exploited by a malicious peer in a Denial of Service attack.
1758 * Added AES consttime code for no-asm configurations
1759 an optional constant time support for AES was added
1760 when building openssl for no-asm.
1761 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1762 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1763 At this time this feature is by default disabled.
1764 It will be enabled by default in 3.0.
1768 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1770 * Revert the change of EOF detection while reading in libssl to avoid
1771 regressions in applications depending on the current way of reporting
1772 the EOF. As the existing method is not fully accurate the change to
1773 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1774 branch and will be present in the 3.0 release.
1778 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1779 when primes for RSA keys are computed.
1780 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1781 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1782 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1783 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1784 This avoids possible fingerprinting of newly generated RSA modules.
1788 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1790 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1791 while reading in libssl then we would report an error back to the
1792 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1793 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1794 therefore give a hint as to what went wrong.
1798 * Check that ed25519 and ed448 are allowed by the security level. Previously
1799 signature algorithms not using an MD were not being checked that they were
1800 allowed by the security level.
1804 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1805 was not quite right. The behaviour was not consistent between resumption
1806 and normal handshakes, and also not quite consistent with historical
1807 behaviour. The behaviour in various scenarios has been clarified and
1808 it has been updated to make it match historical behaviour as closely as
1813 * *[VMS only]* The header files that the VMS compilers include automatically,
1814 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1815 that the C++ compiler doesn't understand. This is a shortcoming in the
1816 compiler, but can be worked around with `__cplusplus` guards.
1818 C++ applications that use OpenSSL libraries must be compiled using the
1819 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1820 functions. Otherwise, only functions with symbols of less than 31
1821 characters can be used, as the linker will not be able to successfully
1822 resolve symbols with longer names.
1826 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1827 The presence of this system service is determined at run-time.
1831 * Added newline escaping functionality to a filename when using openssl dgst.
1832 This output format is to replicate the output format found in the `*sum`
1833 checksum programs. This aims to preserve backward compatibility.
1835 *Matt Eaton, Richard Levitte, and Paul Dale*
1837 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1842 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1844 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1845 number generator (RNG). This was intended to include protection in the
1846 event of a fork() system call in order to ensure that the parent and child
1847 processes did not share the same RNG state. However this protection was not
1848 being used in the default case.
1850 A partial mitigation for this issue is that the output from a high
1851 precision timer is mixed into the RNG state so the likelihood of a parent
1852 and child process sharing state is significantly reduced.
1854 If an application already calls OPENSSL_init_crypto() explicitly using
1855 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1858 *Matthias St. Pierre*
1860 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1861 used even when parsing explicit parameters, when loading a encoded key
1862 or calling `EC_GROUP_new_from_ecpkparameters()`/
1863 `EC_GROUP_new_from_ecparameters()`.
1864 This prevents bypass of security hardening and performance gains,
1865 especially for curves with specialized EC_METHODs.
1866 By default, if a key encoded with explicit parameters is loaded and later
1867 encoded, the output is still encoded with explicit parameters, even if
1868 internally a "named" EC_GROUP is used for computation.
1872 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1873 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1874 NULL. After this change, only the cofactor parameter can be NULL. It also
1875 does some minimal sanity checks on the passed order.
1880 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1881 An attack is simple, if the first CMS_recipientInfo is valid but the
1882 second CMS_recipientInfo is chosen ciphertext. If the second
1883 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1884 encryption key will be replaced by garbage, and the message cannot be
1885 decoded, but if the RSA decryption fails, the correct encryption key is
1886 used and the recipient will not notice the attack.
1887 As a work around for this potential attack the length of the decrypted
1888 key must be equal to the cipher default key length, in case the
1889 certifiate is not given and all recipientInfo are tried out.
1890 The old behaviour can be re-enabled in the CMS code by setting the
1891 CMS_DEBUG_DECRYPT flag.
1896 * Early start up entropy quality from the DEVRANDOM seed source has been
1897 improved for older Linux systems. The RAND subsystem will wait for
1898 /dev/random to be producing output before seeding from /dev/urandom.
1899 The seeded state is stored for future library initialisations using
1900 a system global shared memory segment. The shared memory identifier
1901 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1902 the desired value. The default identifier is 114.
1906 * Correct the extended master secret constant on EBCDIC systems. Without this
1907 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1908 negotiate EMS will fail. Unfortunately this also means that TLS connections
1909 between EBCDIC systems with this fix, and EBCDIC systems without this
1910 fix will fail if they negotiate EMS.
1914 * Use Windows installation paths in the mingw builds
1916 Mingw isn't a POSIX environment per se, which means that Windows
1917 paths should be used for installation.
1922 * Changed DH_check to accept parameters with order q and 2q subgroups.
1923 With order 2q subgroups the bit 0 of the private key is not secret
1924 but DH_generate_key works around that by clearing bit 0 of the
1925 private key for those. This avoids leaking bit 0 of the private key.
1929 * Significantly reduce secure memory usage by the randomness pools.
1933 * Revert the DEVRANDOM_WAIT feature for Linux systems
1935 The DEVRANDOM_WAIT feature added a select() call to wait for the
1936 /dev/random device to become readable before reading from the
1937 /dev/urandom device.
1939 It turned out that this change had negative side effects on
1940 performance which were not acceptable. After some discussion it
1941 was decided to revert this feature and leave it up to the OS
1942 resp. the platform maintainer to ensure a proper initialization
1943 during early boot time.
1945 *Matthias St. Pierre*
1947 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1949 * Add build tests for C++. These are generated files that only do one
1950 thing, to include one public OpenSSL head file each. This tests that
1951 the public header files can be usefully included in a C++ application.
1953 This test isn't enabled by default. It can be enabled with the option
1954 'enable-buildtest-c++'.
1958 * Enable SHA3 pre-hashing for ECDSA and DSA.
1962 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1963 This changes the size when using the `genpkey` command when no size is given.
1964 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1965 generation commands to use 2048 bits by default.
1969 * Reorganize the manual pages to consistently have RETURN VALUES,
1970 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1971 util/fix-doc-nits accordingly.
1973 *Paul Yang, Joshua Lock*
1975 * Add the missing accessor EVP_PKEY_get0_engine()
1979 * Have commands like `s_client` and `s_server` output the signature scheme
1980 along with other cipher suite parameters when debugging.
1984 * Make OPENSSL_config() error agnostic again.
1988 * Do the error handling in RSA decryption constant time.
1992 * Prevent over long nonces in ChaCha20-Poly1305.
1994 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1995 for every encryption operation. RFC 7539 specifies that the nonce value
1996 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1997 and front pads the nonce with 0 bytes if it is less than 12
1998 bytes. However it also incorrectly allows a nonce to be set of up to 16
1999 bytes. In this case only the last 12 bytes are significant and any
2000 additional leading bytes are ignored.
2002 It is a requirement of using this cipher that nonce values are
2003 unique. Messages encrypted using a reused nonce value are susceptible to
2004 serious confidentiality and integrity attacks. If an application changes
2005 the default nonce length to be longer than 12 bytes and then makes a
2006 change to the leading bytes of the nonce expecting the new value to be a
2007 new unique nonce then such an application could inadvertently encrypt
2008 messages with a reused nonce.
2010 Additionally the ignored bytes in a long nonce are not covered by the
2011 integrity guarantee of this cipher. Any application that relies on the
2012 integrity of these ignored leading bytes of a long nonce may be further
2013 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2014 is safe because no such use sets such a long nonce value. However user
2015 applications that use this cipher directly and set a non-default nonce
2016 length to be longer than 12 bytes may be vulnerable.
2018 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2024 * Add DEVRANDOM_WAIT feature for Linux systems
2026 On older Linux systems where the getrandom() system call is not available,
2027 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2028 Contrary to getrandom(), the /dev/urandom device will not block during
2029 early boot when the kernel CSPRNG has not been seeded yet.
2031 To mitigate this known weakness, use select() to wait for /dev/random to
2032 become readable before reading from /dev/urandom.
2034 * Ensure that SM2 only uses SM3 as digest algorithm
2038 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2040 * Change the info callback signals for the start and end of a post-handshake
2041 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2042 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2043 confused by this and assume that a TLSv1.2 renegotiation has started. This
2044 can break KeyUpdate handling. Instead we no longer signal the start and end
2045 of a post handshake message exchange (although the messages themselves are
2046 still signalled). This could break some applications that were expecting
2047 the old signals. However without this KeyUpdate is not usable for many
2052 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2054 * Timing vulnerability in DSA signature generation
2056 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2057 timing side channel attack. An attacker could use variations in the signing
2058 algorithm to recover the private key.
2060 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2065 * Timing vulnerability in ECDSA signature generation
2067 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2068 timing side channel attack. An attacker could use variations in the signing
2069 algorithm to recover the private key.
2071 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2076 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2077 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2078 of two gigabytes and the error handling improved.
2080 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2081 categorized as a normal bug, not a security issue, because the DRBG reseeds
2082 automatically and is fully functional even without additional randomness
2083 provided by the application.
2085 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2087 * Add a new ClientHello callback. Provides a callback interface that gives
2088 the application the ability to adjust the nascent SSL object at the
2089 earliest stage of ClientHello processing, immediately after extensions have
2090 been collected but before they have been processed. In particular, this
2091 callback can adjust the supported TLS versions in response to the contents
2096 * Add SM2 base algorithm support.
2100 * s390x assembly pack: add (improved) hardware-support for the following
2101 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2102 aes-cfb/cfb8, aes-ecb.
2106 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2107 parameter is no longer accepted, as it leads to a corrupt table. NULL
2108 pem_str is reserved for alias entries only.
2112 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2113 step for prime curves. The new implementation is based on formulae from
2114 differential addition-and-doubling in homogeneous projective coordinates
2115 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2116 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2117 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2118 to work in projective coordinates.
2120 *Billy Bob Brumley, Nicola Tuveri*
2122 * Change generating and checking of primes so that the error rate of not
2123 being prime depends on the intended use based on the size of the input.
2124 For larger primes this will result in more rounds of Miller-Rabin.
2125 The maximal error rate for primes with more than 1080 bits is lowered
2128 *Kurt Roeckx, Annie Yousar*
2130 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2134 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2135 moving between systems, and to avoid confusion when a Windows build is
2136 done with mingw vs with MSVC. For POSIX installs, there's still a
2137 symlink or copy named 'tsget' to avoid that confusion as well.
2141 * Revert blinding in ECDSA sign and instead make problematic addition
2142 length-invariant. Switch even to fixed-length Montgomery multiplication.
2146 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2147 step for binary curves. The new implementation is based on formulae from
2148 differential addition-and-doubling in mixed Lopez-Dahab projective
2149 coordinates, modified to independently blind the operands.
2151 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2153 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2154 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2155 EC_METHODs to implement their own specialized "ladder step", to take
2156 advantage of more favorable coordinate systems or more efficient
2157 differential addition-and-doubling algorithms.
2159 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2161 * Modified the random device based seed sources to keep the relevant
2162 file descriptors open rather than reopening them on each access.
2163 This allows such sources to operate in a chroot() jail without
2164 the associated device nodes being available. This behaviour can be
2165 controlled using RAND_keep_random_devices_open().
2169 * Numerous side-channel attack mitigations have been applied. This may have
2170 performance impacts for some algorithms for the benefit of improved
2171 security. Specific changes are noted in this change log by their respective
2176 * AIX shared library support overhaul. Switch to AIX "natural" way of
2177 handling shared libraries, which means collecting shared objects of
2178 different versions and bitnesses in one common archive. This allows to
2179 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2180 doesn't affect the way 3rd party applications are linked, only how
2181 multi-version installation is managed.
2185 * Make ec_group_do_inverse_ord() more robust and available to other
2186 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2187 mitigations are applied to the fallback BN_mod_inverse().
2188 When using this function rather than BN_mod_inverse() directly, new
2189 EC cryptosystem implementations are then safer-by-default.
2193 * Add coordinate blinding for EC_POINT and implement projective
2194 coordinate blinding for generic prime curves as a countermeasure to
2195 chosen point SCA attacks.
2197 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2199 * Add blinding to ECDSA and DSA signatures to protect against side channel
2200 attacks discovered by Keegan Ryan (NCC Group).
2204 * Enforce checking in the `pkeyutl` command to ensure that the input
2205 length does not exceed the maximum supported digest length when performing
2206 a sign, verify or verifyrecover operation.
2210 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2211 I/O in combination with something like select() or poll() will hang. This
2212 can be turned off again using SSL_CTX_clear_mode().
2213 Many applications do not properly handle non-application data records, and
2214 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2215 around the problems in those applications, but can also break some.
2216 It's recommended to read the manpages about SSL_read(), SSL_write(),
2217 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2218 SSL_CTX_set_read_ahead() again.
2222 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2223 now allow empty (zero character) pass phrases.
2227 * Apply blinding to binary field modular inversion and remove patent
2228 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2232 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2233 binary and prime elliptic curves.
2237 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2238 constant time fixed point multiplication.
2242 * Revise elliptic curve scalar multiplication with timing attack
2243 defenses: ec_wNAF_mul redirects to a constant time implementation
2244 when computing fixed point and variable point multiplication (which
2245 in OpenSSL are mostly used with secret scalars in keygen, sign,
2246 ECDH derive operations).
2247 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2250 * Updated CONTRIBUTING
2254 * Updated DRBG / RAND to request nonce and additional low entropy
2255 randomness from the system.
2257 *Matthias St. Pierre*
2259 * Updated 'openssl rehash' to use OpenSSL consistent default.
2263 * Moved the load of the ssl_conf module to libcrypto, which helps
2264 loading engines that libssl uses before libssl is initialised.
2268 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2272 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2274 *Ingo Schwarze, Rich Salz*
2276 * Added output of accepting IP address and port for 'openssl s_server'
2280 * Added a new API for TLSv1.3 ciphersuites:
2281 SSL_CTX_set_ciphersuites()
2282 SSL_set_ciphersuites()
2286 * Memory allocation failures consistently add an error to the error
2291 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2292 in libcrypto when run as setuid/setgid.
2296 * Load any config file by default when libssl is used.
2300 * Added new public header file <openssl/rand_drbg.h> and documentation
2301 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2303 *Matthias St. Pierre*
2305 * QNX support removed (cannot find contributors to get their approval
2306 for the license change).
2310 * TLSv1.3 replay protection for early data has been implemented. See the
2311 SSL_read_early_data() man page for further details.
2315 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2316 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2317 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2318 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2319 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2320 configuration has been separated out. See the ciphers man page or the
2321 SSL_CTX_set_ciphersuites() man page for more information.
2325 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2326 in responder mode now supports the new "-multi" option, which
2327 spawns the specified number of child processes to handle OCSP
2328 requests. The "-timeout" option now also limits the OCSP
2329 responder's patience to wait to receive the full client request
2330 on a newly accepted connection. Child processes are respawned
2331 as needed, and the CA index file is automatically reloaded
2332 when changed. This makes it possible to run the "ocsp" responder
2333 as a long-running service, making the OpenSSL CA somewhat more
2334 feature-complete. In this mode, most diagnostic messages logged
2335 after entering the event loop are logged via syslog(3) rather than
2340 * Added support for X448 and Ed448. Heavily based on original work by
2345 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2346 objects loaded. This adds the functions OSSL_STORE_expect() and
2347 OSSL_STORE_find() as well as needed tools to construct searches and
2348 get the search data out of them.
2352 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2353 version of OpenSSL should review their configuration settings to ensure
2354 that they are still appropriate for TLSv1.3. For further information see:
2355 <https://wiki.openssl.org/index.php/TLS1.3>
2359 * Grand redesign of the OpenSSL random generator
2361 The default RAND method now utilizes an AES-CTR DRBG according to
2362 NIST standard SP 800-90Ar1. The new random generator is essentially
2363 a port of the default random generator from the OpenSSL FIPS 2.0
2364 object module. It is a hybrid deterministic random bit generator
2365 using an AES-CTR bit stream and which seeds and reseeds itself
2366 automatically using trusted system entropy sources.
2368 Some of its new features are:
2369 - Support for multiple DRBG instances with seed chaining.
2370 - The default RAND method makes use of a DRBG.
2371 - There is a public and private DRBG instance.
2372 - The DRBG instances are fork-safe.
2373 - Keep all global DRBG instances on the secure heap if it is enabled.
2374 - The public and private DRBG instance are per thread for lock free
2377 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2379 * Changed Configure so it only says what it does and doesn't dump
2380 so much data. Instead, ./configdata.pm should be used as a script
2381 to display all sorts of configuration data.
2385 * Added processing of "make variables" to Configure.
2389 * Added SHA512/224 and SHA512/256 algorithm support.
2393 * The last traces of Netware support, first removed in 1.1.0, have
2398 * Get rid of Makefile.shared, and in the process, make the processing
2399 of certain files (rc.obj, or the .def/.map/.opt files produced from
2400 the ordinal files) more visible and hopefully easier to trace and
2401 debug (or make silent).
2405 * Make it possible to have environment variable assignments as
2406 arguments to config / Configure.
2410 * Add multi-prime RSA (RFC 8017) support.
2414 * Add SM3 implemented according to GB/T 32905-2016
2415 *Jack Lloyd <jack.lloyd@ribose.com>,*
2416 *Ronald Tse <ronald.tse@ribose.com>,*
2417 *Erick Borsboom <erick.borsboom@ribose.com>*
2419 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2420 as documented in RFC6066.
2421 Based on a patch from Tomasz Moń
2423 *Filipe Raimundo da Silva*
2425 * Add SM4 implemented according to GB/T 32907-2016.
2426 *Jack Lloyd <jack.lloyd@ribose.com>,*
2427 *Ronald Tse <ronald.tse@ribose.com>,*
2428 *Erick Borsboom <erick.borsboom@ribose.com>*
2430 * Reimplement -newreq-nodes and ERR_error_string_n; the
2431 original author does not agree with the license change.
2435 * Add ARIA AEAD TLS support.
2439 * Some macro definitions to support VS6 have been removed. Visual
2440 Studio 6 has not worked since 1.1.0
2444 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2445 without clearing the errors.
2449 * Add "atfork" functions. If building on a system that without
2450 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2451 requirements. The RAND facility now uses/requires this.
2459 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2460 not possible to disable entirely. However, it's still possible to
2461 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2464 To disable, configure with 'no-ui-console'. 'no-ui' is still
2465 possible to use as an alias. Check at compile time with the
2466 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2467 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2471 * Add a STORE module, which implements a uniform and URI based reader of
2472 stores that can contain keys, certificates, CRLs and numerous other
2473 objects. The main API is loosely based on a few stdio functions,
2474 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2475 OSSL_STORE_error and OSSL_STORE_close.
2476 The implementation uses backends called "loaders" to implement arbitrary
2477 URI schemes. There is one built in "loader" for the 'file' scheme.
2481 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2482 then adjusted to work on FreeBSD 8.4 as well.
2483 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2484 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2488 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2489 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2490 error code calls like this:
2492 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2494 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2495 that can be encoded in C. For the foreseeable future, this will only
2498 *Richard Levitte and Tim Hudson*
2500 * Removed BSD cryptodev engine.
2504 * Add a build target 'build_all_generated', to build all generated files
2505 and only that. This can be used to prepare everything that requires
2506 things like perl for a system that lacks perl and then move everything
2507 to that system and do the rest of the build there.
2511 * In the UI interface, make it possible to duplicate the user data. This
2512 can be used by engines that need to retain the data for a longer time
2513 than just the call where this user data is passed.
2517 * Ignore the '-named_curve auto' value for compatibility of applications
2520 *Tomáš Mráz <tmraz@fedoraproject.org>*
2522 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2523 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2524 alerts across multiple records (some of which could be empty). In practice
2525 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2526 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2527 support this at all. Supporting it adds significant complexity to the
2528 record layer, and its removal is unlikely to cause interoperability
2533 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2534 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2535 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2540 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2541 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2543 *Richard Levitte, Andy Polyakov*
2545 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2550 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2551 platform rather than 'mingw'.
2555 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2556 success if they are asked to add an object which already exists
2557 in the store. This change cascades to other functions which load
2558 certificates and CRLs.
2562 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2563 facilitate stack unwinding even from assembly subroutines.
2567 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2568 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2572 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2573 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2574 which is the minimum version we support.
2578 * Certificate time validation (X509_cmp_time) enforces stricter
2579 compliance with RFC 5280. Fractional seconds and timezone offsets
2580 are no longer allowed.
2584 * Add support for ARIA
2588 * s_client will now send the Server Name Indication (SNI) extension by
2589 default unless the new "-noservername" option is used. The server name is
2590 based on the host provided to the "-connect" option unless overridden by
2591 using "-servername".
2595 * Add support for SipHash
2599 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2600 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2601 prevent issues where no progress is being made and the peer continually
2602 sends unrecognised record types, using up resources processing them.
2606 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2607 using the algorithm defined in
2608 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2612 * Heartbeat support has been removed; the ABI is changed for now.
2614 *Richard Levitte, Rich Salz*
2616 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2620 * The RSA "null" method, which was partially supported to avoid patent
2621 issues, has been replaced to always returns NULL.
2628 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2630 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2631 used even when parsing explicit parameters, when loading a encoded key
2632 or calling `EC_GROUP_new_from_ecpkparameters()`/
2633 `EC_GROUP_new_from_ecparameters()`.
2634 This prevents bypass of security hardening and performance gains,
2635 especially for curves with specialized EC_METHODs.
2636 By default, if a key encoded with explicit parameters is loaded and later
2637 encoded, the output is still encoded with explicit parameters, even if
2638 internally a "named" EC_GROUP is used for computation.
2642 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2643 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2644 NULL. After this change, only the cofactor parameter can be NULL. It also
2645 does some minimal sanity checks on the passed order.
2650 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2651 An attack is simple, if the first CMS_recipientInfo is valid but the
2652 second CMS_recipientInfo is chosen ciphertext. If the second
2653 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2654 encryption key will be replaced by garbage, and the message cannot be
2655 decoded, but if the RSA decryption fails, the correct encryption key is
2656 used and the recipient will not notice the attack.
2657 As a work around for this potential attack the length of the decrypted
2658 key must be equal to the cipher default key length, in case the
2659 certifiate is not given and all recipientInfo are tried out.
2660 The old behaviour can be re-enabled in the CMS code by setting the
2661 CMS_DEBUG_DECRYPT flag.
2666 * Use Windows installation paths in the mingw builds
2668 Mingw isn't a POSIX environment per se, which means that Windows
2669 paths should be used for installation.
2674 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2676 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2677 This changes the size when using the `genpkey` command when no size is given.
2678 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2679 generation commands to use 2048 bits by default.
2683 * Prevent over long nonces in ChaCha20-Poly1305.
2685 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2686 for every encryption operation. RFC 7539 specifies that the nonce value
2687 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2688 and front pads the nonce with 0 bytes if it is less than 12
2689 bytes. However it also incorrectly allows a nonce to be set of up to 16
2690 bytes. In this case only the last 12 bytes are significant and any
2691 additional leading bytes are ignored.
2693 It is a requirement of using this cipher that nonce values are
2694 unique. Messages encrypted using a reused nonce value are susceptible to
2695 serious confidentiality and integrity attacks. If an application changes
2696 the default nonce length to be longer than 12 bytes and then makes a
2697 change to the leading bytes of the nonce expecting the new value to be a
2698 new unique nonce then such an application could inadvertently encrypt
2699 messages with a reused nonce.
2701 Additionally the ignored bytes in a long nonce are not covered by the
2702 integrity guarantee of this cipher. Any application that relies on the
2703 integrity of these ignored leading bytes of a long nonce may be further
2704 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2705 is safe because no such use sets such a long nonce value. However user
2706 applications that use this cipher directly and set a non-default nonce
2707 length to be longer than 12 bytes may be vulnerable.
2709 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2715 * Added SCA hardening for modular field inversion in EC_GROUP through
2716 a new dedicated field_inv() pointer in EC_METHOD.
2717 This also addresses a leakage affecting conversions from projective
2718 to affine coordinates.
2720 *Billy Bob Brumley, Nicola Tuveri*
2722 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2723 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2727 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2731 * Remove the 'dist' target and add a tarball building script. The
2732 'dist' target has fallen out of use, and it shouldn't be
2733 necessary to configure just to create a source distribution.
2737 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2739 * Timing vulnerability in DSA signature generation
2741 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2742 timing side channel attack. An attacker could use variations in the signing
2743 algorithm to recover the private key.
2745 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2750 * Timing vulnerability in ECDSA signature generation
2752 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2753 timing side channel attack. An attacker could use variations in the signing
2754 algorithm to recover the private key.
2756 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2761 * Add coordinate blinding for EC_POINT and implement projective
2762 coordinate blinding for generic prime curves as a countermeasure to
2763 chosen point SCA attacks.
2765 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2767 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2769 * Client DoS due to large DH parameter
2771 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2772 malicious server can send a very large prime value to the client. This will
2773 cause the client to spend an unreasonably long period of time generating a
2774 key for this prime resulting in a hang until the client has finished. This
2775 could be exploited in a Denial Of Service attack.
2777 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2782 * Cache timing vulnerability in RSA Key Generation
2784 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2785 a cache timing side channel attack. An attacker with sufficient access to
2786 mount cache timing attacks during the RSA key generation process could
2787 recover the private key.
2789 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2790 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2795 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2796 parameter is no longer accepted, as it leads to a corrupt table. NULL
2797 pem_str is reserved for alias entries only.
2801 * Revert blinding in ECDSA sign and instead make problematic addition
2802 length-invariant. Switch even to fixed-length Montgomery multiplication.
2806 * Change generating and checking of primes so that the error rate of not
2807 being prime depends on the intended use based on the size of the input.
2808 For larger primes this will result in more rounds of Miller-Rabin.
2809 The maximal error rate for primes with more than 1080 bits is lowered
2812 *Kurt Roeckx, Annie Yousar*
2814 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2818 * Add blinding to ECDSA and DSA signatures to protect against side channel
2819 attacks discovered by Keegan Ryan (NCC Group).
2823 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2824 now allow empty (zero character) pass phrases.
2828 * Certificate time validation (X509_cmp_time) enforces stricter
2829 compliance with RFC 5280. Fractional seconds and timezone offsets
2830 are no longer allowed.
2834 * Fixed a text canonicalisation bug in CMS
2836 Where a CMS detached signature is used with text content the text goes
2837 through a canonicalisation process first prior to signing or verifying a
2838 signature. This process strips trailing space at the end of lines, converts
2839 line terminators to CRLF and removes additional trailing line terminators
2840 at the end of a file. A bug in the canonicalisation process meant that
2841 some characters, such as form-feed, were incorrectly treated as whitespace
2842 and removed. This is contrary to the specification (RFC5485). This fix
2843 could mean that detached text data signed with an earlier version of
2844 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2845 signed with a fixed OpenSSL may fail to verify with an earlier version of
2846 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2847 and use the "-binary" flag (for the "cms" command line application) or set
2848 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2852 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2854 * Constructed ASN.1 types with a recursive definition could exceed the stack
2856 Constructed ASN.1 types with a recursive definition (such as can be found
2857 in PKCS7) could eventually exceed the stack given malicious input with
2858 excessive recursion. This could result in a Denial Of Service attack. There
2859 are no such structures used within SSL/TLS that come from untrusted sources
2860 so this is considered safe.
2862 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2868 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2870 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2871 effectively reduced to only comparing the least significant bit of each
2872 byte. This allows an attacker to forge messages that would be considered as
2873 authenticated in an amount of tries lower than that guaranteed by the
2874 security claims of the scheme. The module can only be compiled by the
2875 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2877 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2883 * Add a build target 'build_all_generated', to build all generated files
2884 and only that. This can be used to prepare everything that requires
2885 things like perl for a system that lacks perl and then move everything
2886 to that system and do the rest of the build there.
2890 * Backport SSL_OP_NO_RENGOTIATION
2892 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2893 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2894 changes this is no longer possible in 1.1.0. Therefore the new
2895 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2896 1.1.0 to provide equivalent functionality.
2898 Note that if an application built against 1.1.0h headers (or above) is run
2899 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2900 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2904 * Removed the OS390-Unix config target. It relied on a script that doesn't
2909 * rsaz_1024_mul_avx2 overflow bug on x86_64
2911 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2912 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2913 Analysis suggests that attacks against RSA and DSA as a result of this
2914 defect would be very difficult to perform and are not believed likely.
2915 Attacks against DH1024 are considered just feasible, because most of the
2916 work necessary to deduce information about a private key may be performed
2917 offline. The amount of resources required for such an attack would be
2918 significant. However, for an attack on TLS to be meaningful, the server
2919 would have to share the DH1024 private key among multiple clients, which is
2920 no longer an option since CVE-2016-0701.
2922 This only affects processors that support the AVX2 but not ADX extensions
2923 like Intel Haswell (4th generation).
2925 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2926 was originally found via the OSS-Fuzz project.
2931 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2933 * bn_sqrx8x_internal carry bug on x86_64
2935 There is a carry propagating bug in the x86_64 Montgomery squaring
2936 procedure. No EC algorithms are affected. Analysis suggests that attacks
2937 against RSA and DSA as a result of this defect would be very difficult to
2938 perform and are not believed likely. Attacks against DH are considered just
2939 feasible (although very difficult) because most of the work necessary to
2940 deduce information about a private key may be performed offline. The amount
2941 of resources required for such an attack would be very significant and
2942 likely only accessible to a limited number of attackers. An attacker would
2943 additionally need online access to an unpatched system using the target
2944 private key in a scenario with persistent DH parameters and a private
2945 key that is shared between multiple clients.
2947 This only affects processors that support the BMI1, BMI2 and ADX extensions
2948 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2950 This issue was reported to OpenSSL by the OSS-Fuzz project.
2955 * Malformed X.509 IPAddressFamily could cause OOB read
2957 If an X.509 certificate has a malformed IPAddressFamily extension,
2958 OpenSSL could do a one-byte buffer overread. The most likely result
2959 would be an erroneous display of the certificate in text format.
2961 This issue was reported to OpenSSL by the OSS-Fuzz project.
2966 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2968 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2969 platform rather than 'mingw'.
2973 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2974 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2975 which is the minimum version we support.
2979 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2981 * Encrypt-Then-Mac renegotiation crash
2983 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2984 negotiated where it was not in the original handshake (or vice-versa) then
2985 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2986 and servers are affected.
2988 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2993 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2995 * Truncated packet could crash via OOB read
2997 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2998 cipher is being used, then a truncated packet can cause that host to
2999 perform an out-of-bounds read, usually resulting in a crash.
3001 This issue was reported to OpenSSL by Robert Święcki of Google.
3006 * Bad (EC)DHE parameters cause a client crash
3008 If a malicious server supplies bad parameters for a DHE or ECDHE key
3009 exchange then this can result in the client attempting to dereference a
3010 NULL pointer leading to a client crash. This could be exploited in a Denial
3013 This issue was reported to OpenSSL by Guido Vranken.
3018 * BN_mod_exp may produce incorrect results on x86_64
3020 There is a carry propagating bug in the x86_64 Montgomery squaring
3021 procedure. No EC algorithms are affected. Analysis suggests that attacks
3022 against RSA and DSA as a result of this defect would be very difficult to
3023 perform and are not believed likely. Attacks against DH are considered just
3024 feasible (although very difficult) because most of the work necessary to
3025 deduce information about a private key may be performed offline. The amount
3026 of resources required for such an attack would be very significant and
3027 likely only accessible to a limited number of attackers. An attacker would
3028 additionally need online access to an unpatched system using the target
3029 private key in a scenario with persistent DH parameters and a private
3030 key that is shared between multiple clients. For example this can occur by
3031 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3032 similar to CVE-2015-3193 but must be treated as a separate problem.
3034 This issue was reported to OpenSSL by the OSS-Fuzz project.
3039 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3041 * ChaCha20/Poly1305 heap-buffer-overflow
3043 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3044 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3045 crash. This issue is not considered to be exploitable beyond a DoS.
3047 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3052 * CMS Null dereference
3054 Applications parsing invalid CMS structures can crash with a NULL pointer
3055 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3056 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3057 structure callback if an attempt is made to free certain invalid encodings.
3058 Only CHOICE structures using a callback which do not handle NULL value are
3061 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3066 * Montgomery multiplication may produce incorrect results
3068 There is a carry propagating bug in the Broadwell-specific Montgomery
3069 multiplication procedure that handles input lengths divisible by, but
3070 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3071 and DH private keys are impossible. This is because the subroutine in
3072 question is not used in operations with the private key itself and an input
3073 of the attacker's direct choice. Otherwise the bug can manifest itself as
3074 transient authentication and key negotiation failures or reproducible
3075 erroneous outcome of public-key operations with specially crafted input.
3076 Among EC algorithms only Brainpool P-512 curves are affected and one
3077 presumably can attack ECDH key negotiation. Impact was not analyzed in
3078 detail, because pre-requisites for attack are considered unlikely. Namely
3079 multiple clients have to choose the curve in question and the server has to
3080 share the private key among them, neither of which is default behaviour.
3081 Even then only clients that chose the curve will be affected.
3083 This issue was publicly reported as transient failures and was not
3084 initially recognized as a security issue. Thanks to Richard Morgan for
3085 providing reproducible case.
3090 * Removed automatic addition of RPATH in shared libraries and executables,
3091 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3095 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3097 * Fix Use After Free for large message sizes
3099 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3100 message larger than approx 16k is received then the underlying buffer to
3101 store the incoming message is reallocated and moved. Unfortunately a
3102 dangling pointer to the old location is left which results in an attempt to
3103 write to the previously freed location. This is likely to result in a
3104 crash, however it could potentially lead to execution of arbitrary code.
3106 This issue only affects OpenSSL 1.1.0a.
3108 This issue was reported to OpenSSL by Robert Święcki.
3113 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3115 * OCSP Status Request extension unbounded memory growth
3117 A malicious client can send an excessively large OCSP Status Request
3118 extension. If that client continually requests renegotiation, sending a
3119 large OCSP Status Request extension each time, then there will be unbounded
3120 memory growth on the server. This will eventually lead to a Denial Of
3121 Service attack through memory exhaustion. Servers with a default
3122 configuration are vulnerable even if they do not support OCSP. Builds using
3123 the "no-ocsp" build time option are not affected.
3125 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3130 * SSL_peek() hang on empty record
3132 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3133 sends an empty record. This could be exploited by a malicious peer in a
3134 Denial Of Service attack.
3136 This issue was reported to OpenSSL by Alex Gaynor.
3141 * Excessive allocation of memory in tls_get_message_header() and
3142 dtls1_preprocess_fragment()
3144 A (D)TLS message includes 3 bytes for its length in the header for the
3145 message. This would allow for messages up to 16Mb in length. Messages of
3146 this length are excessive and OpenSSL includes a check to ensure that a
3147 peer is sending reasonably sized messages in order to avoid too much memory
3148 being consumed to service a connection. A flaw in the logic of version
3149 1.1.0 means that memory for the message is allocated too early, prior to
3150 the excessive message length check. Due to way memory is allocated in
3151 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3152 to service a connection. This could lead to a Denial of Service through
3153 memory exhaustion. However, the excessive message length check still takes
3154 place, and this would cause the connection to immediately fail. Assuming
3155 that the application calls SSL_free() on the failed connection in a timely
3156 manner then the 21Mb of allocated memory will then be immediately freed
3157 again. Therefore the excessive memory allocation will be transitory in
3158 nature. This then means that there is only a security impact if:
3160 1) The application does not call SSL_free() in a timely manner in the event
3161 that the connection fails
3163 2) The application is working in a constrained environment where there is
3164 very little free memory
3166 3) The attacker initiates multiple connection attempts such that there are
3167 multiple connections in a state where memory has been allocated for the
3168 connection; SSL_free() has not yet been called; and there is insufficient
3169 memory to service the multiple requests.
3171 Except in the instance of (1) above any Denial Of Service is likely to be
3172 transitory because as soon as the connection fails the memory is
3173 subsequently freed again in the SSL_free() call. However there is an
3174 increased risk during this period of application crashes due to the lack of
3175 memory - which would then mean a more serious Denial of Service.
3177 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3178 (CVE-2016-6307 and CVE-2016-6308)
3182 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3183 had to be removed. Primary reason is that vendor assembler can't
3184 assemble our modules with -KPIC flag. As result it, assembly
3185 support, was not even available as option. But its lack means
3186 lack of side-channel resistant code, which is incompatible with
3187 security by todays standards. Fortunately gcc is readily available
3188 prepackaged option, which we firmly point at...
3192 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3194 * Windows command-line tool supports UTF-8 opt-in option for arguments
3195 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3196 (to any value) allows Windows user to access PKCS#12 file generated
3197 with Windows CryptoAPI and protected with non-ASCII password, as well
3198 as files generated under UTF-8 locale on Linux also protected with
3203 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3204 have been disabled by default and removed from DEFAULT, just like RC4.
3205 See the RC4 item below to re-enable both.
3209 * The method for finding the storage location for the Windows RAND seed file
3210 has changed. First we check %RANDFILE%. If that is not set then we check
3211 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3212 all else fails we fall back to C:\.
3216 * The EVP_EncryptUpdate() function has had its return type changed from void
3217 to int. A return of 0 indicates and error while a return of 1 indicates
3222 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3223 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3224 off the constant time implementation for RSA, DSA and DH have been made
3225 no-ops and deprecated.
3229 * Windows RAND implementation was simplified to only get entropy by
3230 calling CryptGenRandom(). Various other RAND-related tickets
3233 *Joseph Wylie Yandle, Rich Salz*
3235 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3236 and `OPENSSL_LH_`, respectively. The old names are available
3237 with API compatibility. They new names are now completely documented.
3241 * Unify TYPE_up_ref(obj) methods signature.
3242 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3243 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3244 int (instead of void) like all others TYPE_up_ref() methods.
3245 So now these methods also check the return value of CRYPTO_atomic_add(),
3246 and the validity of object reference counter.
3248 *fdasilvayy@gmail.com*
3250 * With Windows Visual Studio builds, the .pdb files are installed
3251 alongside the installed libraries and executables. For a static
3252 library installation, ossl_static.pdb is the associate compiler
3253 generated .pdb file to be used when linking programs.
3257 * Remove openssl.spec. Packaging files belong with the packagers.
3261 * Automatic Darwin/OSX configuration has had a refresh, it will now
3262 recognise x86_64 architectures automatically. You can still decide
3263 to build for a different bitness with the environment variable
3264 KERNEL_BITS (can be 32 or 64), for example:
3266 KERNEL_BITS=32 ./config
3270 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3271 256 bit AES and HMAC with SHA256.
3275 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3279 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3283 * To enable users to have their own config files and build file templates,
3284 Configure looks in the directory indicated by the environment variable
3285 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3286 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3287 name and is used as is.
3291 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3292 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3293 X509_CERT_FILE_CTX was removed.
3297 * "shared" builds are now the default. To create only static libraries use
3298 the "no-shared" Configure option.
3302 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3303 All of these option have not worked for some while and are fundamental
3308 * Make various cleanup routines no-ops and mark them as deprecated. Most
3309 global cleanup functions are no longer required because they are handled
3310 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3311 Explicitly de-initing can cause problems (e.g. where a library that uses
3312 OpenSSL de-inits, but an application is still using it). The affected
3313 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3314 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3315 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3316 COMP_zlib_cleanup().
3320 * --strict-warnings no longer enables runtime debugging options
3321 such as REF_DEBUG. Instead, debug options are automatically
3322 enabled with '--debug' builds.
3324 *Andy Polyakov, Emilia Käsper*
3326 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3327 have been moved out of the public header files. New functions for managing
3328 these have been added.
3332 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3333 objects have been moved out of the public header files. New
3334 functions for managing these have been added.
3338 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3339 have been moved out of the public header files. New functions for managing
3340 these have been added.
3344 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3345 moved out of the public header files. New functions for managing these
3350 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3354 * Removed the mk1mf build scripts.
3358 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3359 it is always safe to #include a header now.
3363 * Removed the aged BC-32 config and all its supporting scripts
3367 * Removed support for Ultrix, Netware, and OS/2.
3371 * Add support for HKDF.
3373 *Alessandro Ghedini*
3375 * Add support for blake2b and blake2s
3379 * Added support for "pipelining". Ciphers that have the
3380 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3381 encryptions/decryptions simultaneously. There are currently no built-in
3382 ciphers with this property but the expectation is that engines will be able
3383 to offer it to significantly improve throughput. Support has been extended
3384 into libssl so that multiple records for a single connection can be
3385 processed in one go (for >=TLS 1.1).
3389 * Added the AFALG engine. This is an async capable engine which is able to
3390 offload work to the Linux kernel. In this initial version it only supports
3391 AES128-CBC. The kernel must be version 4.1.0 or greater.
3395 * OpenSSL now uses a new threading API. It is no longer necessary to
3396 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3397 are two supported threading models: pthreads and windows threads. It is
3398 also possible to configure OpenSSL at compile time for "no-threads". The
3399 old threading API should no longer be used. The functions have been
3400 replaced with "no-op" compatibility macros.
3402 *Alessandro Ghedini, Matt Caswell*
3404 * Modify behavior of ALPN to invoke callback after SNI/servername
3405 callback, such that updates to the SSL_CTX affect ALPN.
3409 * Add SSL_CIPHER queries for authentication and key-exchange.
3413 * Changes to the DEFAULT cipherlist:
3414 - Prefer (EC)DHE handshakes over plain RSA.
3415 - Prefer AEAD ciphers over legacy ciphers.
3416 - Prefer ECDSA over RSA when both certificates are available.
3417 - Prefer TLSv1.2 ciphers/PRF.
3418 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3423 * Change the ECC default curve list to be this, in order: x25519,
3424 secp256r1, secp521r1, secp384r1.
3428 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3429 disabled by default. They can be re-enabled using the
3430 enable-weak-ssl-ciphers option to Configure.
3434 * If the server has ALPN configured, but supports no protocols that the
3435 client advertises, send a fatal "no_application_protocol" alert.
3436 This behaviour is SHALL in RFC 7301, though it isn't universally
3437 implemented by other servers.
3441 * Add X25519 support.
3442 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3443 for public and private key encoding using the format documented in
3444 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3445 key generation and key derivation.
3447 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3452 * Deprecate SRP_VBASE_get_by_user.
3453 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3454 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3455 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3456 seed, even if the seed is configured.
3458 Users should use SRP_VBASE_get1_by_user instead. Note that in
3459 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3460 also that even though configuring the SRP seed attempts to hide
3461 invalid usernames by continuing the handshake with fake
3462 credentials, this behaviour is not constant time and no strong
3463 guarantees are made that the handshake is indistinguishable from
3464 that of a valid user.
3468 * Configuration change; it's now possible to build dynamic engines
3469 without having to build shared libraries and vice versa. This
3470 only applies to the engines in `engines/`, those in `crypto/engine/`
3471 will always be built into libcrypto (i.e. "static").
3473 Building dynamic engines is enabled by default; to disable, use
3474 the configuration option "disable-dynamic-engine".
3476 The only requirements for building dynamic engines are the
3477 presence of the DSO module and building with position independent
3478 code, so they will also automatically be disabled if configuring
3479 with "disable-dso" or "disable-pic".
3481 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3482 are also taken away from openssl/opensslconf.h, as they are
3487 * Configuration change; if there is a known flag to compile
3488 position independent code, it will always be applied on the
3489 libcrypto and libssl object files, and never on the application
3490 object files. This means other libraries that use routines from
3491 libcrypto / libssl can be made into shared libraries regardless
3492 of how OpenSSL was configured.
3494 If this isn't desirable, the configuration options "disable-pic"
3495 or "no-pic" can be used to disable the use of PIC. This will
3496 also disable building shared libraries and dynamic engines.
3500 * Removed JPAKE code. It was experimental and has no wide use.
3504 * The INSTALL_PREFIX Makefile variable has been renamed to
3505 DESTDIR. That makes for less confusion on what this variable
3506 is for. Also, the configuration option --install_prefix is
3511 * Heartbeat for TLS has been removed and is disabled by default
3512 for DTLS; configure with enable-heartbeats. Code that uses the
3513 old #define's might need to be updated.
3515 *Emilia Käsper, Rich Salz*
3517 * Rename REF_CHECK to REF_DEBUG.
3521 * New "unified" build system
3523 The "unified" build system is aimed to be a common system for all
3524 platforms we support. With it comes new support for VMS.
3526 This system builds supports building in a different directory tree
3527 than the source tree. It produces one Makefile (for unix family
3528 or lookalikes), or one descrip.mms (for VMS).
3530 The source of information to make the Makefile / descrip.mms is
3531 small files called 'build.info', holding the necessary
3532 information for each directory with source to compile, and a
3533 template in Configurations, like unix-Makefile.tmpl or
3536 With this change, the library names were also renamed on Windows
3537 and on VMS. They now have names that are closer to the standard
3538 on Unix, and include the major version number, and in certain
3539 cases, the architecture they are built for. See "Notes on shared
3540 libraries" in INSTALL.
3542 We rely heavily on the perl module Text::Template.
3546 * Added support for auto-initialisation and de-initialisation of the library.
3547 OpenSSL no longer requires explicit init or deinit routines to be called,
3548 except in certain circumstances. See the OPENSSL_init_crypto() and
3549 OPENSSL_init_ssl() man pages for further information.
3553 * The arguments to the DTLSv1_listen function have changed. Specifically the
3554 "peer" argument is now expected to be a BIO_ADDR object.
3556 * Rewrite of BIO networking library. The BIO library lacked consistent
3557 support of IPv6, and adding it required some more extensive
3558 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3559 which hold all types of addresses and chains of address information.
3560 It also introduces a new API, with functions like BIO_socket,
3561 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3562 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3563 have been adapted accordingly.
3567 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3572 * CRIME protection: disable compression by default, even if OpenSSL is
3573 compiled with zlib enabled. Applications can still enable compression
3574 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3575 using the SSL_CONF library to configure compression.
3579 * The signature of the session callback configured with
3580 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3581 was explicitly marked as `const unsigned char*` instead of
3586 * Always DPURIFY. Remove the use of uninitialized memory in the
3587 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3591 * Removed many obsolete configuration items, including
3592 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3593 MD2_CHAR, MD2_INT, MD2_LONG
3595 IDEA_SHORT, IDEA_LONG
3596 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3598 *Rich Salz, with advice from Andy Polyakov*
3600 * Many BN internals have been moved to an internal header file.
3602 *Rich Salz with help from Andy Polyakov*
3604 * Configuration and writing out the results from it has changed.
3605 Files such as Makefile include/openssl/opensslconf.h and are now
3606 produced through general templates, such as Makefile.in and
3607 crypto/opensslconf.h.in and some help from the perl module
3610 Also, the center of configuration information is no longer
3611 Makefile. Instead, Configure produces a perl module in
3612 configdata.pm which holds most of the config data (in the hash
3613 table %config), the target data that comes from the target
3614 configuration in one of the `Configurations/*.conf` files (in
3619 * To clarify their intended purposes, the Configure options
3620 --prefix and --openssldir change their semantics, and become more
3621 straightforward and less interdependent.
3623 --prefix shall be used exclusively to give the location INSTALLTOP
3624 where programs, scripts, libraries, include files and manuals are
3625 going to be installed. The default is now /usr/local.
3627 --openssldir shall be used exclusively to give the default
3628 location OPENSSLDIR where certificates, private keys, CRLs are
3629 managed. This is also where the default openssl.cnf gets
3631 If the directory given with this option is a relative path, the
3632 values of both the --prefix value and the --openssldir value will
3633 be combined to become OPENSSLDIR.
3634 The default for --openssldir is INSTALLTOP/ssl.
3636 Anyone who uses --openssldir to specify where OpenSSL is to be
3637 installed MUST change to use --prefix instead.
3641 * The GOST engine was out of date and therefore it has been removed. An up
3642 to date GOST engine is now being maintained in an external repository.
3643 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3644 support for GOST ciphersuites (these are only activated if a GOST engine
3649 * EGD is no longer supported by default; use enable-egd when
3652 *Ben Kaduk and Rich Salz*
3654 * The distribution now has Makefile.in files, which are used to
3655 create Makefile's when Configure is run. *Configure must be run
3656 before trying to build now.*
3660 * The return value for SSL_CIPHER_description() for error conditions
3665 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3667 Obtaining and performing DNSSEC validation of TLSA records is
3668 the application's responsibility. The application provides
3669 the TLSA records of its choice to OpenSSL, and these are then
3670 used to authenticate the peer.
3672 The TLSA records need not even come from DNS. They can, for
3673 example, be used to implement local end-entity certificate or
3674 trust-anchor "pinning", where the "pin" data takes the form
3675 of TLSA records, which can augment or replace verification
3676 based on the usual WebPKI public certification authorities.
3680 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3681 continues to support deprecated interfaces in default builds.
3682 However, applications are strongly advised to compile their
3683 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3684 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3685 or the 1.1.0 releases.
3687 In environments in which all applications have been ported to
3688 not use any deprecated interfaces OpenSSL's Configure script
3689 should be used with the --api=1.1.0 option to entirely remove
3690 support for the deprecated features from the library and
3691 unconditionally disable them in the installed headers.
3692 Essentially the same effect can be achieved with the "no-deprecated"
3693 argument to Configure, except that this will always restrict
3694 the build to just the latest API, rather than a fixed API
3697 As applications are ported to future revisions of the API,
3698 they should update their compile-time OPENSSL_API_COMPAT define
3699 accordingly, but in most cases should be able to continue to
3700 compile with later releases.
3702 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3703 0x10000000L and 0x00908000L, respectively. However those
3704 versions did not support the OPENSSL_API_COMPAT feature, and
3705 so applications are not typically tested for explicit support
3706 of just the undeprecated features of either release.
3710 * Add support for setting the minimum and maximum supported protocol.
3711 It can bet set via the SSL_set_min_proto_version() and
3712 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3713 MaxProtocol. It's recommended to use the new APIs to disable
3714 protocols instead of disabling individual protocols using
3715 SSL_set_options() or SSL_CONF's Protocol. This change also
3716 removes support for disabling TLS 1.2 in the OpenSSL TLS
3717 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3721 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3725 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3726 and integrates ECDSA and ECDH functionality into EC. Implementations can
3727 now redirect key generation and no longer need to convert to or from
3730 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3731 include the ec.h header file instead.
3735 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3736 ciphers who are no longer supported and drops support the ephemeral RSA key
3737 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3741 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3742 opaque. For HMAC_CTX, the following constructors and destructors
3745 HMAC_CTX *HMAC_CTX_new(void);
3746 void HMAC_CTX_free(HMAC_CTX *ctx);
3748 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3749 destroy such methods has been added. See EVP_MD_meth_new(3) and
3750 EVP_CIPHER_meth_new(3) for documentation.
3753 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3754 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3755 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3756 an already created structure.
3757 2) For consistency with the majority of our object creators and
3758 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3759 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3760 for deprecated builds.
3764 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3765 cryptographic operations to be performed asynchronously as long as an
3766 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3767 further details. Libssl has also had this capability integrated with the
3768 introduction of the new mode SSL_MODE_ASYNC and associated error
3769 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3770 pages. This work was developed in partnership with Intel Corp.
3774 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3775 always enabled now. If you want to disable the support you should
3776 exclude it using the list of supported ciphers. This also means that the
3777 "-no_ecdhe" option has been removed from s_server.
3781 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3782 SSL_{CTX_}set1_curves() which can set a list.
3786 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3787 curve you want to support using SSL_{CTX_}set1_curves().
3791 * State machine rewrite. The state machine code has been significantly
3792 refactored in order to remove much duplication of code and solve issues
3793 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3794 further details). This change does have some associated API changes.
3795 Notably the SSL_state() function has been removed and replaced by
3796 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3797 SSL_set_state() has been removed altogether. The previous handshake states
3798 defined in ssl.h and ssl3.h have also been removed.
3802 * All instances of the string "ssleay" in the public API were replaced
3803 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3804 Some error codes related to internal RSA_eay API's were renamed.
3808 * The demo files in crypto/threads were moved to demo/threads.
3812 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3815 *Matt Caswell, Rich Salz*
3817 * New ASN.1 embed macro.
3819 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3820 structure is not allocated: it is part of the parent. That is instead of
3828 This reduces memory fragmentation and make it impossible to accidentally
3829 set a mandatory field to NULL.
3831 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3832 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3833 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3838 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3842 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3843 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3844 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3845 DES and RC4 ciphersuites.
3849 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3850 This changes the decoding behaviour for some invalid messages,
3851 though the change is mostly in the more lenient direction, and
3852 legacy behaviour is preserved as much as possible.
3856 * Fix no-stdio build.
3857 *David Woodhouse <David.Woodhouse@intel.com> and also*
3858 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3860 * New testing framework
3861 The testing framework has been largely rewritten and is now using
3862 perl and the perl modules Test::Harness and an extended variant of
3863 Test::More called OpenSSL::Test to do its work. All test scripts in
3864 test/ have been rewritten into test recipes, and all direct calls to
3865 executables in test/Makefile have become individual recipes using the
3866 simplified testing OpenSSL::Test::Simple.
3868 For documentation on our testing modules, do:
3870 perldoc test/testlib/OpenSSL/Test/Simple.pm
3871 perldoc test/testlib/OpenSSL/Test.pm
3875 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3876 are used; the latter aborts on memory leaks (usually checked on exit).
3877 Some undocumented "set malloc, etc., hooks" functions were removed
3878 and others were changed. All are now documented.
3882 * In DSA_generate_parameters_ex, if the provided seed is too short,
3885 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3887 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3888 from RFC4279, RFC4785, RFC5487, RFC5489.
3890 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3891 original RSA_PSK patch.
3895 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3896 era flag was never set throughout the codebase (only read). Also removed
3897 SSL3_FLAGS_POP_BUFFER which was only used if
3898 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3902 * Changed the default name options in the "ca", "crl", "req" and "x509"
3903 to be "oneline" instead of "compat".
3907 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3908 not aware of clients that still exhibit this bug, and the workaround
3909 hasn't been working properly for a while.
3913 * The return type of BIO_number_read() and BIO_number_written() as well as
3914 the corresponding num_read and num_write members in the BIO structure has
3915 changed from unsigned long to uint64_t. On platforms where an unsigned
3916 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3921 * Given the pervasive nature of TLS extensions it is inadvisable to run
3922 OpenSSL without support for them. It also means that maintaining
3923 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3924 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3928 * Removed support for the two export grade static DH ciphersuites
3929 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3930 were newly added (along with a number of other static DH ciphersuites) to
3931 1.0.2. However the two export ones have *never* worked since they were
3932 introduced. It seems strange in any case to be adding new export
3933 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3937 * Version negotiation has been rewritten. In particular SSLv23_method(),
3938 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3939 and turned into macros which simply call the new preferred function names
3940 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3941 should use the new names instead. Also as part of this change the ssl23.h
3942 header file has been removed.
3946 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3947 code and the associated standard is no longer considered fit-for-purpose.
3951 * RT2547 was closed. When generating a private key, try to make the
3952 output file readable only by the owner. This behavior change might
3953 be noticeable when interacting with other software.
3955 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3960 * Added HTTP GET support to the ocsp command.
3964 * Changed default digest for the dgst and enc commands from MD5 to
3969 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3973 * Added support for TLS extended master secret from
3974 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3975 initial patch which was a great help during development.
3979 * All libssl internal structures have been removed from the public header
3980 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3981 now redundant). Users should not attempt to access internal structures
3982 directly. Instead they should use the provided API functions.
3986 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3987 Access to deprecated functions can be re-enabled by running config with
3988 "enable-deprecated". In addition applications wishing to use deprecated
3989 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3990 will, by default, disable some transitive includes that previously existed
3991 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3995 * Added support for OCB mode. OpenSSL has been granted a patent license
3996 compatible with the OpenSSL license for use of OCB. Details are available
3997 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3998 for OCB can be removed by calling config with no-ocb.
4002 * SSLv2 support has been removed. It still supports receiving a SSLv2
4003 compatible client hello.
4007 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4008 done while fixing the error code for the key-too-small case.
4010 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4012 * CA.sh has been removed; use CA.pl instead.
4016 * Removed old DES API.
4020 * Remove various unsupported platforms:
4026 Sinix/ReliantUNIX RM400
4031 16-bit platforms such as WIN16
4035 * Clean up OPENSSL_NO_xxx #define's
4036 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4037 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4038 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4039 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4040 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4041 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4042 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4043 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4044 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4045 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4049 * Cleaned up dead code
4050 Remove all but one '#ifdef undef' which is to be looked at.
4054 * Clean up calling of xxx_free routines.
4055 Just like free(), fix most of the xxx_free routines to accept
4056 NULL. Remove the non-null checks from callers. Save much code.
4060 * Add secure heap for storage of private keys (when possible).
4061 Add BIO_s_secmem(), CBIGNUM, etc.
4062 Contributed by Akamai Technologies under our Corporate CLA.
4066 * Experimental support for a new, fast, unbiased prime candidate generator,
4067 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4069 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4071 * New output format NSS in the sess_id command line tool. This allows
4072 exporting the session id and the master key in NSS keylog format.
4074 *Martin Kaiser <martin@kaiser.cx>*
4076 * Harmonize version and its documentation. -f flag is used to display
4079 *mancha <mancha1@zoho.com>*
4081 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4082 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4084 *mancha <mancha1@zoho.com>*
4086 * Fix some double frees. These are not thought to be exploitable.
4088 *mancha <mancha1@zoho.com>*
4090 * A missing bounds check in the handling of the TLS heartbeat extension
4091 can be used to reveal up to 64k of memory to a connected client or
4094 Thanks for Neel Mehta of Google Security for discovering this bug and to
4095 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4096 preparing the fix ([CVE-2014-0160])
4098 *Adam Langley, Bodo Moeller*
4100 * Fix for the attack described in the paper "Recovering OpenSSL
4101 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4102 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4103 <http://eprint.iacr.org/2014/140>
4105 Thanks to Yuval Yarom and Naomi Benger for discovering this
4106 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4108 *Yuval Yarom and Naomi Benger*
4110 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4111 this fixes a limitation in previous versions of OpenSSL.
4115 * Experimental encrypt-then-mac support.
4117 Experimental support for encrypt then mac from
4118 draft-gutmann-tls-encrypt-then-mac-02.txt
4120 To enable it set the appropriate extension number (0x42 for the test
4121 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4123 For non-compliant peers (i.e. just about everything) this should have no
4126 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4130 * Add EVP support for key wrapping algorithms, to avoid problems with
4131 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4132 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4133 algorithms and include tests cases.
4137 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4142 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4143 MGF1 digest and OAEP label.
4147 * Make openssl verify return errors.
4149 *Chris Palmer <palmer@google.com> and Ben Laurie*
4151 * New function ASN1_TIME_diff to calculate the difference between two
4152 ASN1_TIME structures or one structure and the current time.
4156 * Update fips_test_suite to support multiple command line options. New
4157 test to induce all self test errors in sequence and check expected
4162 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4163 sign or verify all in one operation.
4167 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4168 test programs and fips_test_suite. Includes functionality to parse
4169 the minimal script output of fipsalgest.pl directly.
4173 * Add authorisation parameter to FIPS_module_mode_set().
4177 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4181 * Use separate DRBG fields for internal and external flags. New function
4182 FIPS_drbg_health_check() to perform on demand health checking. Add
4183 generation tests to fips_test_suite with reduced health check interval to
4184 demonstrate periodic health checking. Add "nodh" option to
4185 fips_test_suite to skip very slow DH test.
4189 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4194 * More extensive health check for DRBG checking many more failure modes.
4195 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4196 combination: call this in fips_test_suite.
4200 * Add support for canonical generation of DSA parameter 'g'. See
4203 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4204 POST to handle HMAC cases.
4208 * Add functions FIPS_module_version() and FIPS_module_version_text()
4209 to return numerical and string versions of the FIPS module number.
4213 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4214 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4215 outside the validated module in the FIPS capable OpenSSL.
4219 * Minor change to DRBG entropy callback semantics. In some cases
4220 there is no multiple of the block length between min_len and
4221 max_len. Allow the callback to return more than max_len bytes
4222 of entropy but discard any extra: it is the callback's responsibility
4223 to ensure that the extra data discarded does not impact the
4224 requested amount of entropy.
4228 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4229 information in FIPS186-3, SP800-57 and SP800-131A.
4233 * CCM support via EVP. Interface is very similar to GCM case except we
4234 must supply all data in one chunk (i.e. no update, final) and the
4235 message length must be supplied if AAD is used. Add algorithm test
4240 * Initial version of POST overhaul. Add POST callback to allow the status
4241 of POST to be monitored and/or failures induced. Modify fips_test_suite
4242 to use callback. Always run all selftests even if one fails.
4246 * XTS support including algorithm test driver in the fips_gcmtest program.
4247 Note: this does increase the maximum key length from 32 to 64 bytes but
4248 there should be no binary compatibility issues as existing applications
4249 will never use XTS mode.
4253 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4254 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4255 performs algorithm blocking for unapproved PRNG types. Also do not
4256 set PRNG type in FIPS_mode_set(): leave this to the application.
4257 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4258 the standard OpenSSL PRNG: set additional data to a date time vector.
4262 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4263 This shouldn't present any incompatibility problems because applications
4264 shouldn't be using these directly and any that are will need to rethink
4265 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4269 * Extensive self tests and health checking required by SP800-90 DRBG.
4270 Remove strength parameter from FIPS_drbg_instantiate and always
4271 instantiate at maximum supported strength.
4275 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4279 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4283 * New function DH_compute_key_padded() to compute a DH key and pad with
4284 leading zeroes if needed: this complies with SP800-56A et al.
4288 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4289 anything, incomplete, subject to change and largely untested at present.
4293 * Modify fipscanisteronly build option to only build the necessary object
4294 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4298 * Add experimental option FIPSSYMS to give all symbols in
4299 fipscanister.o and FIPS or fips prefix. This will avoid
4300 conflicts with future versions of OpenSSL. Add perl script
4301 util/fipsas.pl to preprocess assembly language source files
4302 and rename any affected symbols.
4306 * Add selftest checks and algorithm block of non-fips algorithms in
4307 FIPS mode. Remove DES2 from selftests.
4311 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4312 return internal method without any ENGINE dependencies. Add new
4313 tiny fips sign and verify functions.
4317 * New build option no-ec2m to disable characteristic 2 code.
4321 * New build option "fipscanisteronly". This only builds fipscanister.o
4322 and (currently) associated fips utilities. Uses the file Makefile.fips
4323 instead of Makefile.org as the prototype.
4327 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4328 Update fips_gcmtest to use IV generator.
4332 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4333 setting output buffer to NULL. The `*Final` function must be
4334 called although it will not retrieve any additional data. The tag
4335 can be set or retrieved with a ctrl. The IV length is by default 12
4336 bytes (96 bits) but can be set to an alternative value. If the IV
4337 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4342 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4343 underlying do_cipher function handles all cipher semantics itself
4344 including padding and finalisation. This is useful if (for example)
4345 an ENGINE cipher handles block padding itself. The behaviour of
4346 do_cipher is subtly changed if this flag is set: the return value
4347 is the number of characters written to the output buffer (zero is
4348 no longer an error code) or a negative error code. Also if the
4349 input buffer is NULL and length 0 finalisation should be performed.
4353 * If a candidate issuer certificate is already part of the constructed
4354 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4358 * Improve forward-security support: add functions
4360 void SSL_CTX_set_not_resumable_session_callback(
4361 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4362 void SSL_set_not_resumable_session_callback(
4363 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4365 for use by SSL/TLS servers; the callback function will be called whenever a
4366 new session is created, and gets to decide whether the session may be
4367 cached to make it resumable (return 0) or not (return 1). (As by the
4368 SSL/TLS protocol specifications, the session_id sent by the server will be
4369 empty to indicate that the session is not resumable; also, the server will
4370 not generate RFC 4507 (RFC 5077) session tickets.)
4372 A simple reasonable callback implementation is to return is_forward_secure.
4373 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4374 by the SSL/TLS server library, indicating whether it can provide forward
4377 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4379 * New -verify_name option in command line utilities to set verification
4384 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4385 Add CMAC pkey methods.
4389 * Experimental renegotiation in s_server -www mode. If the client
4390 browses /reneg connection is renegotiated. If /renegcert it is
4391 renegotiated requesting a certificate.
4395 * Add an "external" session cache for debugging purposes to s_server. This
4396 should help trace issues which normally are only apparent in deployed
4397 multi-process servers.
4401 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4402 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4403 BIO_set_cipher() and some obscure PEM functions were changed so they
4404 can now return an error. The RAND changes required a change to the
4405 RAND_METHOD structure.
4409 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4410 a gcc attribute to warn if the result of a function is ignored. This
4411 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4412 whose return value is often ignored.
4416 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4417 These allow SCTs (signed certificate timestamps) to be requested and
4418 validated when establishing a connection.
4420 *Rob Percival <robpercival@google.com>*
4425 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4427 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4428 used even when parsing explicit parameters, when loading a encoded key
4429 or calling `EC_GROUP_new_from_ecpkparameters()`/
4430 `EC_GROUP_new_from_ecparameters()`.
4431 This prevents bypass of security hardening and performance gains,
4432 especially for curves with specialized EC_METHODs.
4433 By default, if a key encoded with explicit parameters is loaded and later
4434 encoded, the output is still encoded with explicit parameters, even if
4435 internally a "named" EC_GROUP is used for computation.
4439 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4440 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4441 NULL. After this change, only the cofactor parameter can be NULL. It also
4442 does some minimal sanity checks on the passed order.
4447 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4448 An attack is simple, if the first CMS_recipientInfo is valid but the
4449 second CMS_recipientInfo is chosen ciphertext. If the second
4450 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4451 encryption key will be replaced by garbage, and the message cannot be
4452 decoded, but if the RSA decryption fails, the correct encryption key is
4453 used and the recipient will not notice the attack.
4454 As a work around for this potential attack the length of the decrypted
4455 key must be equal to the cipher default key length, in case the
4456 certifiate is not given and all recipientInfo are tried out.
4457 The old behaviour can be re-enabled in the CMS code by setting the
4458 CMS_DEBUG_DECRYPT flag.
4463 * Document issue with installation paths in diverse Windows builds
4465 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4466 binaries and run-time config file.
4471 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4473 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4474 This changes the size when using the `genpkey` command when no size is given.
4475 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4476 generation commands to use 2048 bits by default.
4480 * Add FIPS support for Android Arm 64-bit
4482 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4483 Module in Version 2.0.10. For some reason, the corresponding target
4484 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4485 built with FIPS support on Android Arm 64-bit. This omission has been
4488 *Matthias St. Pierre*
4490 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4492 * 0-byte record padding oracle
4494 If an application encounters a fatal protocol error and then calls
4495 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4496 then OpenSSL can respond differently to the calling application if a 0 byte
4497 record is received with invalid padding compared to if a 0 byte record is
4498 received with an invalid MAC. If the application then behaves differently
4499 based on that in a way that is detectable to the remote peer, then this
4500 amounts to a padding oracle that could be used to decrypt data.
4502 In order for this to be exploitable "non-stitched" ciphersuites must be in
4503 use. Stitched ciphersuites are optimised implementations of certain
4504 commonly used ciphersuites. Also the application must call SSL_shutdown()
4505 twice even if a protocol error has occurred (applications should not do
4506 this but some do anyway).
4508 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4509 Aviram, with additional investigation by Steven Collison and Andrew
4510 Hourselt. It was reported to OpenSSL on 10th December 2018.
4515 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4519 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4521 * Microarchitecture timing vulnerability in ECC scalar multiplication
4523 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4524 shown to be vulnerable to a microarchitecture timing side channel attack.
4525 An attacker with sufficient access to mount local timing attacks during
4526 ECDSA signature generation could recover the private key.
4528 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4529 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4535 * Timing vulnerability in DSA signature generation
4537 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4538 timing side channel attack. An attacker could use variations in the signing
4539 algorithm to recover the private key.
4541 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4546 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4547 Module, accidentally introduced while backporting security fixes from the
4548 development branch and hindering the use of ECC in FIPS mode.
4552 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4554 * Client DoS due to large DH parameter
4556 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4557 malicious server can send a very large prime value to the client. This will
4558 cause the client to spend an unreasonably long period of time generating a
4559 key for this prime resulting in a hang until the client has finished. This
4560 could be exploited in a Denial Of Service attack.
4562 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4567 * Cache timing vulnerability in RSA Key Generation
4569 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4570 a cache timing side channel attack. An attacker with sufficient access to
4571 mount cache timing attacks during the RSA key generation process could
4572 recover the private key.
4574 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4575 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4580 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4581 parameter is no longer accepted, as it leads to a corrupt table. NULL
4582 pem_str is reserved for alias entries only.
4586 * Revert blinding in ECDSA sign and instead make problematic addition
4587 length-invariant. Switch even to fixed-length Montgomery multiplication.
4591 * Change generating and checking of primes so that the error rate of not
4592 being prime depends on the intended use based on the size of the input.
4593 For larger primes this will result in more rounds of Miller-Rabin.
4594 The maximal error rate for primes with more than 1080 bits is lowered
4597 *Kurt Roeckx, Annie Yousar*
4599 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4603 * Add blinding to ECDSA and DSA signatures to protect against side channel
4604 attacks discovered by Keegan Ryan (NCC Group).
4608 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4609 now allow empty (zero character) pass phrases.
4613 * Certificate time validation (X509_cmp_time) enforces stricter
4614 compliance with RFC 5280. Fractional seconds and timezone offsets
4615 are no longer allowed.
4619 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4621 * Constructed ASN.1 types with a recursive definition could exceed the stack
4623 Constructed ASN.1 types with a recursive definition (such as can be found
4624 in PKCS7) could eventually exceed the stack given malicious input with
4625 excessive recursion. This could result in a Denial Of Service attack. There
4626 are no such structures used within SSL/TLS that come from untrusted sources
4627 so this is considered safe.
4629 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4635 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4637 * Read/write after SSL object in error state
4639 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4640 mechanism. The intent was that if a fatal error occurred during a handshake
4641 then OpenSSL would move into the error state and would immediately fail if
4642 you attempted to continue the handshake. This works as designed for the
4643 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4644 SSL_connect()), however due to a bug it does not work correctly if
4645 SSL_read() or SSL_write() is called directly. In that scenario, if the
4646 handshake fails then a fatal error will be returned in the initial function
4647 call. If SSL_read()/SSL_write() is subsequently called by the application
4648 for the same SSL object then it will succeed and the data is passed without
4649 being decrypted/encrypted directly from the SSL/TLS record layer.
4651 In order to exploit this issue an application bug would have to be present
4652 that resulted in a call to SSL_read()/SSL_write() being issued after having
4653 already received a fatal error.
4655 This issue was reported to OpenSSL by David Benjamin (Google).
4660 * rsaz_1024_mul_avx2 overflow bug on x86_64
4662 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4663 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4664 Analysis suggests that attacks against RSA and DSA as a result of this
4665 defect would be very difficult to perform and are not believed likely.
4666 Attacks against DH1024 are considered just feasible, because most of the
4667 work necessary to deduce information about a private key may be performed
4668 offline. The amount of resources required for such an attack would be
4669 significant. However, for an attack on TLS to be meaningful, the server
4670 would have to share the DH1024 private key among multiple clients, which is
4671 no longer an option since CVE-2016-0701.
4673 This only affects processors that support the AVX2 but not ADX extensions
4674 like Intel Haswell (4th generation).
4676 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4677 was originally found via the OSS-Fuzz project.
4682 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4684 * bn_sqrx8x_internal carry bug on x86_64
4686 There is a carry propagating bug in the x86_64 Montgomery squaring
4687 procedure. No EC algorithms are affected. Analysis suggests that attacks
4688 against RSA and DSA as a result of this defect would be very difficult to
4689 perform and are not believed likely. Attacks against DH are considered just
4690 feasible (although very difficult) because most of the work necessary to
4691 deduce information about a private key may be performed offline. The amount
4692 of resources required for such an attack would be very significant and
4693 likely only accessible to a limited number of attackers. An attacker would
4694 additionally need online access to an unpatched system using the target
4695 private key in a scenario with persistent DH parameters and a private
4696 key that is shared between multiple clients.
4698 This only affects processors that support the BMI1, BMI2 and ADX extensions
4699 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4701 This issue was reported to OpenSSL by the OSS-Fuzz project.
4706 * Malformed X.509 IPAddressFamily could cause OOB read
4708 If an X.509 certificate has a malformed IPAddressFamily extension,
4709 OpenSSL could do a one-byte buffer overread. The most likely result
4710 would be an erroneous display of the certificate in text format.
4712 This issue was reported to OpenSSL by the OSS-Fuzz project.
4717 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4719 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4720 platform rather than 'mingw'.
4724 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4726 * Truncated packet could crash via OOB read
4728 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4729 cipher is being used, then a truncated packet can cause that host to
4730 perform an out-of-bounds read, usually resulting in a crash.
4732 This issue was reported to OpenSSL by Robert Święcki of Google.
4737 * BN_mod_exp may produce incorrect results on x86_64
4739 There is a carry propagating bug in the x86_64 Montgomery squaring
4740 procedure. No EC algorithms are affected. Analysis suggests that attacks
4741 against RSA and DSA as a result of this defect would be very difficult to
4742 perform and are not believed likely. Attacks against DH are considered just
4743 feasible (although very difficult) because most of the work necessary to
4744 deduce information about a private key may be performed offline. The amount
4745 of resources required for such an attack would be very significant and
4746 likely only accessible to a limited number of attackers. An attacker would
4747 additionally need online access to an unpatched system using the target
4748 private key in a scenario with persistent DH parameters and a private
4749 key that is shared between multiple clients. For example this can occur by
4750 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4751 similar to CVE-2015-3193 but must be treated as a separate problem.
4753 This issue was reported to OpenSSL by the OSS-Fuzz project.
4758 * Montgomery multiplication may produce incorrect results
4760 There is a carry propagating bug in the Broadwell-specific Montgomery
4761 multiplication procedure that handles input lengths divisible by, but
4762 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4763 and DH private keys are impossible. This is because the subroutine in
4764 question is not used in operations with the private key itself and an input
4765 of the attacker's direct choice. Otherwise the bug can manifest itself as
4766 transient authentication and key negotiation failures or reproducible
4767 erroneous outcome of public-key operations with specially crafted input.
4768 Among EC algorithms only Brainpool P-512 curves are affected and one
4769 presumably can attack ECDH key negotiation. Impact was not analyzed in
4770 detail, because pre-requisites for attack are considered unlikely. Namely
4771 multiple clients have to choose the curve in question and the server has to
4772 share the private key among them, neither of which is default behaviour.
4773 Even then only clients that chose the curve will be affected.
4775 This issue was publicly reported as transient failures and was not
4776 initially recognized as a security issue. Thanks to Richard Morgan for
4777 providing reproducible case.
4782 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4783 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4784 prevent issues where no progress is being made and the peer continually
4785 sends unrecognised record types, using up resources processing them.
4789 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4791 * Missing CRL sanity check
4793 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4794 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4795 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4797 This issue only affects the OpenSSL 1.0.2i
4802 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4804 * OCSP Status Request extension unbounded memory growth
4806 A malicious client can send an excessively large OCSP Status Request
4807 extension. If that client continually requests renegotiation, sending a
4808 large OCSP Status Request extension each time, then there will be unbounded
4809 memory growth on the server. This will eventually lead to a Denial Of
4810 Service attack through memory exhaustion. Servers with a default
4811 configuration are vulnerable even if they do not support OCSP. Builds using
4812 the "no-ocsp" build time option are not affected.
4814 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4819 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4822 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4828 * OOB write in MDC2_Update()
4830 An overflow can occur in MDC2_Update() either if called directly or
4831 through the EVP_DigestUpdate() function using MDC2. If an attacker
4832 is able to supply very large amounts of input data after a previous
4833 call to EVP_EncryptUpdate() with a partial block then a length check
4834 can overflow resulting in a heap corruption.
4836 The amount of data needed is comparable to SIZE_MAX which is impractical
4839 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4844 * Malformed SHA512 ticket DoS
4846 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4847 DoS attack where a malformed ticket will result in an OOB read which will
4850 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4851 a custom server callback and ticket lookup mechanism.
4853 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4858 * OOB write in BN_bn2dec()
4860 The function BN_bn2dec() does not check the return value of BN_div_word().
4861 This can cause an OOB write if an application uses this function with an
4862 overly large BIGNUM. This could be a problem if an overly large certificate
4863 or CRL is printed out from an untrusted source. TLS is not affected because
4864 record limits will reject an oversized certificate before it is parsed.
4866 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4871 * OOB read in TS_OBJ_print_bio()
4873 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4874 the total length the OID text representation would use and not the amount
4875 of data written. This will result in OOB reads when large OIDs are
4878 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4883 * Pointer arithmetic undefined behaviour
4885 Avoid some undefined pointer arithmetic
4887 A common idiom in the codebase is to check limits in the following manner:
4890 Where "p" points to some malloc'd data of SIZE bytes and
4893 "len" here could be from some externally supplied data (e.g. from a TLS
4896 The rules of C pointer arithmetic are such that "p + len" is only well
4897 defined where len <= SIZE. Therefore the above idiom is actually
4898 undefined behaviour.
4900 For example this could cause problems if some malloc implementation
4901 provides an address for "p" such that "p + len" actually overflows for
4902 values of len that are too big and therefore p + len < limit.
4904 This issue was reported to OpenSSL by Guido Vranken
4909 * Constant time flag not preserved in DSA signing
4911 Operations in the DSA signing algorithm should run in constant time in
4912 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4913 implementation means that a non-constant time codepath is followed for
4914 certain operations. This has been demonstrated through a cache-timing
4915 attack to be sufficient for an attacker to recover the private DSA key.
4917 This issue was reported by César Pereida (Aalto University), Billy Brumley
4918 (Tampere University of Technology), and Yuval Yarom (The University of
4919 Adelaide and NICTA).
4924 * DTLS buffered message DoS
4926 In a DTLS connection where handshake messages are delivered out-of-order
4927 those messages that OpenSSL is not yet ready to process will be buffered
4928 for later use. Under certain circumstances, a flaw in the logic means that
4929 those messages do not get removed from the buffer even though the handshake
4930 has been completed. An attacker could force up to approx. 15 messages to
4931 remain in the buffer when they are no longer required. These messages will
4932 be cleared when the DTLS connection is closed. The default maximum size for
4933 a message is 100k. Therefore the attacker could force an additional 1500k
4934 to be consumed per connection. By opening many simulataneous connections an
4935 attacker could cause a DoS attack through memory exhaustion.
4937 This issue was reported to OpenSSL by Quan Luo.
4942 * DTLS replay protection DoS
4944 A flaw in the DTLS replay attack protection mechanism means that records
4945 that arrive for future epochs update the replay protection "window" before
4946 the MAC for the record has been validated. This could be exploited by an
4947 attacker by sending a record for the next epoch (which does not have to
4948 decrypt or have a valid MAC), with a very large sequence number. This means
4949 that all subsequent legitimate packets are dropped causing a denial of
4950 service for a specific DTLS connection.
4952 This issue was reported to OpenSSL by the OCAP audit team.
4957 * Certificate message OOB reads
4959 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4960 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4961 theoretical DoS risk but this has not been observed in practice on common
4964 The messages affected are client certificate, client certificate request
4965 and server certificate. As a result the attack can only be performed
4966 against a client or a server which enables client authentication.
4968 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4973 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4975 * Prevent padding oracle in AES-NI CBC MAC check
4977 A MITM attacker can use a padding oracle attack to decrypt traffic
4978 when the connection uses an AES CBC cipher and the server support
4981 This issue was introduced as part of the fix for Lucky 13 padding
4982 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4983 constant time by making sure that always the same bytes are read and
4984 compared against either the MAC or padding bytes. But it no longer
4985 checked that there was enough data to have both the MAC and padding
4988 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4993 * Fix EVP_EncodeUpdate overflow
4995 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4996 Base64 encoding of binary data. If an attacker is able to supply very large
4997 amounts of input data then a length check can overflow resulting in a heap
5000 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5001 the `PEM_write_bio*` family of functions. These are mainly used within the
5002 OpenSSL command line applications, so any application which processes data
5003 from an untrusted source and outputs it as a PEM file should be considered
5004 vulnerable to this issue. User applications that call these APIs directly
5005 with large amounts of untrusted data may also be vulnerable.
5007 This issue was reported by Guido Vranken.
5012 * Fix EVP_EncryptUpdate overflow
5014 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5015 is able to supply very large amounts of input data after a previous call to
5016 EVP_EncryptUpdate() with a partial block then a length check can overflow
5017 resulting in a heap corruption. Following an analysis of all OpenSSL
5018 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5019 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5020 the first called function after an EVP_EncryptInit(), and therefore that
5021 specific call must be safe. The second form is where the length passed to
5022 EVP_EncryptUpdate() can be seen from the code to be some small value and
5023 therefore there is no possibility of an overflow. Since all instances are
5024 one of these two forms, it is believed that there can be no overflows in
5025 internal code due to this problem. It should be noted that
5026 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5027 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5028 of these calls have also been analysed too and it is believed there are no
5029 instances in internal usage where an overflow could occur.
5031 This issue was reported by Guido Vranken.
5036 * Prevent ASN.1 BIO excessive memory allocation
5038 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5039 a short invalid encoding can cause allocation of large amounts of memory
5040 potentially consuming excessive resources or exhausting memory.
5042 Any application parsing untrusted data through d2i BIO functions is
5043 affected. The memory based functions such as d2i_X509() are *not* affected.
5044 Since the memory based functions are used by the TLS library, TLS
5045 applications are not affected.
5047 This issue was reported by Brian Carpenter.
5054 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5055 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5056 in arbitrary stack data being returned in the buffer.
5058 This issue was reported by Guido Vranken.
5063 * Modify behavior of ALPN to invoke callback after SNI/servername
5064 callback, such that updates to the SSL_CTX affect ALPN.
5068 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5073 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5074 methods are enabled and ssl2 is disabled the methods return NULL.
5078 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5080 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5081 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5082 provide any "EXPORT" or "LOW" strength ciphers.
5086 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5087 is by default disabled at build-time. Builds that are not configured with
5088 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5089 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5090 will need to explicitly call either of:
5092 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5094 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5096 as appropriate. Even if either of those is used, or the application
5097 explicitly uses the version-specific SSLv2_method() or its client and
5098 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5099 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5100 ciphers, and SSLv2 56-bit DES are no longer available.
5105 * Fix a double-free in DSA code
5107 A double free bug was discovered when OpenSSL parses malformed DSA private
5108 keys and could lead to a DoS attack or memory corruption for applications
5109 that receive DSA private keys from untrusted sources. This scenario is
5112 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5118 * Disable SRP fake user seed to address a server memory leak.
5120 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5122 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5123 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5124 was changed to ignore the "fake user" SRP seed, even if the seed
5127 Users should use SRP_VBASE_get1_by_user instead. Note that in
5128 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5129 also that even though configuring the SRP seed attempts to hide
5130 invalid usernames by continuing the handshake with fake
5131 credentials, this behaviour is not constant time and no strong
5132 guarantees are made that the handshake is indistinguishable from
5133 that of a valid user.
5138 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5140 In the BN_hex2bn function the number of hex digits is calculated using an
5141 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5142 large values of `i` this can result in `bn_expand` not allocating any
5143 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5144 field as NULL leading to a subsequent NULL ptr deref. For very large values
5145 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5146 In this case memory is allocated to the internal BIGNUM data field, but it
5147 is insufficiently sized leading to heap corruption. A similar issue exists
5148 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5149 is ever called by user applications with very large untrusted hex/dec data.
5150 This is anticipated to be a rare occurrence.
5152 All OpenSSL internal usage of these functions use data that is not expected
5153 to be untrusted, e.g. config file data or application command line
5154 arguments. If user developed applications generate config file data based
5155 on untrusted data then it is possible that this could also lead to security
5156 consequences. This is also anticipated to be rare.
5158 This issue was reported to OpenSSL by Guido Vranken.
5163 * Fix memory issues in `BIO_*printf` functions
5165 The internal `fmtstr` function used in processing a "%s" format string in
5166 the `BIO_*printf` functions could overflow while calculating the length of a
5167 string and cause an OOB read when printing very long strings.
5169 Additionally the internal `doapr_outch` function can attempt to write to an
5170 OOB memory location (at an offset from the NULL pointer) in the event of a
5171 memory allocation failure. In 1.0.2 and below this could be caused where
5172 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5173 could be in processing a very long "%s" format string. Memory leaks can
5176 The first issue may mask the second issue dependent on compiler behaviour.
5177 These problems could enable attacks where large amounts of untrusted data
5178 is passed to the `BIO_*printf` functions. If applications use these functions
5179 in this way then they could be vulnerable. OpenSSL itself uses these
5180 functions when printing out human-readable dumps of ASN.1 data. Therefore
5181 applications that print this data could be vulnerable if the data is from
5182 untrusted sources. OpenSSL command line applications could also be
5183 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5184 as command line arguments.
5186 Libssl is not considered directly vulnerable. Additionally certificates etc
5187 received via remote connections via libssl are also unlikely to be able to
5188 trigger these issues because of message size limits enforced within libssl.
5190 This issue was reported to OpenSSL Guido Vranken.
5195 * Side channel attack on modular exponentiation
5197 A side-channel attack was found which makes use of cache-bank conflicts on
5198 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5199 of RSA keys. The ability to exploit this issue is limited as it relies on
5200 an attacker who has control of code in a thread running on the same
5201 hyper-threaded core as the victim thread which is performing decryptions.
5203 This issue was reported to OpenSSL by Yuval Yarom, The University of
5204 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5205 Nadia Heninger, University of Pennsylvania with more information at
5206 <http://cachebleed.info>.
5211 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5212 if no keysize is specified with default_bits. This fixes an
5213 omission in an earlier change that changed all RSA/DSA key generation
5214 commands to use 2048 bits by default.
5218 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5220 * DH small subgroups
5222 Historically OpenSSL only ever generated DH parameters based on "safe"
5223 primes. More recently (in version 1.0.2) support was provided for
5224 generating X9.42 style parameter files such as those required for RFC 5114
5225 support. The primes used in such files may not be "safe". Where an
5226 application is using DH configured with parameters based on primes that are
5227 not "safe" then an attacker could use this fact to find a peer's private
5228 DH exponent. This attack requires that the attacker complete multiple
5229 handshakes in which the peer uses the same private DH exponent. For example
5230 this could be used to discover a TLS server's private DH exponent if it's
5231 reusing the private DH exponent or it's using a static DH ciphersuite.
5233 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5234 TLS. It is not on by default. If the option is not set then the server
5235 reuses the same private DH exponent for the life of the server process and
5236 would be vulnerable to this attack. It is believed that many popular
5237 applications do set this option and would therefore not be at risk.
5239 The fix for this issue adds an additional check where a "q" parameter is
5240 available (as is the case in X9.42 based parameters). This detects the
5241 only known attack, and is the only possible defense for static DH
5242 ciphersuites. This could have some performance impact.
5244 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5245 default and cannot be disabled. This could have some performance impact.
5247 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5252 * SSLv2 doesn't block disabled ciphers
5254 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5255 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5256 been disabled, provided that the SSLv2 protocol was not also disabled via
5259 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5260 and Sebastian Schinzel.
5265 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5267 * BN_mod_exp may produce incorrect results on x86_64
5269 There is a carry propagating bug in the x86_64 Montgomery squaring
5270 procedure. No EC algorithms are affected. Analysis suggests that attacks
5271 against RSA and DSA as a result of this defect would be very difficult to
5272 perform and are not believed likely. Attacks against DH are considered just
5273 feasible (although very difficult) because most of the work necessary to
5274 deduce information about a private key may be performed offline. The amount
5275 of resources required for such an attack would be very significant and
5276 likely only accessible to a limited number of attackers. An attacker would
5277 additionally need online access to an unpatched system using the target
5278 private key in a scenario with persistent DH parameters and a private
5279 key that is shared between multiple clients. For example this can occur by
5280 default in OpenSSL DHE based SSL/TLS ciphersuites.
5282 This issue was reported to OpenSSL by Hanno Böck.
5287 * Certificate verify crash with missing PSS parameter
5289 The signature verification routines will crash with a NULL pointer
5290 dereference if presented with an ASN.1 signature using the RSA PSS
5291 algorithm and absent mask generation function parameter. Since these
5292 routines are used to verify certificate signature algorithms this can be
5293 used to crash any certificate verification operation and exploited in a
5294 DoS attack. Any application which performs certificate verification is
5295 vulnerable including OpenSSL clients and servers which enable client
5298 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5303 * X509_ATTRIBUTE memory leak
5305 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5306 memory. This structure is used by the PKCS#7 and CMS routines so any
5307 application which reads PKCS#7 or CMS data from untrusted sources is
5308 affected. SSL/TLS is not affected.
5310 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5316 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5317 This changes the decoding behaviour for some invalid messages,
5318 though the change is mostly in the more lenient direction, and
5319 legacy behaviour is preserved as much as possible.
5323 * In DSA_generate_parameters_ex, if the provided seed is too short,
5326 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5328 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5330 * Alternate chains certificate forgery
5332 During certificate verification, OpenSSL will attempt to find an
5333 alternative certificate chain if the first attempt to build such a chain
5334 fails. An error in the implementation of this logic can mean that an
5335 attacker could cause certain checks on untrusted certificates to be
5336 bypassed, such as the CA flag, enabling them to use a valid leaf
5337 certificate to act as a CA and "issue" an invalid certificate.
5339 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5344 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5346 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5347 incompatibility in the handling of HMAC. The previous ABI has now been
5352 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5354 * Malformed ECParameters causes infinite loop
5356 When processing an ECParameters structure OpenSSL enters an infinite loop
5357 if the curve specified is over a specially malformed binary polynomial
5360 This can be used to perform denial of service against any
5361 system which processes public keys, certificate requests or
5362 certificates. This includes TLS clients and TLS servers with
5363 client authentication enabled.
5365 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5370 * Exploitable out-of-bounds read in X509_cmp_time
5372 X509_cmp_time does not properly check the length of the ASN1_TIME
5373 string and can read a few bytes out of bounds. In addition,
5374 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5377 An attacker can use this to craft malformed certificates and CRLs of
5378 various sizes and potentially cause a segmentation fault, resulting in
5379 a DoS on applications that verify certificates or CRLs. TLS clients
5380 that verify CRLs are affected. TLS clients and servers with client
5381 authentication enabled may be affected if they use custom verification
5384 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5385 independently by Hanno Böck.
5390 * PKCS7 crash with missing EnvelopedContent
5392 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5393 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5394 with missing content and trigger a NULL pointer dereference on parsing.
5396 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5397 structures from untrusted sources are affected. OpenSSL clients and
5398 servers are not affected.
5400 This issue was reported to OpenSSL by Michal Zalewski (Google).
5405 * CMS verify infinite loop with unknown hash function
5407 When verifying a signedData message the CMS code can enter an infinite loop
5408 if presented with an unknown hash function OID. This can be used to perform
5409 denial of service against any system which verifies signedData messages using
5411 This issue was reported to OpenSSL by Johannes Bauer.
5416 * Race condition handling NewSessionTicket
5418 If a NewSessionTicket is received by a multi-threaded client when attempting to
5419 reuse a previous ticket then a race condition can occur potentially leading to
5420 a double free of the ticket data.
5425 * Only support 256-bit or stronger elliptic curves with the
5426 'ecdh_auto' setting (server) or by default (client). Of supported
5427 curves, prefer P-256 (both).
5431 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5433 * ClientHello sigalgs DoS fix
5435 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5436 invalid signature algorithms extension a NULL pointer dereference will
5437 occur. This can be exploited in a DoS attack against the server.
5439 This issue was was reported to OpenSSL by David Ramos of Stanford
5443 *Stephen Henson and Matt Caswell*
5445 * Multiblock corrupted pointer fix
5447 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5448 feature only applies on 64 bit x86 architecture platforms that support AES
5449 NI instructions. A defect in the implementation of "multiblock" can cause
5450 OpenSSL's internal write buffer to become incorrectly set to NULL when
5451 using non-blocking IO. Typically, when the user application is using a
5452 socket BIO for writing, this will only result in a failed connection.
5453 However if some other BIO is used then it is likely that a segmentation
5454 fault will be triggered, thus enabling a potential DoS attack.
5456 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5461 * Segmentation fault in DTLSv1_listen fix
5463 The DTLSv1_listen function is intended to be stateless and processes the
5464 initial ClientHello from many peers. It is common for user code to loop
5465 over the call to DTLSv1_listen until a valid ClientHello is received with
5466 an associated cookie. A defect in the implementation of DTLSv1_listen means
5467 that state is preserved in the SSL object from one invocation to the next
5468 that can lead to a segmentation fault. Errors processing the initial
5469 ClientHello can trigger this scenario. An example of such an error could be
5470 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5473 This issue was reported to OpenSSL by Per Allansson.
5478 * Segmentation fault in ASN1_TYPE_cmp fix
5480 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5481 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5482 certificate signature algorithm consistency this can be used to crash any
5483 certificate verification operation and exploited in a DoS attack. Any
5484 application which performs certificate verification is vulnerable including
5485 OpenSSL clients and servers which enable client authentication.
5490 * Segmentation fault for invalid PSS parameters fix
5492 The signature verification routines will crash with a NULL pointer
5493 dereference if presented with an ASN.1 signature using the RSA PSS
5494 algorithm and invalid parameters. Since these routines are used to verify
5495 certificate signature algorithms this can be used to crash any
5496 certificate verification operation and exploited in a DoS attack. Any
5497 application which performs certificate verification is vulnerable including
5498 OpenSSL clients and servers which enable client authentication.
5500 This issue was was reported to OpenSSL by Brian Carpenter.
5505 * ASN.1 structure reuse memory corruption fix
5507 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5508 memory corruption via an invalid write. Such reuse is and has been
5509 strongly discouraged and is believed to be rare.
5511 Applications that parse structures containing CHOICE or ANY DEFINED BY
5512 components may be affected. Certificate parsing (d2i_X509 and related
5513 functions) are however not affected. OpenSSL clients and servers are
5519 * PKCS7 NULL pointer dereferences fix
5521 The PKCS#7 parsing code does not handle missing outer ContentInfo
5522 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5523 missing content and trigger a NULL pointer dereference on parsing.
5525 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5526 otherwise parse PKCS#7 structures from untrusted sources are
5527 affected. OpenSSL clients and servers are not affected.
5529 This issue was reported to OpenSSL by Michal Zalewski (Google).
5534 * DoS via reachable assert in SSLv2 servers fix
5536 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5537 servers that both support SSLv2 and enable export cipher suites by sending
5538 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5540 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5541 (OpenSSL development team).
5546 * Empty CKE with client auth and DHE fix
5548 If client auth is used then a server can seg fault in the event of a DHE
5549 ciphersuite being selected and a zero length ClientKeyExchange message
5550 being sent by the client. This could be exploited in a DoS attack.
5555 * Handshake with unseeded PRNG fix
5557 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5558 with an unseeded PRNG. The conditions are:
5559 - The client is on a platform where the PRNG has not been seeded
5560 automatically, and the user has not seeded manually
5561 - A protocol specific client method version has been used (i.e. not
5562 SSL_client_methodv23)
5563 - A ciphersuite is used that does not require additional random data from
5564 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5566 If the handshake succeeds then the client random that has been used will
5567 have been generated from a PRNG with insufficient entropy and therefore the
5568 output may be predictable.
5570 For example using the following command with an unseeded openssl will
5571 succeed on an unpatched platform:
5573 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5578 * Use After Free following d2i_ECPrivatekey error fix
5580 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5581 could cause a use after free condition. This, in turn, could cause a double
5582 free in several private key parsing functions (such as d2i_PrivateKey
5583 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5584 for applications that receive EC private keys from untrusted
5585 sources. This scenario is considered rare.
5587 This issue was discovered by the BoringSSL project and fixed in their
5593 * X509_to_X509_REQ NULL pointer deref fix
5595 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5596 the certificate key is invalid. This function is rarely used in practice.
5598 This issue was discovered by Brian Carpenter.
5603 * Removed the export ciphers from the DEFAULT ciphers
5607 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5609 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5610 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5611 So far those who have to target multiple platforms would compromise
5612 and argue that binary targeting say ARMv5 would still execute on
5613 ARMv8. "Universal" build resolves this compromise by providing
5614 near-optimal performance even on newer platforms.
5618 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5619 (other platforms pending).
5621 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5623 * Add support for the SignedCertificateTimestampList certificate and
5624 OCSP response extensions from RFC6962.
5628 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5629 for corner cases. (Certain input points at infinity could lead to
5630 bogus results, with non-infinity inputs mapped to infinity too.)
5634 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5635 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5636 common cases are optimized and there still is room for further
5637 improvements. Vector Permutation AES for Altivec is also added.
5641 * Add support for little-endian ppc64 Linux target.
5643 *Marcelo Cerri (IBM)*
5645 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5646 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5647 are optimized and there still is room for further improvements.
5648 Both 32- and 64-bit modes are supported.
5650 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5652 * Improved ARMv7 NEON support.
5656 * Support for SPARC Architecture 2011 crypto extensions, first
5657 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5658 SHA256/512, MD5, GHASH and modular exponentiation.
5660 *Andy Polyakov, David Miller*
5662 * Accelerated modular exponentiation for Intel processors, a.k.a.
5665 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5667 * Support for new and upcoming Intel processors, including AVX2,
5668 BMI and SHA ISA extensions. This includes additional "stitched"
5669 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5672 This work was sponsored by Intel Corp.
5676 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5677 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5678 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5682 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5683 this fixes a limitation in previous versions of OpenSSL.
5687 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5688 MGF1 digest and OAEP label.
5692 * Add EVP support for key wrapping algorithms, to avoid problems with
5693 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5694 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5695 algorithms and include tests cases.
5699 * Add functions to allocate and set the fields of an ECDSA_METHOD
5702 *Douglas E. Engert, Steve Henson*
5704 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5705 difference in days and seconds between two tm or ASN1_TIME structures.
5709 * Add -rev test option to s_server to just reverse order of characters
5710 received by client and send back to server. Also prints an abbreviated
5711 summary of the connection parameters.
5715 * New option -brief for s_client and s_server to print out a brief summary
5716 of connection parameters.
5720 * Add callbacks for arbitrary TLS extensions.
5722 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5724 * New option -crl_download in several openssl utilities to download CRLs
5725 from CRLDP extension in certificates.
5729 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5733 * New function X509_CRL_diff to generate a delta CRL from the difference
5734 of two full CRLs. Add support to "crl" utility.
5738 * New functions to set lookup_crls function and to retrieve
5739 X509_STORE from X509_STORE_CTX.
5743 * Print out deprecated issuer and subject unique ID fields in
5748 * Extend OCSP I/O functions so they can be used for simple general purpose
5749 HTTP as well as OCSP. New wrapper function which can be used to download
5750 CRLs using the OCSP API.
5754 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5758 * `SSL_CONF*` functions. These provide a common framework for application
5759 configuration using configuration files or command lines.
5763 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5764 message callback and prints the results. Needs compile time option
5765 "enable-ssl-trace". New options to s_client and s_server to enable
5770 * New ctrl and macro to retrieve supported points extensions.
5771 Print out extension in s_server and s_client.
5775 * New functions to retrieve certificate signature and signature
5780 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5785 * New Suite B modes for TLS code. These use and enforce the requirements
5786 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5787 only use Suite B curves. The Suite B modes can be set by using the
5788 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5792 * New chain verification flags for Suite B levels of security. Check
5793 algorithms are acceptable when flags are set in X509_verify_cert.
5797 * Make tls1_check_chain return a set of flags indicating checks passed
5798 by a certificate chain. Add additional tests to handle client
5799 certificates: checks for matching certificate type and issuer name
5804 * If an attempt is made to use a signature algorithm not in the peer
5805 preference list abort the handshake. If client has no suitable
5806 signature algorithms in response to a certificate request do not
5807 use the certificate.
5811 * If server EC tmp key is not in client preference list abort handshake.
5815 * Add support for certificate stores in CERT structure. This makes it
5816 possible to have different stores per SSL structure or one store in
5817 the parent SSL_CTX. Include distinct stores for certificate chain
5818 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5819 to build and store a certificate chain in CERT structure: returning
5820 an error if the chain cannot be built: this will allow applications
5821 to test if a chain is correctly configured.
5823 Note: if the CERT based stores are not set then the parent SSL_CTX
5824 store is used to retain compatibility with existing behaviour.
5828 * New function ssl_set_client_disabled to set a ciphersuite disabled
5829 mask based on the current session, check mask when sending client
5830 hello and checking the requested ciphersuite.
5834 * New ctrls to retrieve and set certificate types in a certificate
5835 request message. Print out received values in s_client. If certificate
5836 types is not set with custom values set sensible values based on
5837 supported signature algorithms.
5841 * Support for distinct client and server supported signature algorithms.
5845 * Add certificate callback. If set this is called whenever a certificate
5846 is required by client or server. An application can decide which
5847 certificate chain to present based on arbitrary criteria: for example
5848 supported signature algorithms. Add very simple example to s_server.
5849 This fixes many of the problems and restrictions of the existing client
5850 certificate callback: for example you can now clear an existing
5851 certificate and specify the whole chain.
5855 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5856 the certificate can be used for (if anything). Set valid_flags field
5857 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5858 to have similar checks in it.
5860 Add new "cert_flags" field to CERT structure and include a "strict mode".
5861 This enforces some TLS certificate requirements (such as only permitting
5862 certificate signature algorithms contained in the supported algorithms
5863 extension) which some implementations ignore: this option should be used
5864 with caution as it could cause interoperability issues.
5868 * Update and tidy signature algorithm extension processing. Work out
5869 shared signature algorithms based on preferences and peer algorithms
5870 and print them out in s_client and s_server. Abort handshake if no
5871 shared signature algorithms.
5875 * Add new functions to allow customised supported signature algorithms
5876 for SSL and SSL_CTX structures. Add options to s_client and s_server
5881 * New function SSL_certs_clear() to delete all references to certificates
5882 from an SSL structure. Before this once a certificate had been added
5883 it couldn't be removed.
5887 * Integrate hostname, email address and IP address checking with certificate
5888 verification. New verify options supporting checking in openssl utility.
5892 * Fixes and wildcard matching support to hostname and email checking
5893 functions. Add manual page.
5895 *Florian Weimer (Red Hat Product Security Team)*
5897 * New functions to check a hostname email or IP address against a
5898 certificate. Add options x509 utility to print results of checks against
5903 * Fix OCSP checking.
5905 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5907 * Initial experimental support for explicitly trusted non-root CAs.
5908 OpenSSL still tries to build a complete chain to a root but if an
5909 intermediate CA has a trust setting included that is used. The first
5910 setting is used: whether to trust (e.g., -addtrust option to the x509
5915 * Add -trusted_first option which attempts to find certificates in the
5916 trusted store even if an untrusted chain is also supplied.
5920 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5921 platform support for Linux and Android.
5925 * Support for linux-x32, ILP32 environment in x86_64 framework.
5929 * Experimental multi-implementation support for FIPS capable OpenSSL.
5930 When in FIPS mode the approved implementations are used as normal,
5931 when not in FIPS mode the internal unapproved versions are used instead.
5932 This means that the FIPS capable OpenSSL isn't forced to use the
5933 (often lower performance) FIPS implementations outside FIPS mode.
5937 * Transparently support X9.42 DH parameters when calling
5938 PEM_read_bio_DHparameters. This means existing applications can handle
5939 the new parameter format automatically.
5943 * Initial experimental support for X9.42 DH parameter format: mainly
5944 to support use of 'q' parameter for RFC5114 parameters.
5948 * Add DH parameters from RFC5114 including test data to dhtest.
5952 * Support for automatic EC temporary key parameter selection. If enabled
5953 the most preferred EC parameters are automatically used instead of
5954 hardcoded fixed parameters. Now a server just has to call:
5955 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5956 support ECDH and use the most appropriate parameters.
5960 * Enhance and tidy EC curve and point format TLS extension code. Use
5961 static structures instead of allocation if default values are used.
5962 New ctrls to set curves we wish to support and to retrieve shared curves.
5963 Print out shared curves in s_server. New options to s_server and s_client
5964 to set list of supported curves.
5968 * New ctrls to retrieve supported signature algorithms and
5969 supported curve values as an array of NIDs. Extend openssl utility
5970 to print out received values.
5974 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5975 between NIDs and the more common NIST names such as "P-256". Enhance
5976 ecparam utility and ECC method to recognise the NIST names for curves.
5980 * Enhance SSL/TLS certificate chain handling to support different
5981 chains for each certificate instead of one chain in the parent SSL_CTX.
5985 * Support for fixed DH ciphersuite client authentication: where both
5986 server and client use DH certificates with common parameters.
5990 * Support for fixed DH ciphersuites: those requiring DH server
5995 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5997 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5998 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5999 X509_CINF_get_signature were reverted post internal team review.
6004 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6006 * OCSP Status Request extension unbounded memory growth
6008 A malicious client can send an excessively large OCSP Status Request
6009 extension. If that client continually requests renegotiation, sending a
6010 large OCSP Status Request extension each time, then there will be unbounded
6011 memory growth on the server. This will eventually lead to a Denial Of
6012 Service attack through memory exhaustion. Servers with a default
6013 configuration are vulnerable even if they do not support OCSP. Builds using
6014 the "no-ocsp" build time option are not affected.
6016 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6021 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6024 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6030 * OOB write in MDC2_Update()
6032 An overflow can occur in MDC2_Update() either if called directly or
6033 through the EVP_DigestUpdate() function using MDC2. If an attacker
6034 is able to supply very large amounts of input data after a previous
6035 call to EVP_EncryptUpdate() with a partial block then a length check
6036 can overflow resulting in a heap corruption.
6038 The amount of data needed is comparable to SIZE_MAX which is impractical
6041 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6046 * Malformed SHA512 ticket DoS
6048 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6049 DoS attack where a malformed ticket will result in an OOB read which will
6052 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6053 a custom server callback and ticket lookup mechanism.
6055 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6060 * OOB write in BN_bn2dec()
6062 The function BN_bn2dec() does not check the return value of BN_div_word().
6063 This can cause an OOB write if an application uses this function with an
6064 overly large BIGNUM. This could be a problem if an overly large certificate
6065 or CRL is printed out from an untrusted source. TLS is not affected because
6066 record limits will reject an oversized certificate before it is parsed.
6068 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6073 * OOB read in TS_OBJ_print_bio()
6075 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6076 the total length the OID text representation would use and not the amount
6077 of data written. This will result in OOB reads when large OIDs are
6080 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6085 * Pointer arithmetic undefined behaviour
6087 Avoid some undefined pointer arithmetic
6089 A common idiom in the codebase is to check limits in the following manner:
6092 Where "p" points to some malloc'd data of SIZE bytes and
6095 "len" here could be from some externally supplied data (e.g. from a TLS
6098 The rules of C pointer arithmetic are such that "p + len" is only well
6099 defined where len <= SIZE. Therefore the above idiom is actually
6100 undefined behaviour.
6102 For example this could cause problems if some malloc implementation
6103 provides an address for "p" such that "p + len" actually overflows for
6104 values of len that are too big and therefore p + len < limit.
6106 This issue was reported to OpenSSL by Guido Vranken
6111 * Constant time flag not preserved in DSA signing
6113 Operations in the DSA signing algorithm should run in constant time in
6114 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6115 implementation means that a non-constant time codepath is followed for
6116 certain operations. This has been demonstrated through a cache-timing
6117 attack to be sufficient for an attacker to recover the private DSA key.
6119 This issue was reported by César Pereida (Aalto University), Billy Brumley
6120 (Tampere University of Technology), and Yuval Yarom (The University of
6121 Adelaide and NICTA).
6126 * DTLS buffered message DoS
6128 In a DTLS connection where handshake messages are delivered out-of-order
6129 those messages that OpenSSL is not yet ready to process will be buffered
6130 for later use. Under certain circumstances, a flaw in the logic means that
6131 those messages do not get removed from the buffer even though the handshake
6132 has been completed. An attacker could force up to approx. 15 messages to
6133 remain in the buffer when they are no longer required. These messages will
6134 be cleared when the DTLS connection is closed. The default maximum size for
6135 a message is 100k. Therefore the attacker could force an additional 1500k
6136 to be consumed per connection. By opening many simulataneous connections an
6137 attacker could cause a DoS attack through memory exhaustion.
6139 This issue was reported to OpenSSL by Quan Luo.
6144 * DTLS replay protection DoS
6146 A flaw in the DTLS replay attack protection mechanism means that records
6147 that arrive for future epochs update the replay protection "window" before
6148 the MAC for the record has been validated. This could be exploited by an
6149 attacker by sending a record for the next epoch (which does not have to
6150 decrypt or have a valid MAC), with a very large sequence number. This means
6151 that all subsequent legitimate packets are dropped causing a denial of
6152 service for a specific DTLS connection.
6154 This issue was reported to OpenSSL by the OCAP audit team.
6159 * Certificate message OOB reads
6161 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6162 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6163 theoretical DoS risk but this has not been observed in practice on common
6166 The messages affected are client certificate, client certificate request
6167 and server certificate. As a result the attack can only be performed
6168 against a client or a server which enables client authentication.
6170 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6175 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6177 * Prevent padding oracle in AES-NI CBC MAC check
6179 A MITM attacker can use a padding oracle attack to decrypt traffic
6180 when the connection uses an AES CBC cipher and the server support
6183 This issue was introduced as part of the fix for Lucky 13 padding
6184 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6185 constant time by making sure that always the same bytes are read and
6186 compared against either the MAC or padding bytes. But it no longer
6187 checked that there was enough data to have both the MAC and padding
6190 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6195 * Fix EVP_EncodeUpdate overflow
6197 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6198 Base64 encoding of binary data. If an attacker is able to supply very large
6199 amounts of input data then a length check can overflow resulting in a heap
6202 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6203 the `PEM_write_bio*` family of functions. These are mainly used within the
6204 OpenSSL command line applications, so any application which processes data
6205 from an untrusted source and outputs it as a PEM file should be considered
6206 vulnerable to this issue. User applications that call these APIs directly
6207 with large amounts of untrusted data may also be vulnerable.
6209 This issue was reported by Guido Vranken.
6214 * Fix EVP_EncryptUpdate overflow
6216 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6217 is able to supply very large amounts of input data after a previous call to
6218 EVP_EncryptUpdate() with a partial block then a length check can overflow
6219 resulting in a heap corruption. Following an analysis of all OpenSSL
6220 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6221 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6222 the first called function after an EVP_EncryptInit(), and therefore that
6223 specific call must be safe. The second form is where the length passed to
6224 EVP_EncryptUpdate() can be seen from the code to be some small value and
6225 therefore there is no possibility of an overflow. Since all instances are
6226 one of these two forms, it is believed that there can be no overflows in
6227 internal code due to this problem. It should be noted that
6228 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6229 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6230 of these calls have also been analysed too and it is believed there are no
6231 instances in internal usage where an overflow could occur.
6233 This issue was reported by Guido Vranken.
6238 * Prevent ASN.1 BIO excessive memory allocation
6240 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6241 a short invalid encoding can casuse allocation of large amounts of memory
6242 potentially consuming excessive resources or exhausting memory.
6244 Any application parsing untrusted data through d2i BIO functions is
6245 affected. The memory based functions such as d2i_X509() are *not* affected.
6246 Since the memory based functions are used by the TLS library, TLS
6247 applications are not affected.
6249 This issue was reported by Brian Carpenter.
6256 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6257 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6258 in arbitrary stack data being returned in the buffer.
6260 This issue was reported by Guido Vranken.
6265 * Modify behavior of ALPN to invoke callback after SNI/servername
6266 callback, such that updates to the SSL_CTX affect ALPN.
6270 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6275 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6276 methods are enabled and ssl2 is disabled the methods return NULL.
6280 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6282 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6283 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6284 provide any "EXPORT" or "LOW" strength ciphers.
6288 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6289 is by default disabled at build-time. Builds that are not configured with
6290 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6291 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6292 will need to explicitly call either of:
6294 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6296 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6298 as appropriate. Even if either of those is used, or the application
6299 explicitly uses the version-specific SSLv2_method() or its client and
6300 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6301 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6302 ciphers, and SSLv2 56-bit DES are no longer available.
6307 * Fix a double-free in DSA code
6309 A double free bug was discovered when OpenSSL parses malformed DSA private
6310 keys and could lead to a DoS attack or memory corruption for applications
6311 that receive DSA private keys from untrusted sources. This scenario is
6314 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6320 * Disable SRP fake user seed to address a server memory leak.
6322 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6324 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6325 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6326 was changed to ignore the "fake user" SRP seed, even if the seed
6329 Users should use SRP_VBASE_get1_by_user instead. Note that in
6330 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6331 also that even though configuring the SRP seed attempts to hide
6332 invalid usernames by continuing the handshake with fake
6333 credentials, this behaviour is not constant time and no strong
6334 guarantees are made that the handshake is indistinguishable from
6335 that of a valid user.
6340 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6342 In the BN_hex2bn function the number of hex digits is calculated using an
6343 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6344 large values of `i` this can result in `bn_expand` not allocating any
6345 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6346 field as NULL leading to a subsequent NULL ptr deref. For very large values
6347 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6348 In this case memory is allocated to the internal BIGNUM data field, but it
6349 is insufficiently sized leading to heap corruption. A similar issue exists
6350 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6351 is ever called by user applications with very large untrusted hex/dec data.
6352 This is anticipated to be a rare occurrence.
6354 All OpenSSL internal usage of these functions use data that is not expected
6355 to be untrusted, e.g. config file data or application command line
6356 arguments. If user developed applications generate config file data based
6357 on untrusted data then it is possible that this could also lead to security
6358 consequences. This is also anticipated to be rare.
6360 This issue was reported to OpenSSL by Guido Vranken.
6365 * Fix memory issues in `BIO_*printf` functions
6367 The internal `fmtstr` function used in processing a "%s" format string in
6368 the `BIO_*printf` functions could overflow while calculating the length of a
6369 string and cause an OOB read when printing very long strings.
6371 Additionally the internal `doapr_outch` function can attempt to write to an
6372 OOB memory location (at an offset from the NULL pointer) in the event of a
6373 memory allocation failure. In 1.0.2 and below this could be caused where
6374 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6375 could be in processing a very long "%s" format string. Memory leaks can
6378 The first issue may mask the second issue dependent on compiler behaviour.
6379 These problems could enable attacks where large amounts of untrusted data
6380 is passed to the `BIO_*printf` functions. If applications use these functions
6381 in this way then they could be vulnerable. OpenSSL itself uses these
6382 functions when printing out human-readable dumps of ASN.1 data. Therefore
6383 applications that print this data could be vulnerable if the data is from
6384 untrusted sources. OpenSSL command line applications could also be
6385 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6386 as command line arguments.
6388 Libssl is not considered directly vulnerable. Additionally certificates etc
6389 received via remote connections via libssl are also unlikely to be able to
6390 trigger these issues because of message size limits enforced within libssl.
6392 This issue was reported to OpenSSL Guido Vranken.
6397 * Side channel attack on modular exponentiation
6399 A side-channel attack was found which makes use of cache-bank conflicts on
6400 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6401 of RSA keys. The ability to exploit this issue is limited as it relies on
6402 an attacker who has control of code in a thread running on the same
6403 hyper-threaded core as the victim thread which is performing decryptions.
6405 This issue was reported to OpenSSL by Yuval Yarom, The University of
6406 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6407 Nadia Heninger, University of Pennsylvania with more information at
6408 <http://cachebleed.info>.
6413 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6414 if no keysize is specified with default_bits. This fixes an
6415 omission in an earlier change that changed all RSA/DSA key generation
6416 commands to use 2048 bits by default.
6420 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6422 * Protection for DH small subgroup attacks
6424 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6425 switched on by default and cannot be disabled. This could have some
6430 * SSLv2 doesn't block disabled ciphers
6432 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6433 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6434 been disabled, provided that the SSLv2 protocol was not also disabled via
6437 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6438 and Sebastian Schinzel.
6443 * Reject DH handshakes with parameters shorter than 1024 bits.
6447 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6449 * Certificate verify crash with missing PSS parameter
6451 The signature verification routines will crash with a NULL pointer
6452 dereference if presented with an ASN.1 signature using the RSA PSS
6453 algorithm and absent mask generation function parameter. Since these
6454 routines are used to verify certificate signature algorithms this can be
6455 used to crash any certificate verification operation and exploited in a
6456 DoS attack. Any application which performs certificate verification is
6457 vulnerable including OpenSSL clients and servers which enable client
6460 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6465 * X509_ATTRIBUTE memory leak
6467 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6468 memory. This structure is used by the PKCS#7 and CMS routines so any
6469 application which reads PKCS#7 or CMS data from untrusted sources is
6470 affected. SSL/TLS is not affected.
6472 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6478 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6479 This changes the decoding behaviour for some invalid messages,
6480 though the change is mostly in the more lenient direction, and
6481 legacy behaviour is preserved as much as possible.
6485 * In DSA_generate_parameters_ex, if the provided seed is too short,
6486 use a random seed, as already documented.
6488 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6490 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6492 * Alternate chains certificate forgery
6494 During certificate verfification, OpenSSL will attempt to find an
6495 alternative certificate chain if the first attempt to build such a chain
6496 fails. An error in the implementation of this logic can mean that an
6497 attacker could cause certain checks on untrusted certificates to be
6498 bypassed, such as the CA flag, enabling them to use a valid leaf
6499 certificate to act as a CA and "issue" an invalid certificate.
6501 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6507 * Race condition handling PSK identify hint
6509 If PSK identity hints are received by a multi-threaded client then
6510 the values are wrongly updated in the parent SSL_CTX structure. This can
6511 result in a race condition potentially leading to a double free of the
6517 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6519 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6520 incompatibility in the handling of HMAC. The previous ABI has now been
6523 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6525 * Malformed ECParameters causes infinite loop
6527 When processing an ECParameters structure OpenSSL enters an infinite loop
6528 if the curve specified is over a specially malformed binary polynomial
6531 This can be used to perform denial of service against any
6532 system which processes public keys, certificate requests or
6533 certificates. This includes TLS clients and TLS servers with
6534 client authentication enabled.
6536 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6541 * Exploitable out-of-bounds read in X509_cmp_time
6543 X509_cmp_time does not properly check the length of the ASN1_TIME
6544 string and can read a few bytes out of bounds. In addition,
6545 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6548 An attacker can use this to craft malformed certificates and CRLs of
6549 various sizes and potentially cause a segmentation fault, resulting in
6550 a DoS on applications that verify certificates or CRLs. TLS clients
6551 that verify CRLs are affected. TLS clients and servers with client
6552 authentication enabled may be affected if they use custom verification
6555 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6556 independently by Hanno Böck.
6561 * PKCS7 crash with missing EnvelopedContent
6563 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6564 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6565 with missing content and trigger a NULL pointer dereference on parsing.
6567 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6568 structures from untrusted sources are affected. OpenSSL clients and
6569 servers are not affected.
6571 This issue was reported to OpenSSL by Michal Zalewski (Google).
6576 * CMS verify infinite loop with unknown hash function
6578 When verifying a signedData message the CMS code can enter an infinite loop
6579 if presented with an unknown hash function OID. This can be used to perform
6580 denial of service against any system which verifies signedData messages using
6582 This issue was reported to OpenSSL by Johannes Bauer.
6587 * Race condition handling NewSessionTicket
6589 If a NewSessionTicket is received by a multi-threaded client when attempting to
6590 reuse a previous ticket then a race condition can occur potentially leading to
6591 a double free of the ticket data.
6596 * Reject DH handshakes with parameters shorter than 768 bits.
6598 *Kurt Roeckx and Emilia Kasper*
6600 * dhparam: generate 2048-bit parameters by default.
6602 *Kurt Roeckx and Emilia Kasper*
6604 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6606 * Segmentation fault in ASN1_TYPE_cmp fix
6608 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6609 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6610 certificate signature algorithm consistency this can be used to crash any
6611 certificate verification operation and exploited in a DoS attack. Any
6612 application which performs certificate verification is vulnerable including
6613 OpenSSL clients and servers which enable client authentication.
6618 * ASN.1 structure reuse memory corruption fix
6620 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6621 memory corruption via an invalid write. Such reuse is and has been
6622 strongly discouraged and is believed to be rare.
6624 Applications that parse structures containing CHOICE or ANY DEFINED BY
6625 components may be affected. Certificate parsing (d2i_X509 and related
6626 functions) are however not affected. OpenSSL clients and servers are
6632 * PKCS7 NULL pointer dereferences fix
6634 The PKCS#7 parsing code does not handle missing outer ContentInfo
6635 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6636 missing content and trigger a NULL pointer dereference on parsing.
6638 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6639 otherwise parse PKCS#7 structures from untrusted sources are
6640 affected. OpenSSL clients and servers are not affected.
6642 This issue was reported to OpenSSL by Michal Zalewski (Google).
6647 * DoS via reachable assert in SSLv2 servers fix
6649 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6650 servers that both support SSLv2 and enable export cipher suites by sending
6651 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6653 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6654 (OpenSSL development team).
6659 * Use After Free following d2i_ECPrivatekey error fix
6661 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6662 could cause a use after free condition. This, in turn, could cause a double
6663 free in several private key parsing functions (such as d2i_PrivateKey
6664 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6665 for applications that receive EC private keys from untrusted
6666 sources. This scenario is considered rare.
6668 This issue was discovered by the BoringSSL project and fixed in their
6674 * X509_to_X509_REQ NULL pointer deref fix
6676 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6677 the certificate key is invalid. This function is rarely used in practice.
6679 This issue was discovered by Brian Carpenter.
6684 * Removed the export ciphers from the DEFAULT ciphers
6688 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6690 * Build fixes for the Windows and OpenVMS platforms
6692 *Matt Caswell and Richard Levitte*
6694 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6696 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6697 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6698 dereference. This could lead to a Denial Of Service attack. Thanks to
6699 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6704 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6705 dtls1_buffer_record function under certain conditions. In particular this
6706 could occur if an attacker sent repeated DTLS records with the same
6707 sequence number but for the next epoch. The memory leak could be exploited
6708 by an attacker in a Denial of Service attack through memory exhaustion.
6709 Thanks to Chris Mueller for reporting this issue.
6714 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6715 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6716 method would be set to NULL which could later result in a NULL pointer
6717 dereference. Thanks to Frank Schmirler for reporting this issue.
6722 * Abort handshake if server key exchange message is omitted for ephemeral
6725 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6726 reporting this issue.
6731 * Remove non-export ephemeral RSA code on client and server. This code
6732 violated the TLS standard by allowing the use of temporary RSA keys in
6733 non-export ciphersuites and could be used by a server to effectively
6734 downgrade the RSA key length used to a value smaller than the server
6735 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6736 INRIA or reporting this issue.
6741 * Fixed issue where DH client certificates are accepted without verification.
6742 An OpenSSL server will accept a DH certificate for client authentication
6743 without the certificate verify message. This effectively allows a client to
6744 authenticate without the use of a private key. This only affects servers
6745 which trust a client certificate authority which issues certificates
6746 containing DH keys: these are extremely rare and hardly ever encountered.
6747 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6753 * Ensure that the session ID context of an SSL is updated when its
6754 SSL_CTX is updated via SSL_set_SSL_CTX.
6756 The session ID context is typically set from the parent SSL_CTX,
6757 and can vary with the CTX.
6761 * Fix various certificate fingerprint issues.
6763 By using non-DER or invalid encodings outside the signed portion of a
6764 certificate the fingerprint can be changed without breaking the signature.
6765 Although no details of the signed portion of the certificate can be changed
6766 this can cause problems with some applications: e.g. those using the
6767 certificate fingerprint for blacklists.
6769 1. Reject signatures with non zero unused bits.
6771 If the BIT STRING containing the signature has non zero unused bits reject
6772 the signature. All current signature algorithms require zero unused bits.
6774 2. Check certificate algorithm consistency.
6776 Check the AlgorithmIdentifier inside TBS matches the one in the
6777 certificate signature. NB: this will result in signature failure
6778 errors for some broken certificates.
6780 Thanks to Konrad Kraszewski from Google for reporting this issue.
6782 3. Check DSA/ECDSA signatures use DER.
6784 Re-encode DSA/ECDSA signatures and compare with the original received
6785 signature. Return an error if there is a mismatch.
6787 This will reject various cases including garbage after signature
6788 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6789 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6790 (negative or with leading zeroes).
6792 Further analysis was conducted and fixes were developed by Stephen Henson
6793 of the OpenSSL core team.
6799 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6800 results on some platforms, including x86_64. This bug occurs at random
6801 with a very low probability, and is not known to be exploitable in any
6802 way, though its exact impact is difficult to determine. Thanks to Pieter
6803 Wuille (Blockstream) who reported this issue and also suggested an initial
6804 fix. Further analysis was conducted by the OpenSSL development team and
6805 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6806 the OpenSSL core team.
6811 * Do not resume sessions on the server if the negotiated protocol
6812 version does not match the session's version. Resuming with a different
6813 version, while not strictly forbidden by the RFC, is of questionable
6814 sanity and breaks all known clients.
6816 *David Benjamin, Emilia Käsper*
6818 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6819 early CCS messages during renegotiation. (Note that because
6820 renegotiation is encrypted, this early CCS was not exploitable.)
6824 * Tighten client-side session ticket handling during renegotiation:
6825 ensure that the client only accepts a session ticket if the server sends
6826 the extension anew in the ServerHello. Previously, a TLS client would
6827 reuse the old extension state and thus accept a session ticket if one was
6828 announced in the initial ServerHello.
6830 Similarly, ensure that the client requires a session ticket if one
6831 was advertised in the ServerHello. Previously, a TLS client would
6832 ignore a missing NewSessionTicket message.
6836 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6840 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6841 sends a carefully crafted handshake message, to cause OpenSSL to fail
6842 to free up to 64k of memory causing a memory leak. This could be
6843 exploited in a Denial Of Service attack. This issue affects OpenSSL
6844 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6845 whether SRTP is used or configured. Implementations of OpenSSL that
6846 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6848 The fix was developed by the OpenSSL team.
6853 * Session Ticket Memory Leak.
6855 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6856 integrity of that ticket is first verified. In the event of a session
6857 ticket integrity check failing, OpenSSL will fail to free memory
6858 causing a memory leak. By sending a large number of invalid session
6859 tickets an attacker could exploit this issue in a Denial Of Service
6865 * Build option no-ssl3 is incomplete.
6867 When OpenSSL is configured with "no-ssl3" as a build option, servers
6868 could accept and complete a SSL 3.0 handshake, and clients could be
6869 configured to send them.
6872 *Akamai and the OpenSSL team*
6874 * Add support for TLS_FALLBACK_SCSV.
6875 Client applications doing fallback retries should call
6876 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6879 *Adam Langley, Bodo Moeller*
6881 * Add additional DigestInfo checks.
6883 Re-encode DigestInto in DER and check against the original when
6884 verifying RSA signature: this will reject any improperly encoded
6885 DigestInfo structures.
6887 Note: this is a precautionary measure and no attacks are currently known.
6891 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6893 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6894 SRP code can be overrun an internal buffer. Add sanity check that
6895 g, A, B < N to SRP code.
6897 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6898 Group for discovering this issue.
6903 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6904 TLS 1.0 instead of higher protocol versions when the ClientHello message
6905 is badly fragmented. This allows a man-in-the-middle attacker to force a
6906 downgrade to TLS 1.0 even if both the server and the client support a
6907 higher protocol version, by modifying the client's TLS records.
6909 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6910 researching this issue.
6915 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6916 to a denial of service attack. A malicious server can crash the client
6917 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6918 ciphersuite and sending carefully crafted handshake messages.
6920 Thanks to Felix Gröbert (Google) for discovering and researching this
6926 * By sending carefully crafted DTLS packets an attacker could cause openssl
6927 to leak memory. This can be exploited through a Denial of Service attack.
6928 Thanks to Adam Langley for discovering and researching this issue.
6933 * An attacker can force openssl to consume large amounts of memory whilst
6934 processing DTLS handshake messages. This can be exploited through a
6935 Denial of Service attack.
6936 Thanks to Adam Langley for discovering and researching this issue.
6941 * An attacker can force an error condition which causes openssl to crash
6942 whilst processing DTLS packets due to memory being freed twice. This
6943 can be exploited through a Denial of Service attack.
6944 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6950 * If a multithreaded client connects to a malicious server using a resumed
6951 session and the server sends an ec point format extension it could write
6952 up to 255 bytes to freed memory.
6954 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6960 * A malicious server can crash an OpenSSL client with a null pointer
6961 dereference (read) by specifying an SRP ciphersuite even though it was not
6962 properly negotiated with the client. This can be exploited through a
6963 Denial of Service attack.
6965 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6966 discovering and researching this issue.
6971 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6972 X509_name_oneline, X509_name_print_ex et al. to leak some information
6973 from the stack. Applications may be affected if they echo pretty printing
6974 output to the attacker.
6976 Thanks to Ivan Fratric (Google) for discovering this issue.
6979 *Emilia Käsper, and Steve Henson*
6981 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6982 for corner cases. (Certain input points at infinity could lead to
6983 bogus results, with non-infinity inputs mapped to infinity too.)
6987 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6989 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6990 handshake can force the use of weak keying material in OpenSSL
6991 SSL/TLS clients and servers.
6993 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6994 researching this issue. ([CVE-2014-0224])
6996 *KIKUCHI Masashi, Steve Henson*
6998 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6999 OpenSSL DTLS client the code can be made to recurse eventually crashing
7002 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7005 *Imre Rad, Steve Henson*
7007 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7008 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7009 client or server. This is potentially exploitable to run arbitrary
7010 code on a vulnerable client or server.
7012 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7014 *Jüri Aedla, Steve Henson*
7016 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7017 are subject to a denial of service attack.
7019 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7020 this issue. ([CVE-2014-3470])
7022 *Felix Gröbert, Ivan Fratric, Steve Henson*
7024 * Harmonize version and its documentation. -f flag is used to display
7027 *mancha <mancha1@zoho.com>*
7029 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7030 in i2d_ECPrivateKey.
7032 *mancha <mancha1@zoho.com>*
7034 * Fix some double frees. These are not thought to be exploitable.
7036 *mancha <mancha1@zoho.com>*
7038 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7040 * A missing bounds check in the handling of the TLS heartbeat extension
7041 can be used to reveal up to 64k of memory to a connected client or
7044 Thanks for Neel Mehta of Google Security for discovering this bug and to
7045 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7046 preparing the fix ([CVE-2014-0160])
7048 *Adam Langley, Bodo Moeller*
7050 * Fix for the attack described in the paper "Recovering OpenSSL
7051 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7052 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7053 <http://eprint.iacr.org/2014/140>
7055 Thanks to Yuval Yarom and Naomi Benger for discovering this
7056 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7058 *Yuval Yarom and Naomi Benger*
7060 * TLS pad extension: draft-agl-tls-padding-03
7062 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7063 TLS client Hello record length value would otherwise be > 255 and
7064 less that 512 pad with a dummy extension containing zeroes so it
7065 is at least 512 bytes long.
7067 *Adam Langley, Steve Henson*
7069 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7071 * Fix for TLS record tampering bug. A carefully crafted invalid
7072 handshake could crash OpenSSL with a NULL pointer exception.
7073 Thanks to Anton Johansson for reporting this issues.
7076 * Keep original DTLS digest and encryption contexts in retransmission
7077 structures so we can use the previous session parameters if they need
7078 to be resent. ([CVE-2013-6450])
7082 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7083 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7084 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7085 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7086 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7087 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7089 *Rob Stradling, Adam Langley*
7091 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7093 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7094 supporting platforms or when small records were transferred.
7096 *Andy Polyakov, Steve Henson*
7098 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7100 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7102 This addresses the flaw in CBC record processing discovered by
7103 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7104 at: <http://www.isg.rhul.ac.uk/tls/>
7106 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7107 Security Group at Royal Holloway, University of London
7108 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7109 Emilia Käsper for the initial patch.
7112 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7114 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7115 ciphersuites which can be exploited in a denial of service attack.
7116 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7117 and detecting this bug and to Wolfgang Ettlinger
7118 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7123 * Return an error when checking OCSP signatures when key is NULL.
7124 This fixes a DoS attack. ([CVE-2013-0166])
7128 * Make openssl verify return errors.
7130 *Chris Palmer <palmer@google.com> and Ben Laurie*
7132 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7133 the right response is stapled. Also change SSL_get_certificate()
7134 so it returns the certificate actually sent.
7135 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7137 *Rob Stradling <rob.stradling@comodo.com>*
7139 * Fix possible deadlock when decoding public keys.
7143 * Don't use TLS 1.0 record version number in initial client hello
7148 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7150 * Sanity check record length before skipping explicit IV in TLS
7151 1.2, 1.1 and DTLS to fix DoS attack.
7153 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7154 fuzzing as a service testing platform.
7159 * Initialise tkeylen properly when encrypting CMS messages.
7160 Thanks to Solar Designer of Openwall for reporting this issue.
7164 * In FIPS mode don't try to use composite ciphers as they are not
7169 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7171 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7172 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7173 mean any application compiled against OpenSSL 1.0.0 headers setting
7174 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7175 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7176 0x10000000L Any application which was previously compiled against
7177 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7178 will need to be recompiled as a result. Letting be results in
7179 inability to disable specifically TLS 1.1 and in client context,
7180 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7184 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7185 disable just protocol X, but all protocols above X *if* there are
7186 protocols *below* X still enabled. In more practical terms it means
7187 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7188 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7189 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7194 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7196 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7197 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7198 in CRYPTO_realloc_clean.
7200 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7201 issue and to Adam Langley <agl@chromium.org> for fixing it.
7204 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7206 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7210 * Workarounds for some broken servers that "hang" if a client hello
7211 record length exceeds 255 bytes.
7213 1. Do not use record version number > TLS 1.0 in initial client
7214 hello: some (but not all) hanging servers will now work.
7215 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7216 the number of ciphers sent in the client hello. This should be
7217 set to an even number, such as 50, for example by passing:
7218 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7219 Most broken servers should now work.
7220 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7221 TLS 1.2 client support entirely.
7225 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7229 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7231 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7232 STRING form instead of a DigestInfo.
7236 * The format used for MDC2 RSA signatures is inconsistent between EVP
7237 and the RSA_sign/RSA_verify functions. This was made more apparent when
7238 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7239 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7240 the correct format in RSA_verify so both forms transparently work.
7244 * Some servers which support TLS 1.0 can choke if we initially indicate
7245 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7246 encrypted premaster secret. As a workaround use the maximum permitted
7247 client version in client hello, this should keep such servers happy
7248 and still work with previous versions of OpenSSL.
7252 * Add support for TLS/DTLS heartbeats.
7254 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7256 * Add support for SCTP.
7258 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7260 * Improved PRNG seeding for VOS.
7262 *Paul Green <Paul.Green@stratus.com>*
7264 * Extensive assembler packs updates, most notably:
7266 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7267 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7268 - x86_64: bit-sliced AES implementation;
7269 - ARM: NEON support, contemporary platforms optimizations;
7270 - s390x: z196 support;
7271 - `*`: GHASH and GF(2^m) multiplication implementations;
7275 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7276 (removal of unnecessary code)
7278 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7280 * Add TLS key material exporter from RFC 5705.
7284 * Add DTLS-SRTP negotiation from RFC 5764.
7288 * Add Next Protocol Negotiation,
7289 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7290 disabled with a no-npn flag to config or Configure. Code donated
7293 *Adam Langley <agl@google.com> and Ben Laurie*
7295 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7296 NIST-P256, NIST-P521, with constant-time single point multiplication on
7297 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7298 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7299 Code made available under Apache License version 2.0.
7301 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7302 line to include this in your build of OpenSSL, and run "make depend" (or
7303 "make update"). This enables the following EC_METHODs:
7305 EC_GFp_nistp224_method()
7306 EC_GFp_nistp256_method()
7307 EC_GFp_nistp521_method()
7309 EC_GROUP_new_by_curve_name() will automatically use these (while
7310 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7313 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7315 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7316 all platforms. Move ssize_t definition from e_os.h to the public
7317 header file e_os2.h as it now appears in public header file cms.h
7321 * New -sigopt option to the ca, req and x509 utilities. Additional
7322 signature parameters can be passed using this option and in
7327 * Add RSA PSS signing function. This will generate and set the
7328 appropriate AlgorithmIdentifiers for PSS based on those in the
7329 corresponding EVP_MD_CTX structure. No application support yet.
7333 * Support for companion algorithm specific ASN1 signing routines.
7334 New function ASN1_item_sign_ctx() signs a pre-initialised
7335 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7336 the appropriate parameters.
7340 * Add new algorithm specific ASN1 verification initialisation function
7341 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7342 handling will be the same no matter what EVP_PKEY_METHOD is used.
7343 Add a PSS handler to support verification of PSS signatures: checked
7344 against a number of sample certificates.
7348 * Add signature printing for PSS. Add PSS OIDs.
7350 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7352 * Add algorithm specific signature printing. An individual ASN1 method
7353 can now print out signatures instead of the standard hex dump.
7355 More complex signatures (e.g. PSS) can print out more meaningful
7356 information. Include DSA version that prints out the signature
7361 * Password based recipient info support for CMS library: implementing
7366 * Split password based encryption into PBES2 and PBKDF2 functions. This
7367 neatly separates the code into cipher and PBE sections and is required
7368 for some algorithms that split PBES2 into separate pieces (such as
7369 password based CMS).
7373 * Session-handling fixes:
7374 - Fix handling of connections that are resuming with a session ID,
7375 but also support Session Tickets.
7376 - Fix a bug that suppressed issuing of a new ticket if the client
7377 presented a ticket with an expired session.
7378 - Try to set the ticket lifetime hint to something reasonable.
7379 - Make tickets shorter by excluding irrelevant information.
7380 - On the client side, don't ignore renewed tickets.
7382 *Adam Langley, Bodo Moeller (Google)*
7384 * Fix PSK session representation.
7388 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7390 This work was sponsored by Intel.
7394 * Add GCM support to TLS library. Some custom code is needed to split
7395 the IV between the fixed (from PRF) and explicit (from TLS record)
7396 portions. This adds all GCM ciphersuites supported by RFC5288 and
7397 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7398 add a special AESGCM string for GCM only.
7402 * Expand range of ctrls for AES GCM. Permit setting invocation
7403 field on decrypt and retrieval of invocation field only on encrypt.
7407 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7408 As required by RFC5289 these ciphersuites cannot be used if for
7409 versions of TLS earlier than 1.2.
7413 * For FIPS capable OpenSSL interpret a NULL default public key method
7414 as unset and return the appropriate default but do *not* set the default.
7415 This means we can return the appropriate method in applications that
7416 switch between FIPS and non-FIPS modes.
7420 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7421 ENGINE is used then we cannot handle that in the FIPS module so we
7422 keep original code iff non-FIPS operations are allowed.
7426 * Add -attime option to openssl utilities.
7428 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7430 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7434 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7435 FIPS EC methods unconditionally for now.
7439 * New build option no-ec2m to disable characteristic 2 code.
7443 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7444 all cases can be covered as some introduce binary incompatibilities.
7448 * Redirect RSA operations to FIPS module including keygen,
7449 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7453 * Add similar low level API blocking to ciphers.
7457 * Low level digest APIs are not approved in FIPS mode: any attempt
7458 to use these will cause a fatal error. Applications that *really* want
7459 to use them can use the `private_*` version instead.
7463 * Redirect cipher operations to FIPS module for FIPS builds.
7467 * Redirect digest operations to FIPS module for FIPS builds.
7471 * Update build system to add "fips" flag which will link in fipscanister.o
7472 for static and shared library builds embedding a signature if needed.
7476 * Output TLS supported curves in preference order instead of numerical
7477 order. This is currently hardcoded for the highest order curves first.
7478 This should be configurable so applications can judge speed vs strength.
7482 * Add TLS v1.2 server support for client authentication.
7486 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7491 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7492 FIPS modules versions.
7496 * Add TLS v1.2 client side support for client authentication. Keep cache
7497 of handshake records longer as we don't know the hash algorithm to use
7498 until after the certificate request message is received.
7502 * Initial TLS v1.2 client support. Add a default signature algorithms
7503 extension including all the algorithms we support. Parse new signature
7504 format in client key exchange. Relax some ECC signing restrictions for
7505 TLS v1.2 as indicated in RFC5246.
7509 * Add server support for TLS v1.2 signature algorithms extension. Switch
7510 to new signature format when needed using client digest preference.
7511 All server ciphersuites should now work correctly in TLS v1.2. No client
7512 support yet and no support for client certificates.
7516 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7517 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7518 ciphersuites. At present only RSA key exchange ciphersuites work with
7519 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7520 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7521 and version checking.
7525 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7526 with this defined it will not be affected by any changes to ssl internal
7527 structures. Add several utility functions to allow openssl application
7528 to work with OPENSSL_NO_SSL_INTERN defined.
7532 * A long standing patch to add support for SRP from EdelWeb (Peter
7533 Sylvester and Christophe Renou) was integrated.
7534 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7535 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7538 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7542 * Permit abbreviated handshakes when renegotiating using the function
7543 SSL_renegotiate_abbreviated().
7545 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7547 * Add call to ENGINE_register_all_complete() to
7548 ENGINE_load_builtin_engines(), so some implementations get used
7549 automatically instead of needing explicit application support.
7553 * Add support for TLS key exporter as described in RFC5705.
7555 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7557 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7558 a few changes are required:
7560 Add SSL_OP_NO_TLSv1_1 flag.
7561 Add TLSv1_1 methods.
7562 Update version checking logic to handle version 1.1.
7563 Add explicit IV handling (ported from DTLS code).
7564 Add command line options to s_client/s_server.
7571 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7573 * X509_ATTRIBUTE memory leak
7575 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7576 memory. This structure is used by the PKCS#7 and CMS routines so any
7577 application which reads PKCS#7 or CMS data from untrusted sources is
7578 affected. SSL/TLS is not affected.
7580 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7586 * Race condition handling PSK identify hint
7588 If PSK identity hints are received by a multi-threaded client then
7589 the values are wrongly updated in the parent SSL_CTX structure. This can
7590 result in a race condition potentially leading to a double free of the
7596 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7598 * Malformed ECParameters causes infinite loop
7600 When processing an ECParameters structure OpenSSL enters an infinite loop
7601 if the curve specified is over a specially malformed binary polynomial
7604 This can be used to perform denial of service against any
7605 system which processes public keys, certificate requests or
7606 certificates. This includes TLS clients and TLS servers with
7607 client authentication enabled.
7609 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7614 * Exploitable out-of-bounds read in X509_cmp_time
7616 X509_cmp_time does not properly check the length of the ASN1_TIME
7617 string and can read a few bytes out of bounds. In addition,
7618 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7621 An attacker can use this to craft malformed certificates and CRLs of
7622 various sizes and potentially cause a segmentation fault, resulting in
7623 a DoS on applications that verify certificates or CRLs. TLS clients
7624 that verify CRLs are affected. TLS clients and servers with client
7625 authentication enabled may be affected if they use custom verification
7628 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7629 independently by Hanno Böck.
7634 * PKCS7 crash with missing EnvelopedContent
7636 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7637 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7638 with missing content and trigger a NULL pointer dereference on parsing.
7640 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7641 structures from untrusted sources are affected. OpenSSL clients and
7642 servers are not affected.
7644 This issue was reported to OpenSSL by Michal Zalewski (Google).
7649 * CMS verify infinite loop with unknown hash function
7651 When verifying a signedData message the CMS code can enter an infinite loop
7652 if presented with an unknown hash function OID. This can be used to perform
7653 denial of service against any system which verifies signedData messages using
7655 This issue was reported to OpenSSL by Johannes Bauer.
7660 * Race condition handling NewSessionTicket
7662 If a NewSessionTicket is received by a multi-threaded client when attempting to
7663 reuse a previous ticket then a race condition can occur potentially leading to
7664 a double free of the ticket data.
7669 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7671 * Segmentation fault in ASN1_TYPE_cmp fix
7673 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7674 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7675 certificate signature algorithm consistency this can be used to crash any
7676 certificate verification operation and exploited in a DoS attack. Any
7677 application which performs certificate verification is vulnerable including
7678 OpenSSL clients and servers which enable client authentication.
7683 * ASN.1 structure reuse memory corruption fix
7685 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7686 memory corruption via an invalid write. Such reuse is and has been
7687 strongly discouraged and is believed to be rare.
7689 Applications that parse structures containing CHOICE or ANY DEFINED BY
7690 components may be affected. Certificate parsing (d2i_X509 and related
7691 functions) are however not affected. OpenSSL clients and servers are
7697 * PKCS7 NULL pointer dereferences fix
7699 The PKCS#7 parsing code does not handle missing outer ContentInfo
7700 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7701 missing content and trigger a NULL pointer dereference on parsing.
7703 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7704 otherwise parse PKCS#7 structures from untrusted sources are
7705 affected. OpenSSL clients and servers are not affected.
7707 This issue was reported to OpenSSL by Michal Zalewski (Google).
7712 * DoS via reachable assert in SSLv2 servers fix
7714 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7715 servers that both support SSLv2 and enable export cipher suites by sending
7716 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7718 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7719 (OpenSSL development team).
7724 * Use After Free following d2i_ECPrivatekey error fix
7726 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7727 could cause a use after free condition. This, in turn, could cause a double
7728 free in several private key parsing functions (such as d2i_PrivateKey
7729 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7730 for applications that receive EC private keys from untrusted
7731 sources. This scenario is considered rare.
7733 This issue was discovered by the BoringSSL project and fixed in their
7739 * X509_to_X509_REQ NULL pointer deref fix
7741 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7742 the certificate key is invalid. This function is rarely used in practice.
7744 This issue was discovered by Brian Carpenter.
7749 * Removed the export ciphers from the DEFAULT ciphers
7753 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7755 * Build fixes for the Windows and OpenVMS platforms
7757 *Matt Caswell and Richard Levitte*
7759 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7761 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7762 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7763 dereference. This could lead to a Denial Of Service attack. Thanks to
7764 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7769 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7770 dtls1_buffer_record function under certain conditions. In particular this
7771 could occur if an attacker sent repeated DTLS records with the same
7772 sequence number but for the next epoch. The memory leak could be exploited
7773 by an attacker in a Denial of Service attack through memory exhaustion.
7774 Thanks to Chris Mueller for reporting this issue.
7779 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7780 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7781 method would be set to NULL which could later result in a NULL pointer
7782 dereference. Thanks to Frank Schmirler for reporting this issue.
7787 * Abort handshake if server key exchange message is omitted for ephemeral
7790 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7791 reporting this issue.
7796 * Remove non-export ephemeral RSA code on client and server. This code
7797 violated the TLS standard by allowing the use of temporary RSA keys in
7798 non-export ciphersuites and could be used by a server to effectively
7799 downgrade the RSA key length used to a value smaller than the server
7800 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7801 INRIA or reporting this issue.
7806 * Fixed issue where DH client certificates are accepted without verification.
7807 An OpenSSL server will accept a DH certificate for client authentication
7808 without the certificate verify message. This effectively allows a client to
7809 authenticate without the use of a private key. This only affects servers
7810 which trust a client certificate authority which issues certificates
7811 containing DH keys: these are extremely rare and hardly ever encountered.
7812 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7818 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7819 results on some platforms, including x86_64. This bug occurs at random
7820 with a very low probability, and is not known to be exploitable in any
7821 way, though its exact impact is difficult to determine. Thanks to Pieter
7822 Wuille (Blockstream) who reported this issue and also suggested an initial
7823 fix. Further analysis was conducted by the OpenSSL development team and
7824 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7825 the OpenSSL core team.
7830 * Fix various certificate fingerprint issues.
7832 By using non-DER or invalid encodings outside the signed portion of a
7833 certificate the fingerprint can be changed without breaking the signature.
7834 Although no details of the signed portion of the certificate can be changed
7835 this can cause problems with some applications: e.g. those using the
7836 certificate fingerprint for blacklists.
7838 1. Reject signatures with non zero unused bits.
7840 If the BIT STRING containing the signature has non zero unused bits reject
7841 the signature. All current signature algorithms require zero unused bits.
7843 2. Check certificate algorithm consistency.
7845 Check the AlgorithmIdentifier inside TBS matches the one in the
7846 certificate signature. NB: this will result in signature failure
7847 errors for some broken certificates.
7849 Thanks to Konrad Kraszewski from Google for reporting this issue.
7851 3. Check DSA/ECDSA signatures use DER.
7853 Reencode DSA/ECDSA signatures and compare with the original received
7854 signature. Return an error if there is a mismatch.
7856 This will reject various cases including garbage after signature
7857 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7858 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7859 (negative or with leading zeroes).
7861 Further analysis was conducted and fixes were developed by Stephen Henson
7862 of the OpenSSL core team.
7868 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7870 * Session Ticket Memory Leak.
7872 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7873 integrity of that ticket is first verified. In the event of a session
7874 ticket integrity check failing, OpenSSL will fail to free memory
7875 causing a memory leak. By sending a large number of invalid session
7876 tickets an attacker could exploit this issue in a Denial Of Service
7882 * Build option no-ssl3 is incomplete.
7884 When OpenSSL is configured with "no-ssl3" as a build option, servers
7885 could accept and complete a SSL 3.0 handshake, and clients could be
7886 configured to send them.
7889 *Akamai and the OpenSSL team*
7891 * Add support for TLS_FALLBACK_SCSV.
7892 Client applications doing fallback retries should call
7893 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7896 *Adam Langley, Bodo Moeller*
7898 * Add additional DigestInfo checks.
7900 Reencode DigestInto in DER and check against the original when
7901 verifying RSA signature: this will reject any improperly encoded
7902 DigestInfo structures.
7904 Note: this is a precautionary measure and no attacks are currently known.
7908 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7910 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7911 to a denial of service attack. A malicious server can crash the client
7912 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7913 ciphersuite and sending carefully crafted handshake messages.
7915 Thanks to Felix Gröbert (Google) for discovering and researching this
7921 * By sending carefully crafted DTLS packets an attacker could cause openssl
7922 to leak memory. This can be exploited through a Denial of Service attack.
7923 Thanks to Adam Langley for discovering and researching this issue.
7928 * An attacker can force openssl to consume large amounts of memory whilst
7929 processing DTLS handshake messages. This can be exploited through a
7930 Denial of Service attack.
7931 Thanks to Adam Langley for discovering and researching this issue.
7936 * An attacker can force an error condition which causes openssl to crash
7937 whilst processing DTLS packets due to memory being freed twice. This
7938 can be exploited through a Denial of Service attack.
7939 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7945 * If a multithreaded client connects to a malicious server using a resumed
7946 session and the server sends an ec point format extension it could write
7947 up to 255 bytes to freed memory.
7949 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7955 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7956 X509_name_oneline, X509_name_print_ex et al. to leak some information
7957 from the stack. Applications may be affected if they echo pretty printing
7958 output to the attacker.
7960 Thanks to Ivan Fratric (Google) for discovering this issue.
7963 *Emilia Käsper, and Steve Henson*
7965 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7966 for corner cases. (Certain input points at infinity could lead to
7967 bogus results, with non-infinity inputs mapped to infinity too.)
7971 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7973 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7974 handshake can force the use of weak keying material in OpenSSL
7975 SSL/TLS clients and servers.
7977 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7978 researching this issue. ([CVE-2014-0224])
7980 *KIKUCHI Masashi, Steve Henson*
7982 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7983 OpenSSL DTLS client the code can be made to recurse eventually crashing
7986 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7989 *Imre Rad, Steve Henson*
7991 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7992 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7993 client or server. This is potentially exploitable to run arbitrary
7994 code on a vulnerable client or server.
7996 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7998 *Jüri Aedla, Steve Henson*
8000 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8001 are subject to a denial of service attack.
8003 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8004 this issue. ([CVE-2014-3470])
8006 *Felix Gröbert, Ivan Fratric, Steve Henson*
8008 * Harmonize version and its documentation. -f flag is used to display
8011 *mancha <mancha1@zoho.com>*
8013 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8014 in i2d_ECPrivateKey.
8016 *mancha <mancha1@zoho.com>*
8018 * Fix some double frees. These are not thought to be exploitable.
8020 *mancha <mancha1@zoho.com>*
8022 * Fix for the attack described in the paper "Recovering OpenSSL
8023 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8024 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8025 <http://eprint.iacr.org/2014/140>
8027 Thanks to Yuval Yarom and Naomi Benger for discovering this
8028 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8030 *Yuval Yarom and Naomi Benger*
8032 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8034 * Keep original DTLS digest and encryption contexts in retransmission
8035 structures so we can use the previous session parameters if they need
8036 to be resent. ([CVE-2013-6450])
8040 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8041 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8042 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8043 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8044 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8045 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8047 *Rob Stradling, Adam Langley*
8049 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8051 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8053 This addresses the flaw in CBC record processing discovered by
8054 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8055 at: <http://www.isg.rhul.ac.uk/tls/>
8057 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8058 Security Group at Royal Holloway, University of London
8059 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8060 Emilia Käsper for the initial patch.
8063 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8065 * Return an error when checking OCSP signatures when key is NULL.
8066 This fixes a DoS attack. ([CVE-2013-0166])
8070 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8071 the right response is stapled. Also change SSL_get_certificate()
8072 so it returns the certificate actually sent.
8073 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8074 (This is a backport)
8076 *Rob Stradling <rob.stradling@comodo.com>*
8078 * Fix possible deadlock when decoding public keys.
8082 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8084 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8087 * Sanity check record length before skipping explicit IV in DTLS
8090 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8091 fuzzing as a service testing platform.
8096 * Initialise tkeylen properly when encrypting CMS messages.
8097 Thanks to Solar Designer of Openwall for reporting this issue.
8101 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8103 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8104 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8105 in CRYPTO_realloc_clean.
8107 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8108 issue and to Adam Langley <agl@chromium.org> for fixing it.
8111 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8113 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8115 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8116 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8117 content decryption and always return the same error. Note: this attack
8118 needs on average 2^20 messages so it only affects automated senders. The
8119 old behaviour can be re-enabled in the CMS code by setting the
8120 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8121 an MMA defence is not necessary.
8122 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8123 this issue. ([CVE-2012-0884])
8127 * Fix CVE-2011-4619: make sure we really are receiving a
8128 client hello before rejecting multiple SGC restarts. Thanks to
8129 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8133 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8135 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8136 Thanks to Antonio Martin, Enterprise Secure Access Research and
8137 Development, Cisco Systems, Inc. for discovering this bug and
8138 preparing a fix. ([CVE-2012-0050])
8142 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8144 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8145 of the Vaudenay padding oracle attack on CBC mode encryption
8146 which enables an efficient plaintext recovery attack against
8147 the OpenSSL implementation of DTLS. Their attack exploits timing
8148 differences arising during decryption processing. A research
8149 paper describing this attack can be found at:
8150 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8151 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8152 Security Group at Royal Holloway, University of London
8153 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8154 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8155 for preparing the fix. ([CVE-2011-4108])
8157 *Robin Seggelmann, Michael Tuexen*
8159 * Clear bytes used for block padding of SSL 3.0 records.
8162 *Adam Langley (Google)*
8164 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8165 Kadianakis <desnacked@gmail.com> for discovering this issue and
8166 Adam Langley for preparing the fix. ([CVE-2011-4619])
8168 *Adam Langley (Google)*
8170 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8172 *Andrey Kulikov <amdeich@gmail.com>*
8174 * Prevent malformed RFC3779 data triggering an assertion failure.
8175 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8176 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8178 *Rob Austein <sra@hactrn.net>*
8180 * Improved PRNG seeding for VOS.
8182 *Paul Green <Paul.Green@stratus.com>*
8184 * Fix ssl_ciph.c set-up race.
8186 *Adam Langley (Google)*
8188 * Fix spurious failures in ecdsatest.c.
8190 *Emilia Käsper (Google)*
8192 * Fix the BIO_f_buffer() implementation (which was mixing different
8193 interpretations of the `..._len` fields).
8195 *Adam Langley (Google)*
8197 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8198 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8199 threads won't reuse the same blinding coefficients.
8201 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8202 lock to call BN_BLINDING_invert_ex, and avoids one use of
8203 BN_BLINDING_update for each BN_BLINDING structure (previously,
8204 the last update always remained unused).
8206 *Emilia Käsper (Google)*
8208 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8210 *Bob Buckholz (Google)*
8212 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8214 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8215 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8217 *Kaspar Brand <ossl@velox.ch>*
8219 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8220 for multi-threaded use of ECDH. ([CVE-2011-3210])
8222 *Adam Langley (Google)*
8224 * Fix x509_name_ex_d2i memory leak on bad inputs.
8228 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8229 signature public key algorithm by using OID xref utilities instead.
8230 Before this you could only use some ECC ciphersuites with SHA1 only.
8234 * Add protection against ECDSA timing attacks as mentioned in the paper
8235 by Billy Bob Brumley and Nicola Tuveri, see:
8236 <http://eprint.iacr.org/2011/232.pdf>
8238 *Billy Bob Brumley and Nicola Tuveri*
8240 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8242 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8244 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8246 * Fix bug in string printing code: if *any* escaping is enabled we must
8247 escape the escape character (backslash) or the resulting string is
8252 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8254 * Disable code workaround for ancient and obsolete Netscape browsers
8255 and servers: an attacker can use it in a ciphersuite downgrade attack.
8256 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8260 * Fixed J-PAKE implementation error, originally discovered by
8261 Sebastien Martini, further info and confirmation from Stefan
8262 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8266 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8268 * Fix extension code to avoid race conditions which can result in a buffer
8269 overrun vulnerability: resumed sessions must not be modified as they can
8270 be shared by multiple threads. CVE-2010-3864
8274 * Fix WIN32 build system to correctly link an ENGINE directory into
8279 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8281 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8284 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8286 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8288 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8289 context. The operation can be customised via the ctrl mechanism in
8290 case ENGINEs want to include additional functionality.
8294 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8298 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8299 output hashes compatible with older versions of OpenSSL.
8301 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8303 * Fix compression algorithm handling: if resuming a session use the
8304 compression algorithm of the resumed session instead of determining
8305 it from client hello again. Don't allow server to change algorithm.
8309 * Add load_crls() function to commands tidying load_certs() too. Add option
8310 to verify utility to allow additional CRLs to be included.
8314 * Update OCSP request code to permit adding custom headers to the request:
8315 some responders need this.
8319 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8322 *Julia Lawall <julia@diku.dk>*
8324 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8325 needlessly dereferenced structures, used obsolete functions and
8326 didn't handle all updated verify codes correctly.
8330 * Disable MD2 in the default configuration.
8334 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8335 indicate the initial BIO being pushed or popped. This makes it possible
8336 to determine whether the BIO is the one explicitly called or as a result
8337 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8338 it handles reference counts correctly and doesn't zero out the I/O bio
8339 when it is not being explicitly popped. WARNING: applications which
8340 included workarounds for the old buggy behaviour will need to be modified
8341 or they could free up already freed BIOs.
8345 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8346 renaming to all platforms (within the 0.9.8 branch, this was
8347 done conditionally on Netware platforms to avoid a name clash).
8349 *Guenter <lists@gknw.net>*
8351 * Add ECDHE and PSK support to DTLS.
8353 *Michael Tuexen <tuexen@fh-muenster.de>*
8355 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8360 * Add "missing" function EVP_MD_flags() (without this the only way to
8361 retrieve a digest flags is by accessing the structure directly. Update
8362 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8363 or cipher is registered as in the "from" argument. Print out all
8364 registered digests in the dgst usage message instead of manually
8365 attempting to work them out.
8369 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8370 this allows the use of compression and extensions. Change default cipher
8371 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8372 by default unless an application cipher string requests it.
8376 * Alter match criteria in PKCS12_parse(). It used to try to use local
8377 key ids to find matching certificates and keys but some PKCS#12 files
8378 don't follow the (somewhat unwritten) rules and this strategy fails.
8379 Now just gather all certificates together and the first private key
8380 then look for the first certificate that matches the key.
8384 * Support use of registered digest and cipher names for dgst and cipher
8385 commands instead of having to add each one as a special case. So now
8392 openssl dgst -sha256 foo
8394 and this works for ENGINE based algorithms too.
8398 * Update Gost ENGINE to support parameter files.
8400 *Victor B. Wagner <vitus@cryptocom.ru>*
8402 * Support GeneralizedTime in ca utility.
8404 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8406 * Enhance the hash format used for certificate directory links. The new
8407 form uses the canonical encoding (meaning equivalent names will work
8408 even if they aren't identical) and uses SHA1 instead of MD5. This form
8409 is incompatible with the older format and as a result c_rehash should
8410 be used to rebuild symbolic links.
8414 * Make PKCS#8 the default write format for private keys, replacing the
8415 traditional format. This form is standardised, more secure and doesn't
8416 include an implicit MD5 dependency.
8420 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8421 committed to OpenSSL should pass this lot as a minimum.
8425 * Add session ticket override functionality for use by EAP-FAST.
8427 *Jouni Malinen <j@w1.fi>*
8429 * Modify HMAC functions to return a value. Since these can be implemented
8430 in an ENGINE errors can occur.
8434 * Type-checked OBJ_bsearch_ex.
8438 * Type-checked OBJ_bsearch. Also some constification necessitated
8439 by type-checking. Still to come: TXT_DB, bsearch(?),
8440 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8445 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8446 seconds to a tm structure directly, instead of going through OS
8447 specific date routines. This avoids any issues with OS routines such
8448 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8449 and X509_time_adj_ex() to cover the extended range. The existing
8450 X509_time_adj() is still usable and will no longer have any date issues.
8454 * Delta CRL support. New use deltas option which will attempt to locate
8455 and search any appropriate delta CRLs available.
8457 This work was sponsored by Google.
8461 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8462 code and add additional score elements. Validate alternate CRL paths
8463 as part of the CRL checking and indicate a new error "CRL path validation
8464 error" in this case. Applications wanting additional details can use
8465 the verify callback and check the new "parent" field. If this is not
8466 NULL CRL path validation is taking place. Existing applications won't
8467 see this because it requires extended CRL support which is off by
8470 This work was sponsored by Google.
8474 * Support for freshest CRL extension.
8476 This work was sponsored by Google.
8480 * Initial indirect CRL support. Currently only supported in the CRLs
8481 passed directly and not via lookup. Process certificate issuer
8482 CRL entry extension and lookup CRL entries by bother issuer name
8483 and serial number. Check and process CRL issuer entry in IDP extension.
8485 This work was sponsored by Google.
8489 * Add support for distinct certificate and CRL paths. The CRL issuer
8490 certificate is validated separately in this case. Only enabled if
8491 an extended CRL support flag is set: this flag will enable additional
8492 CRL functionality in future.
8494 This work was sponsored by Google.
8498 * Add support for policy mappings extension.
8500 This work was sponsored by Google.
8504 * Fixes to pathlength constraint, self issued certificate handling,
8505 policy processing to align with RFC3280 and PKITS tests.
8507 This work was sponsored by Google.
8511 * Support for name constraints certificate extension. DN, email, DNS
8512 and URI types are currently supported.
8514 This work was sponsored by Google.
8518 * To cater for systems that provide a pointer-based thread ID rather
8519 than numeric, deprecate the current numeric thread ID mechanism and
8520 replace it with a structure and associated callback type. This
8521 mechanism allows a numeric "hash" to be extracted from a thread ID in
8522 either case, and on platforms where pointers are larger than 'long',
8523 mixing is done to help ensure the numeric 'hash' is usable even if it
8524 can't be guaranteed unique. The default mechanism is to use "&errno"
8525 as a pointer-based thread ID to distinguish between threads.
8527 Applications that want to provide their own thread IDs should now use
8528 CRYPTO_THREADID_set_callback() to register a callback that will call
8529 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8531 Note that ERR_remove_state() is now deprecated, because it is tied
8532 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8533 to free the current thread's error state should be replaced by
8534 ERR_remove_thread_state(NULL).
8536 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8537 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8538 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8539 application was previously providing a numeric thread callback that
8540 was inappropriate for distinguishing threads, then uniqueness might
8541 have been obtained with &errno that happened immediately in the
8542 intermediate development versions of OpenSSL; this is no longer the
8543 case, the numeric thread callback will now override the automatic use
8546 *Geoff Thorpe, with help from Bodo Moeller*
8548 * Initial support for different CRL issuing certificates. This covers a
8549 simple case where the self issued certificates in the chain exist and
8550 the real CRL issuer is higher in the existing chain.
8552 This work was sponsored by Google.
8556 * Removed effectively defunct crypto/store from the build.
8560 * Revamp of STACK to provide stronger type-checking. Still to come:
8561 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8562 ASN1_STRING, CONF_VALUE.
8566 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8567 RAM on SSL connections. This option can save about 34k per idle SSL.
8571 * Revamp of LHASH to provide stronger type-checking. Still to come:
8572 STACK, TXT_DB, bsearch, qsort.
8576 * Initial support for Cryptographic Message Syntax (aka CMS) based
8577 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8578 support for data, signedData, compressedData, digestedData and
8579 encryptedData, envelopedData types included. Scripts to check against
8580 RFC4134 examples draft and interop and consistency checks of many
8581 content types and variants.
8585 * Add options to enc utility to support use of zlib compression BIO.
8589 * Extend mk1mf to support importing of options and assembly language
8590 files from Configure script, currently only included in VC-WIN32.
8591 The assembly language rules can now optionally generate the source
8592 files from the associated perl scripts.
8596 * Implement remaining functionality needed to support GOST ciphersuites.
8597 Interop testing has been performed using CryptoPro implementations.
8599 *Victor B. Wagner <vitus@cryptocom.ru>*
8601 * s390x assembler pack.
8605 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8610 * Implement Opaque PRF Input TLS extension as specified in
8611 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8612 official specification yet and no extension type assignment by
8613 IANA exists, this extension (for now) will have to be explicitly
8614 enabled when building OpenSSL by providing the extension number
8615 to use. For example, specify an option
8617 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8619 to the "config" or "Configure" script to enable the extension,
8620 assuming extension number 0x9527 (which is a completely arbitrary
8621 and unofficial assignment based on the MD5 hash of the Internet
8622 Draft). Note that by doing so, you potentially lose
8623 interoperability with other TLS implementations since these might
8624 be using the same extension number for other purposes.
8626 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8627 opaque PRF input value to use in the handshake. This will create
8628 an internal copy of the length-'len' string at 'src', and will
8629 return non-zero for success.
8631 To get more control and flexibility, provide a callback function
8634 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8635 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8639 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8642 Callback function 'cb' will be called in handshakes, and is
8643 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8644 Argument 'arg' is for application purposes (the value as given to
8645 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8646 be provided to the callback function). The callback function
8647 has to return non-zero to report success: usually 1 to use opaque
8648 PRF input just if possible, or 2 to enforce use of the opaque PRF
8649 input. In the latter case, the library will abort the handshake
8650 if opaque PRF input is not successfully negotiated.
8652 Arguments 'peerinput' and 'len' given to the callback function
8653 will always be NULL and 0 in the case of a client. A server will
8654 see the client's opaque PRF input through these variables if
8655 available (NULL and 0 otherwise). Note that if the server
8656 provides an opaque PRF input, the length must be the same as the
8657 length of the client's opaque PRF input.
8659 Note that the callback function will only be called when creating
8660 a new session (session resumption can resume whatever was
8661 previously negotiated), and will not be called in SSL 2.0
8662 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8663 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8664 for applications that need to enforce opaque PRF input.
8668 * Update ssl code to support digests other than SHA1+MD5 for handshake
8671 *Victor B. Wagner <vitus@cryptocom.ru>*
8673 * Add RFC4507 support to OpenSSL. This includes the corrections in
8674 RFC4507bis. The encrypted ticket format is an encrypted encoded
8675 SSL_SESSION structure, that way new session features are automatically
8678 If a client application caches session in an SSL_SESSION structure
8679 support is transparent because tickets are now stored in the encoded
8682 The SSL_CTX structure automatically generates keys for ticket
8683 protection in servers so again support should be possible
8684 with no application modification.
8686 If a client or server wishes to disable RFC4507 support then the option
8687 SSL_OP_NO_TICKET can be set.
8689 Add a TLS extension debugging callback to allow the contents of any client
8690 or server extensions to be examined.
8692 This work was sponsored by Google.
8696 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8697 OpenSSL should now compile cleanly on gcc 4.2
8699 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8701 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8702 support including streaming MAC support: this is required for GOST
8703 ciphersuite support.
8705 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8707 * Add option -stream to use PKCS#7 streaming in smime utility. New
8708 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8709 to output in BER and PEM format.
8713 * Experimental support for use of HMAC via EVP_PKEY interface. This
8714 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8715 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8716 ENGINE support for HMAC keys which are unextractable. New -mac and
8717 -macopt options to dgst utility.
8721 * New option -sigopt to dgst utility. Update dgst to use
8722 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8723 alternative signing parameters such as X9.31 or PSS in the dgst
8728 * Change ssl_cipher_apply_rule(), the internal function that does
8729 the work each time a ciphersuite string requests enabling
8730 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8731 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8732 the order of disabled ciphersuites such that those ciphersuites
8733 that most recently went from enabled to disabled not only stay
8734 in order with respect to each other, but also have higher priority
8735 than other disabled ciphersuites the next time ciphersuites are
8738 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8739 the same ciphersuites as with "HIGH" alone, but in a specific
8740 order where the PSK ciphersuites come first (since they are the
8741 most recently disabled ciphersuites when "HIGH" is parsed).
8743 Also, change ssl_create_cipher_list() (using this new
8744 functionality) such that between otherwise identical
8745 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8750 * Change ssl_create_cipher_list() so that it automatically
8751 arranges the ciphersuites in reasonable order before starting
8752 to process the rule string. Thus, the definition for "DEFAULT"
8753 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8754 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8755 This makes it much easier to arrive at a reasonable default order
8756 in applications for which anonymous ciphers are OK (meaning
8757 that you can't actually use DEFAULT).
8759 *Bodo Moeller; suggested by Victor Duchovni*
8761 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8762 processing) into multiple integers instead of setting
8763 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8764 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8765 (These masks as well as the individual bit definitions are hidden
8766 away into the non-exported interface ssl/ssl_locl.h, so this
8767 change to the definition of the SSL_CIPHER structure shouldn't
8768 affect applications.) This give us more bits for each of these
8769 categories, so there is no longer a need to coagulate AES128 and
8770 AES256 into a single algorithm bit, and to coagulate Camellia128
8771 and Camellia256 into a single algorithm bit, which has led to all
8774 Thus, among other things, the kludge introduced in 0.9.7m and
8775 0.9.8e for masking out AES256 independently of AES128 or masking
8776 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8778 With the change, we also introduce new ciphersuite aliases that
8779 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8784 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8785 Use the leftmost N bytes of the signature input if the input is
8786 larger than the prime q (with N being the size in bytes of q).
8790 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8791 it yet and it is largely untested.
8795 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8799 * Initial incomplete changes to avoid need for function casts in OpenSSL
8800 some compilers (gcc 4.2 and later) reject their use. Safestack is
8801 reimplemented. Update ASN1 to avoid use of legacy functions.
8805 * Win32/64 targets are linked with Winsock2.
8809 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8810 to external functions. This can be used to increase CRL handling
8811 efficiency especially when CRLs are very large by (for example) storing
8812 the CRL revoked certificates in a database.
8816 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8817 new CRLs added to a directory can be used. New command line option
8818 -verify_return_error to s_client and s_server. This causes real errors
8819 to be returned by the verify callback instead of carrying on no matter
8820 what. This reflects the way a "real world" verify callback would behave.
8824 * GOST engine, supporting several GOST algorithms and public key formats.
8825 Kindly donated by Cryptocom.
8829 * Partial support for Issuing Distribution Point CRL extension. CRLs
8830 partitioned by DP are handled but no indirect CRL or reason partitioning
8831 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8832 selected via a scoring technique which handles IDP and AKID in CRLs.
8836 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8837 will ultimately be used for all verify operations: this will remove the
8838 X509_STORE dependency on certificate verification and allow alternative
8839 lookup methods. X509_STORE based implementations of these two callbacks.
8843 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8844 Modify get_crl() to find a valid (unexpired) CRL if possible.
8848 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8849 this would be called X509_CRL_cmp() but that name is already used by
8850 a function that just compares CRL issuer names. Cache several CRL
8851 extensions in X509_CRL structure and cache CRLDP in X509.
8855 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8856 this maps equivalent X509_NAME structures into a consistent structure.
8857 Name comparison can then be performed rapidly using memcmp().
8861 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8866 * Allow digests to supply their own micalg string for S/MIME type using
8867 the ctrl EVP_MD_CTRL_MICALG.
8871 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8872 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8873 ctrl. It can then customise the structure before and/or after signing
8878 * New function OBJ_add_sigid() to allow application defined signature OIDs
8879 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8880 to free up any added signature OIDs.
8884 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8885 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8886 digest and cipher tables. New options added to openssl utility:
8887 list-message-digest-algorithms and list-cipher-algorithms.
8891 * Change the array representation of binary polynomials: the list
8892 of degrees of non-zero coefficients is now terminated with -1.
8893 Previously it was terminated with 0, which was also part of the
8894 value; thus, the array representation was not applicable to
8895 polynomials where t^0 has coefficient zero. This change makes
8896 the array representation useful in a more general context.
8900 * Various modifications and fixes to SSL/TLS cipher string
8901 handling. For ECC, the code now distinguishes between fixed ECDH
8902 with RSA certificates on the one hand and with ECDSA certificates
8903 on the other hand, since these are separate ciphersuites. The
8904 unused code for Fortezza ciphersuites has been removed.
8906 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8907 (not "ECDHE"). For consistency with the code for DH
8908 certificates, use of ECDH certificates is now considered ECDH
8909 authentication, not RSA or ECDSA authentication (the latter is
8910 merely the CA's signing algorithm and not actively used in the
8913 The temporary ciphersuite alias "ECCdraft" is no longer
8914 available, and ECC ciphersuites are no longer excluded from "ALL"
8915 and "DEFAULT". The following aliases now exist for RFC 4492
8916 ciphersuites, most of these by analogy with the DH case:
8918 kECDHr - ECDH cert, signed with RSA
8919 kECDHe - ECDH cert, signed with ECDSA
8920 kECDH - ECDH cert (signed with either RSA or ECDSA)
8921 kEECDH - ephemeral ECDH
8922 ECDH - ECDH cert or ephemeral ECDH
8928 AECDH - anonymous ECDH
8929 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8933 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8934 Use correct micalg parameters depending on digest(s) in signed message.
8938 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8939 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8943 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8944 an engine to register a method. Add ENGINE lookups for methods and
8945 functional reference processing.
8949 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8950 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8955 * New -resign option to smime utility. This adds one or more signers
8956 to an existing PKCS#7 signedData structure. Also -md option to use an
8957 alternative message digest algorithm for signing.
8961 * Tidy up PKCS#7 routines and add new functions to make it easier to
8962 create PKCS7 structures containing multiple signers. Update smime
8963 application to support multiple signers.
8967 * New -macalg option to pkcs12 utility to allow setting of an alternative
8972 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8973 Reorganize PBE internals to lookup from a static table using NIDs,
8974 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8975 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8976 PRF which will be automatically used with PBES2.
8980 * Replace the algorithm specific calls to generate keys in "req" with the
8985 * Update PKCS#7 enveloped data routines to use new API. This is now
8986 supported by any public key method supporting the encrypt operation. A
8987 ctrl is added to allow the public key algorithm to examine or modify
8988 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8993 * Add a ctrl to asn1 method to allow a public key algorithm to express
8994 a default digest type to use. In most cases this will be SHA1 but some
8995 algorithms (such as GOST) need to specify an alternative digest. The
8996 return value indicates how strong the preference is 1 means optional and
8997 2 is mandatory (that is it is the only supported type). Modify
8998 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8999 use the default md. Update openssl utilities to use the default digest
9000 type for signing if it is not explicitly indicated.
9004 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9005 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9006 signing method from the key type. This effectively removes the link
9007 between digests and public key types.
9011 * Add an OID cross reference table and utility functions. Its purpose is to
9012 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9013 rsaEncryption. This will allow some of the algorithm specific hackery
9014 needed to use the correct OID to be removed.
9018 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9019 structures for PKCS7_sign(). They are now set up by the relevant public
9024 * Add provisional EC pkey method with support for ECDSA and ECDH.
9028 * Add support for key derivation (agreement) in the API, DH method and
9033 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9034 public and private key formats. As a side effect these add additional
9035 command line functionality not previously available: DSA signatures can be
9036 generated and verified using pkeyutl and DH key support and generation in
9043 *Oliver Tappe <zooey@hirschkaefer.de>*
9045 * New make target "install_html_docs" installs HTML renditions of the
9048 *Oliver Tappe <zooey@hirschkaefer.de>*
9050 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9051 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9052 support key and parameter generation and add initial key generation
9053 functionality for RSA.
9057 * Add functions for main EVP_PKEY_method operations. The undocumented
9058 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9059 `EVP_PKEY_{encrypt,decrypt}_old`.
9063 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9064 key API, doesn't do much yet.
9068 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9069 public key algorithms. New option to openssl utility:
9070 "list-public-key-algorithms" to print out info.
9074 * Implement the Supported Elliptic Curves Extension for
9075 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9079 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9080 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9084 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9085 utilities such as rsa, dsa, dsaparam etc except they process any key
9090 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9091 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9092 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9097 * Initial support for pluggable public key ASN1.
9098 De-spaghettify the public key ASN1 handling. Move public and private
9099 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9100 algorithm specific handling to a single module within the relevant
9101 algorithm directory. Add functions to allow (near) opaque processing
9102 of public and private key structures.
9106 * Implement the Supported Point Formats Extension for
9107 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9111 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9112 for the psk identity [hint] and the psk callback functions to the
9113 SSL_SESSION, SSL and SSL_CTX structure.
9116 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9120 SSL_CTX_use_psk_identity_hint
9121 SSL_get_psk_identity_hint
9122 SSL_get_psk_identity
9123 SSL_use_psk_identity_hint
9125 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9127 * Add RFC 3161 compliant time stamp request creation, response generation
9128 and response verification functionality.
9130 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9132 * Add initial support for TLS extensions, specifically for the server_name
9133 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9134 have new members for a host name. The SSL data structure has an
9135 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9136 stored in that context to allow for session resumption, even after the
9137 SSL has been switched to a new SSL_CTX in reaction to a client's
9138 server_name extension.
9140 New functions (subject to change):
9142 SSL_get_servername()
9143 SSL_get_servername_type()
9146 New CTRL codes and macros (subject to change):
9148 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9149 - SSL_CTX_set_tlsext_servername_callback()
9150 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9151 - SSL_CTX_set_tlsext_servername_arg()
9152 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9154 openssl s_client has a new '-servername ...' option.
9156 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9157 '-key2 ...', '-servername_fatal' (subject to change). This allows
9158 testing the HostName extension for a specific single host name ('-cert'
9159 and '-key' remain fallbacks for handshakes without HostName
9160 negotiation). If the unrecognized_name alert has to be sent, this by
9161 default is a warning; it becomes fatal with the '-servername_fatal'
9164 *Peter Sylvester, Remy Allais, Christophe Renou*
9166 * Whirlpool hash implementation is added.
9170 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9171 bn(64,32). Because of instruction set limitations it doesn't have
9172 any negative impact on performance. This was done mostly in order
9173 to make it possible to share assembler modules, such as bn_mul_mont
9174 implementations, between 32- and 64-bit builds without hassle.
9178 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9179 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9184 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9185 dedicated Montgomery multiplication procedure, is introduced.
9186 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9187 "64-bit" performance on certain 32-bit targets.
9191 * New option SSL_OP_NO_COMP to disable use of compression selectively
9192 in SSL structures. New SSL ctrl to set maximum send fragment size.
9193 Save memory by setting the I/O buffer sizes dynamically instead of
9194 using the maximum available value.
9198 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9199 in addition to the text details.
9203 * Very, very preliminary EXPERIMENTAL support for printing of general
9204 ASN1 structures. This currently produces rather ugly output and doesn't
9205 handle several customised structures at all.
9209 * Integrated support for PVK file format and some related formats such
9210 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9211 these in the 'rsa' and 'dsa' utilities.
9215 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9219 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9220 place for the (very old) "NETSCAPE" format certificates which are now
9221 handled using new ASN1 code equivalents.
9225 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9226 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9227 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9231 * Modify CRL distribution points extension code to print out previously
9232 unsupported fields. Enhance extension setting code to allow setting of
9237 * Add print and set support for Issuing Distribution Point CRL extension.
9241 * Change 'Configure' script to enable Camellia by default.
9248 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9250 * When rejecting SSL/TLS records due to an incorrect version number, never
9251 update s->server with a new major version number. As of
9252 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9253 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9254 the previous behavior could result in a read attempt at NULL when
9255 receiving specific incorrect SSL/TLS records once record payload
9256 protection is active. ([CVE-2010-0740])
9258 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9260 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9261 could be crashed if the relevant tables were not present (e.g. chrooted).
9263 *Tomas Hoger <thoger@redhat.com>*
9265 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9267 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9269 *Martin Olsson, Neel Mehta*
9271 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9272 accommodate for stack sorting, always a write lock!).
9276 * On some versions of WIN32 Heap32Next is very slow. This can cause
9277 excessive delays in the RAND_poll(): over a minute. As a workaround
9278 include a time check in the inner Heap32Next loop too.
9282 * The code that handled flushing of data in SSL/TLS originally used the
9283 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9284 the problem outlined in PR#1949. The fix suggested there however can
9285 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9286 of Apache). So instead simplify the code to flush unconditionally.
9287 This should be fine since flushing with no data to flush is a no op.
9291 * Handle TLS versions 2.0 and later properly and correctly use the
9292 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9293 off ancient servers have a habit of sticking around for a while...
9297 * Modify compression code so it frees up structures without using the
9298 ex_data callbacks. This works around a problem where some applications
9299 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9300 restarting) then use compression (e.g. SSL with compression) later.
9301 This results in significant per-connection memory leaks and
9302 has caused some security issues including CVE-2008-1678 and
9307 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9308 change when encrypting or decrypting.
9312 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9313 connect and renegotiate with servers which do not support RI.
9314 Until RI is more widely deployed this option is enabled by default.
9318 * Add "missing" ssl ctrls to clear options and mode.
9322 * If client attempts to renegotiate and doesn't support RI respond with
9323 a no_renegotiation alert as required by RFC5746. Some renegotiating
9324 TLS clients will continue a connection gracefully when they receive
9325 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9326 waiting for a server hello which it will never receive. Now we treat a
9327 received no_renegotiation alert as a fatal error. This is because
9328 applications requesting a renegotiation might well expect it to succeed
9329 and would have no code in place to handle the server denying it so the
9330 only safe thing to do is to terminate the connection.
9334 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9335 peer supports secure renegotiation and 0 otherwise. Print out peer
9336 renegotiation support in s_client/s_server.
9340 * Replace the highly broken and deprecated SPKAC certification method with
9341 the updated NID creation version. This should correctly handle UTF8.
9345 * Implement RFC5746. Re-enable renegotiation but require the extension
9346 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9347 turns out to be a bad idea. It has been replaced by
9348 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9349 SSL_CTX_set_options(). This is really not recommended unless you
9350 know what you are doing.
9352 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9354 * Fixes to stateless session resumption handling. Use initial_ctx when
9355 issuing and attempting to decrypt tickets in case it has changed during
9356 servername handling. Use a non-zero length session ID when attempting
9357 stateless session resumption: this makes it possible to determine if
9358 a resumption has occurred immediately after receiving server hello
9359 (several places in OpenSSL subtly assume this) instead of later in
9364 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9365 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9366 fixes for a few places where the return code is not checked
9369 *Julia Lawall <julia@diku.dk>*
9371 * Add --strict-warnings option to Configure script to include devteam
9372 warnings in other configurations.
9376 * Add support for --libdir option and LIBDIR variable in makefiles. This
9377 makes it possible to install openssl libraries in locations which
9378 have names other than "lib", for example "/usr/lib64" which some
9381 *Steve Henson, based on patch from Jeremy Utley*
9383 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9384 X690 8.9.12 and can produce some misleading textual output of OIDs.
9386 *Steve Henson, reported by Dan Kaminsky*
9388 * Delete MD2 from algorithm tables. This follows the recommendation in
9389 several standards that it is not used in new applications due to
9390 several cryptographic weaknesses. For binary compatibility reasons
9391 the MD2 API is still compiled in by default.
9395 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9400 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9401 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9404 *Guenter <lists@gknw.net>*
9406 * Fix the server certificate chain building code to use X509_verify_cert(),
9407 it used to have an ad-hoc builder which was unable to cope with anything
9408 other than a simple chain.
9410 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9412 * Don't check self signed certificate signatures in X509_verify_cert()
9413 by default (a flag can override this): it just wastes time without
9414 adding any security. As a useful side effect self signed root CAs
9415 with non-FIPS digests are now usable in FIPS mode.
9419 * In dtls1_process_out_of_seq_message() the check if the current message
9420 is already buffered was missing. For every new message was memory
9421 allocated, allowing an attacker to perform an denial of service attack
9422 with sending out of seq handshake messages until there is no memory
9423 left. Additionally every future message was buffered, even if the
9424 sequence number made no sense and would be part of another handshake.
9425 So only messages with sequence numbers less than 10 in advance will be
9426 buffered. ([CVE-2009-1378])
9428 *Robin Seggelmann, discovered by Daniel Mentz*
9430 * Records are buffered if they arrive with a future epoch to be
9431 processed after finishing the corresponding handshake. There is
9432 currently no limitation to this buffer allowing an attacker to perform
9433 a DOS attack with sending records with future epochs until there is no
9434 memory left. This patch adds the pqueue_size() function to determine
9435 the size of a buffer and limits the record buffer to 100 entries.
9438 *Robin Seggelmann, discovered by Daniel Mentz*
9440 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9441 parent structure is freed. ([CVE-2009-1379])
9445 * Handle non-blocking I/O properly in SSL_shutdown() call.
9447 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9449 * Add `2.5.4.*` OIDs
9451 *Ilya O. <vrghost@gmail.com>*
9453 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9455 * Disable renegotiation completely - this fixes a severe security
9456 problem ([CVE-2009-3555]) at the cost of breaking all
9457 renegotiation. Renegotiation can be re-enabled by setting
9458 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9459 run-time. This is really not recommended unless you know what
9464 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9466 * Don't set val to NULL when freeing up structures, it is freed up by
9467 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9468 zeroing past the valid field. ([CVE-2009-0789])
9470 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9472 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9473 checked correctly. This would allow some invalid signed attributes to
9474 appear to verify correctly. ([CVE-2009-0591])
9476 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9478 * Reject UniversalString and BMPString types with invalid lengths. This
9479 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9480 a legal length. ([CVE-2009-0590])
9484 * Set S/MIME signing as the default purpose rather than setting it
9485 unconditionally. This allows applications to override it at the store
9490 * Permit restricted recursion of ASN1 strings. This is needed in practice
9491 to handle some structures.
9495 * Improve efficiency of mem_gets: don't search whole buffer each time
9498 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9500 * New -hex option for openssl rand.
9504 * Print out UTF8String and NumericString when parsing ASN1.
9508 * Support NumericString type for name components.
9512 * Allow CC in the environment to override the automatically chosen
9513 compiler. Note that nothing is done to ensure flags work with the
9518 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9520 * Properly check EVP_VerifyFinal() and similar return values
9523 *Ben Laurie, Bodo Moeller, Google Security Team*
9525 * Enable TLS extensions by default.
9529 * Allow the CHIL engine to be loaded, whether the application is
9530 multithreaded or not. (This does not release the developer from the
9531 obligation to set up the dynamic locking callbacks.)
9533 *Sander Temme <sander@temme.net>*
9535 * Use correct exit code if there is an error in dgst command.
9537 *Steve Henson; problem pointed out by Roland Dirlewanger*
9539 * Tweak Configure so that you need to say "experimental-jpake" to enable
9540 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9544 * Add experimental JPAKE support, including demo authentication in
9545 s_client and s_server.
9549 * Set the comparison function in v3_addr_canonize().
9551 *Rob Austein <sra@hactrn.net>*
9553 * Add support for XMPP STARTTLS in s_client.
9555 *Philip Paeps <philip@freebsd.org>*
9557 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9558 to ensure that even with this option, only ciphersuites in the
9559 server's preference list will be accepted. (Note that the option
9560 applies only when resuming a session, so the earlier behavior was
9561 just about the algorithm choice for symmetric cryptography.)
9565 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9567 * Fix NULL pointer dereference if a DTLS server received
9568 ChangeCipherSpec as first record ([CVE-2009-1386]).
9572 * Fix a state transition in s3_srvr.c and d1_srvr.c
9573 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9577 * The fix in 0.9.8c that supposedly got rid of unsafe
9578 double-checked locking was incomplete for RSA blinding,
9579 addressing just one layer of what turns out to have been
9580 doubly unsafe triple-checked locking.
9582 So now fix this for real by retiring the MONT_HELPER macro
9583 in crypto/rsa/rsa_eay.c.
9585 *Bodo Moeller; problem pointed out by Marius Schilder*
9587 * Various precautionary measures:
9589 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9591 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9592 (NB: This would require knowledge of the secret session ticket key
9593 to exploit, in which case you'd be SOL either way.)
9595 - Change bn_nist.c so that it will properly handle input BIGNUMs
9596 outside the expected range.
9598 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9601 *Neel Mehta, Bodo Moeller*
9603 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9604 the load fails. Useful for distros.
9606 *Ben Laurie and the FreeBSD team*
9608 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9612 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9616 * Expand ENGINE to support engine supplied SSL client certificate functions.
9618 This work was sponsored by Logica.
9622 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9623 keystores. Support for SSL/TLS client authentication too.
9624 Not compiled unless enable-capieng specified to Configure.
9626 This work was sponsored by Logica.
9630 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9631 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9632 attribute creation routines such as certificate requests and PKCS#12
9637 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9639 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9640 handshake which could lead to a client crash as found using the
9641 Codenomicon TLS test suite ([CVE-2008-1672])
9643 *Steve Henson, Mark Cox*
9645 * Fix double free in TLS server name extensions which could lead to
9646 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9650 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9652 Clear the error queue to ensure that error entries left from
9653 older function calls do not interfere with the correct operation.
9655 *Lutz Jaenicke, Erik de Castro Lopo*
9657 * Remove root CA certificates of commercial CAs:
9659 The OpenSSL project does not recommend any specific CA and does not
9660 have any policy with respect to including or excluding any CA.
9661 Therefore it does not make any sense to ship an arbitrary selection
9662 of root CA certificates with the OpenSSL software.
9666 * RSA OAEP patches to fix two separate invalid memory reads.
9667 The first one involves inputs when 'lzero' is greater than
9668 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9669 before the beginning of from). The second one involves inputs where
9670 the 'db' section contains nothing but zeroes (there is a one-byte
9671 invalid read after the end of 'db').
9673 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9675 * Partial backport from 0.9.9-dev:
9677 Introduce bn_mul_mont (dedicated Montgomery multiplication
9678 procedure) as a candidate for BIGNUM assembler implementation.
9679 While 0.9.9-dev uses assembler for various architectures, only
9680 x86_64 is available by default here in the 0.9.8 branch, and
9681 32-bit x86 is available through a compile-time setting.
9683 To try the 32-bit x86 assembler implementation, use Configure
9684 option "enable-montasm" (which exists only for this backport).
9686 As "enable-montasm" for 32-bit x86 disclaims code stability
9687 anyway, in this constellation we activate additional code
9688 backported from 0.9.9-dev for further performance improvements,
9689 namely BN_from_montgomery_word. (To enable this otherwise,
9690 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9692 *Andy Polyakov (backport partially by Bodo Moeller)*
9694 * Add TLS session ticket callback. This allows an application to set
9695 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9696 values. This is useful for key rollover for example where several key
9697 sets may exist with different names.
9701 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9702 This was broken until now in 0.9.8 releases, such that the only way
9703 a registered ENGINE could be used (assuming it initialises
9704 successfully on the host) was to explicitly set it as the default
9705 for the relevant algorithms. This is in contradiction with 0.9.7
9706 behaviour and the documentation. With this fix, when an ENGINE is
9707 registered into a given algorithm's table of implementations, the
9708 'uptodate' flag is reset so that auto-discovery will be used next
9709 time a new context for that algorithm attempts to select an
9712 *Ian Lister (tweaked by Geoff Thorpe)*
9714 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9715 implementation in the following ways:
9717 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9720 Lack of BER streaming support means one pass streaming processing is
9721 only supported if data is detached: setting the streaming flag is
9722 ignored for embedded content.
9724 CMS support is disabled by default and must be explicitly enabled
9725 with the enable-cms configuration option.
9729 * Update the GMP engine glue to do direct copies between BIGNUM and
9730 mpz_t when openssl and GMP use the same limb size. Otherwise the
9731 existing "conversion via a text string export" trick is still used.
9733 *Paul Sheer <paulsheer@gmail.com>*
9735 * Zlib compression BIO. This is a filter BIO which compressed and
9736 uncompresses any data passed through it.
9740 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9741 RFC3394 compatible AES key wrapping.
9745 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9746 sets string data without copying. X509_ALGOR_set0() and
9747 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9748 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9749 from an X509_ATTRIBUTE structure optionally checking it occurs only
9750 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9755 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9756 to get the expected BN_FLG_CONSTTIME behavior.
9758 *Bodo Moeller (Google)*
9762 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9763 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9764 - added some more tests to do_tests.pl
9765 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9766 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9767 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9768 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9769 - various changes to netware.pl to enable gcc-cross builds on Win32
9771 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9772 - various changes to fix missing prototype warnings
9773 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9774 - added AES, WHIRLPOOL and CPUID assembler code to build files
9775 - added missing AES assembler make rules to mk1mf.pl
9776 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9778 *Guenter Knauf <eflash@gmx.net>*
9780 * Implement certificate status request TLS extension defined in RFC3546.
9781 A client can set the appropriate parameters and receive the encoded
9782 OCSP response via a callback. A server can query the supplied parameters
9783 and set the encoded OCSP response in the callback. Add simplified examples
9784 to s_client and s_server.
9788 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9791 + Binary incompatibility of ssl_ctx_st structure
9792 + DTLS interoperation with non-compliant servers
9793 + Don't call get_session_cb() without proposed session
9794 + Fix ia64 assembler code
9796 *Andy Polyakov, Steve Henson*
9798 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9800 * DTLS Handshake overhaul. There were longstanding issues with
9801 OpenSSL DTLS implementation, which were making it impossible for
9802 RFC 4347 compliant client to communicate with OpenSSL server.
9803 Unfortunately just fixing these incompatibilities would "cut off"
9804 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9805 server keeps tolerating non RFC compliant syntax. The opposite is
9806 not true, 0.9.8f client can not communicate with earlier server.
9807 This update even addresses CVE-2007-4995.
9811 * Changes to avoid need for function casts in OpenSSL: some compilers
9812 (gcc 4.2 and later) reject their use.
9813 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9816 * Add RFC4507 support to OpenSSL. This includes the corrections in
9817 RFC4507bis. The encrypted ticket format is an encrypted encoded
9818 SSL_SESSION structure, that way new session features are automatically
9821 If a client application caches session in an SSL_SESSION structure
9822 support is transparent because tickets are now stored in the encoded
9825 The SSL_CTX structure automatically generates keys for ticket
9826 protection in servers so again support should be possible
9827 with no application modification.
9829 If a client or server wishes to disable RFC4507 support then the option
9830 SSL_OP_NO_TICKET can be set.
9832 Add a TLS extension debugging callback to allow the contents of any client
9833 or server extensions to be examined.
9835 This work was sponsored by Google.
9839 * Add initial support for TLS extensions, specifically for the server_name
9840 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9841 have new members for a host name. The SSL data structure has an
9842 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9843 stored in that context to allow for session resumption, even after the
9844 SSL has been switched to a new SSL_CTX in reaction to a client's
9845 server_name extension.
9847 New functions (subject to change):
9849 SSL_get_servername()
9850 SSL_get_servername_type()
9853 New CTRL codes and macros (subject to change):
9855 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9856 - SSL_CTX_set_tlsext_servername_callback()
9857 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9858 - SSL_CTX_set_tlsext_servername_arg()
9859 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9861 openssl s_client has a new '-servername ...' option.
9863 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9864 '-key2 ...', '-servername_fatal' (subject to change). This allows
9865 testing the HostName extension for a specific single host name ('-cert'
9866 and '-key' remain fallbacks for handshakes without HostName
9867 negotiation). If the unrecognized_name alert has to be sent, this by
9868 default is a warning; it becomes fatal with the '-servername_fatal'
9871 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9873 * Add AES and SSE2 assembly language support to VC++ build.
9877 * Mitigate attack on final subtraction in Montgomery reduction.
9881 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9882 (which previously caused an internal error).
9886 * Squeeze another 10% out of IGE mode when in != out.
9890 * AES IGE mode speedup.
9892 *Dean Gaudet (Google)*
9894 * Add the Korean symmetric 128-bit cipher SEED (see
9895 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9896 add SEED ciphersuites from RFC 4162:
9898 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9899 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9900 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9901 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9903 To minimize changes between patchlevels in the OpenSSL 0.9.8
9904 series, SEED remains excluded from compilation unless OpenSSL
9905 is configured with 'enable-seed'.
9907 *KISA, Bodo Moeller*
9909 * Mitigate branch prediction attacks, which can be practical if a
9910 single processor is shared, allowing a spy process to extract
9911 information. For detailed background information, see
9912 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9913 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9914 and Necessary Software Countermeasures"). The core of the change
9915 are new versions BN_div_no_branch() and
9916 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9917 respectively, which are slower, but avoid the security-relevant
9918 conditional branches. These are automatically called by BN_div()
9919 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9920 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9921 remove a conditional branch.
9923 BN_FLG_CONSTTIME is the new name for the previous
9924 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9925 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9926 in the exponent causes BN_mod_exp_mont() to use the alternative
9927 implementation in BN_mod_exp_mont_consttime().) The old name
9928 remains as a deprecated alias.
9930 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9931 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9932 constant-time implementations for more than just exponentiation.
9933 Here too the old name is kept as a deprecated alias.
9935 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9936 the BN_BLINDING structure gets an independent copy of the
9937 modulus. This means that the previous `BIGNUM *m` argument to
9938 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9939 essentially becomes `const BIGNUM *m`, although we can't actually
9940 change this in the header file before 0.9.9. It allows
9941 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9942 enable BN_FLG_CONSTTIME.
9944 *Matthew D Wood (Intel Corp)*
9946 * In the SSL/TLS server implementation, be strict about session ID
9947 context matching (which matters if an application uses a single
9948 external cache for different purposes). Previously,
9949 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9950 set. This did ensure strict client verification, but meant that,
9951 with applications using a single external cache for quite
9952 different requirements, clients could circumvent ciphersuite
9953 restrictions for a given session ID context by starting a session
9954 in a different context.
9958 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9959 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9960 authentication-only ciphersuites.
9964 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9965 not complete and could lead to a possible single byte overflow
9966 ([CVE-2007-5135]) [Ben Laurie]
9968 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9970 * Since AES128 and AES256 (and similarly Camellia128 and
9971 Camellia256) share a single mask bit in the logic of
9972 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9973 kludge to work properly if AES128 is available and AES256 isn't
9974 (or if Camellia128 is available and Camellia256 isn't).
9978 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9979 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9980 When a point or a seed is encoded in a BIT STRING, we need to
9981 prevent the removal of trailing zero bits to get the proper DER
9982 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9983 of a NamedBitList, for which trailing 0 bits need to be removed.)
9987 * Have SSL/TLS server implementation tolerate "mismatched" record
9988 protocol version while receiving ClientHello even if the
9989 ClientHello is fragmented. (The server can't insist on the
9990 particular protocol version it has chosen before the ServerHello
9991 message has informed the client about his choice.)
9995 * Add RFC 3779 support.
9997 *Rob Austein for ARIN, Ben Laurie*
9999 * Load error codes if they are not already present instead of using a
10000 static variable. This allows them to be cleanly unloaded and reloaded.
10001 Improve header file function name parsing.
10005 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10006 or CAPABILITY handshake as required by RFCs.
10008 *Goetz Babin-Ebell*
10010 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10012 * Introduce limits to prevent malicious keys being able to
10013 cause a denial of service. ([CVE-2006-2940])
10015 *Steve Henson, Bodo Moeller*
10017 * Fix ASN.1 parsing of certain invalid structures that can result
10018 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10020 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10021 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10023 * Fix SSL client code which could crash if connecting to a
10024 malicious SSLv2 server. ([CVE-2006-4343])
10026 *Tavis Ormandy and Will Drewry, Google Security Team*
10028 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10029 match only those. Before that, "AES256-SHA" would be interpreted
10030 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10031 the same strength classification in 0.9.7h) as we currently only
10032 have a single AES bit in the ciphersuite description bitmap.
10033 That change, however, also applied to ciphersuite strings such as
10034 "RC4-MD5" that intentionally matched multiple ciphersuites --
10035 namely, SSL 2.0 ciphersuites in addition to the more common ones
10036 from SSL 3.0/TLS 1.0.
10038 So we change the selection algorithm again: Naming an explicit
10039 ciphersuite selects this one ciphersuite, and any other similar
10040 ciphersuite (same bitmap) from *other* protocol versions.
10041 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10042 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10044 Since SSL 2.0 does not have any ciphersuites for which the
10045 128/256 bit distinction would be relevant, this works for now.
10046 The proper fix will be to use different bits for AES128 and
10047 AES256, which would have avoided the problems from the beginning;
10048 however, bits are scarce, so we can only do this in a new release
10049 (not just a patchlevel) when we can change the SSL_CIPHER
10050 definition to split the single 'unsigned long mask' bitmap into
10051 multiple values to extend the available space.
10055 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10057 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10058 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10060 * Add AES IGE and biIGE modes.
10064 * Change the Unix randomness entropy gathering to use poll() when
10065 possible instead of select(), since the latter has some
10066 undesirable limitations.
10068 *Darryl Miles via Richard Levitte and Bodo Moeller*
10070 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10071 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10072 cannot be implicitly activated as part of, e.g., the "AES" alias.
10073 However, please upgrade to OpenSSL 0.9.9[-dev] for
10074 non-experimental use of the ECC ciphersuites to get TLS extension
10075 support, which is required for curve and point format negotiation
10076 to avoid potential handshake problems.
10080 * Disable rogue ciphersuites:
10082 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10083 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10084 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10086 The latter two were purportedly from
10087 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10090 Also deactivate the remaining ciphersuites from
10091 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10092 unofficial, and the ID has long expired.
10096 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10097 dual-core machines) and other potential thread-safety issues.
10101 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10102 versions), which is now available for royalty-free use
10103 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10104 Also, add Camellia TLS ciphersuites from RFC 4132.
10106 To minimize changes between patchlevels in the OpenSSL 0.9.8
10107 series, Camellia remains excluded from compilation unless OpenSSL
10108 is configured with 'enable-camellia'.
10112 * Disable the padding bug check when compression is in use. The padding
10113 bug check assumes the first packet is of even length, this is not
10114 necessarily true if compression is enabled and can result in false
10115 positives causing handshake failure. The actual bug test is ancient
10116 code so it is hoped that implementations will either have fixed it by
10117 now or any which still have the bug do not support compression.
10121 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10123 * When applying a cipher rule check to see if string match is an explicit
10124 cipher suite and only match that one cipher suite if it is.
10128 * Link in manifests for VC++ if needed.
10130 *Austin Ziegler <halostatue@gmail.com>*
10132 * Update support for ECC-based TLS ciphersuites according to
10133 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10134 TLS extensions, which are supported starting with the 0.9.9
10135 branch, not in the OpenSSL 0.9.8 branch).
10139 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10140 opaque EVP_CIPHER_CTX handling.
10144 * Fixes and enhancements to zlib compression code. We now only use
10145 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10146 to conform with the standards mentioned here:
10147 <http://www.zlib.net/DLL_FAQ.txt>
10148 Static zlib linking now works on Windows and the new --with-zlib-include
10149 --with-zlib-lib options to Configure can be used to supply the location
10150 of the headers and library. Gracefully handle case where zlib library
10155 * Several fixes and enhancements to the OID generation code. The old code
10156 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10157 handle numbers larger than ULONG_MAX, truncated printing and had a
10158 non standard OBJ_obj2txt() behaviour.
10162 * Add support for building of engines under engine/ as shared libraries
10163 under VC++ build system.
10167 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10168 Hopefully, we will not see any false combination of paths any more.
10172 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10174 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10175 (part of SSL_OP_ALL). This option used to disable the
10176 countermeasure against man-in-the-middle protocol-version
10177 rollback in the SSL 2.0 server implementation, which is a bad
10178 idea. ([CVE-2005-2969])
10180 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10181 for Information Security, National Institute of Advanced Industrial
10182 Science and Technology [AIST], Japan)*
10184 * Add two function to clear and return the verify parameter flags.
10188 * Keep cipherlists sorted in the source instead of sorting them at
10189 runtime, thus removing the need for a lock.
10193 * Avoid some small subgroup attacks in Diffie-Hellman.
10195 *Nick Mathewson and Ben Laurie*
10197 * Add functions for well-known primes.
10201 * Extended Windows CE support.
10203 *Satoshi Nakamura and Andy Polyakov*
10205 * Initialize SSL_METHOD structures at compile time instead of during
10206 runtime, thus removing the need for a lock.
10210 * Make PKCS7_decrypt() work even if no certificate is supplied by
10211 attempting to decrypt each encrypted key in turn. Add support to
10216 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10218 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10221 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10225 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10226 key into the same file any more.
10230 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10234 * Add -utf8 command line and config file option to 'ca'.
10236 *Stefan <stf@udoma.org*
10238 * Removed the macro des_crypt(), as it seems to conflict with some
10239 libraries. Use DES_crypt().
10243 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10244 involves renaming the source and generated shared-libs for
10245 both. The engines will accept the corrected or legacy ids
10246 ('ncipher' and '4758_cca' respectively) when binding. NB,
10247 this only applies when building 'shared'.
10249 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10251 * Add attribute functions to EVP_PKEY structure. Modify
10252 PKCS12_create() to recognize a CSP name attribute and
10253 use it. Make -CSP option work again in pkcs12 utility.
10257 * Add new functionality to the bn blinding code:
10258 - automatic re-creation of the BN_BLINDING parameters after
10259 a fixed number of uses (currently 32)
10260 - add new function for parameter creation
10261 - introduce flags to control the update behaviour of the
10262 BN_BLINDING parameters
10263 - hide BN_BLINDING structure
10264 Add a second BN_BLINDING slot to the RSA structure to improve
10265 performance when a single RSA object is shared among several
10270 * Add support for DTLS.
10272 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10274 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10275 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10279 * Remove buggy and incomplete DH cert support from
10280 ssl/ssl_rsa.c and ssl/s3_both.c
10284 * Use SHA-1 instead of MD5 as the default digest algorithm for
10285 the `apps/openssl` commands.
10289 * Compile clean with "-Wall -Wmissing-prototypes
10290 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10291 DEBUG_SAFESTACK must also be set.
10295 * Change ./Configure so that certain algorithms can be disabled by default.
10296 The new counterpiece to "no-xxx" is "enable-xxx".
10298 The patented RC5 and MDC2 algorithms will now be disabled unless
10299 "enable-rc5" and "enable-mdc2", respectively, are specified.
10301 (IDEA remains enabled despite being patented. This is because IDEA
10302 is frequently required for interoperability, and there is no license
10303 fee for non-commercial use. As before, "no-idea" can be used to
10304 avoid this algorithm.)
10308 * Add processing of proxy certificates (see RFC 3820). This work was
10309 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10310 EGEE (Enabling Grids for E-science in Europe).
10314 * RC4 performance overhaul on modern architectures/implementations, such
10315 as Intel P4, IA-64 and AMD64.
10319 * New utility extract-section.pl. This can be used specify an alternative
10320 section number in a pod file instead of having to treat each file as
10321 a separate case in Makefile. This can be done by adding two lines to the
10324 =for comment openssl_section:XXX
10326 The blank line is mandatory.
10330 * New arguments -certform, -keyform and -pass for s_client and s_server
10331 to allow alternative format key and certificate files and passphrase
10336 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10337 update associated structures and add various utility functions.
10339 Add new policy related verify parameters, include policy checking in
10340 standard verify code. Enhance 'smime' application with extra parameters
10341 to support policy checking and print out.
10345 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10346 Nehemiah processors. These extensions support AES encryption in hardware
10347 as well as RNG (though RNG support is currently disabled).
10349 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10351 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10355 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10357 *Andy Polyakov and a number of other people*
10359 * Improved PowerPC platform support. Most notably BIGNUM assembler
10360 implementation contributed by IBM.
10362 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10364 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10365 exponent rather than 'unsigned long'. There is a corresponding change to
10366 the new 'rsa_keygen' element of the RSA_METHOD structure.
10368 *Jelte Jansen, Geoff Thorpe*
10370 * Functionality for creating the initial serial number file is now
10371 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10373 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10374 number file to 1, which is bound to cause problems. To avoid
10375 the problems while respecting compatibility between different 0.9.7
10376 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10377 CA.pl for serial number initialization. With the new release 0.9.8,
10378 we can fix the problem directly in the 'ca' utility.)
10382 * Reduced header interdependencies by declaring more opaque objects in
10383 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10384 give fewer recursive includes, which could break lazy source code - so
10385 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10386 developers should define this symbol when building and using openssl to
10387 ensure they track the recommended behaviour, interfaces, [etc], but
10388 backwards-compatible behaviour prevails when this isn't defined.
10392 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10396 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10397 This will generate a random key of the appropriate length based on the
10398 cipher context. The EVP_CIPHER can provide its own random key generation
10399 routine to support keys of a specific form. This is used in the des and
10400 3des routines to generate a key of the correct parity. Update S/MIME
10401 code to use new functions and hence generate correct parity DES keys.
10402 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10403 valid (weak or incorrect parity).
10407 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10408 as looking them up. This is useful when the verified structure may contain
10409 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10410 present unless the new PKCS7_NO_CRL flag is asserted.
10414 * Extend ASN1 oid configuration module. It now additionally accepts the
10417 shortName = some long name, 1.2.3.4
10421 * Reimplemented the BN_CTX implementation. There is now no more static
10422 limitation on the number of variables it can handle nor the depth of the
10423 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10424 information can now expand as required, and rather than having a single
10425 static array of bignums, BN_CTX now uses a linked-list of such arrays
10426 allowing it to expand on demand whilst maintaining the usefulness of
10427 BN_CTX's "bundling".
10431 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10432 to allow all RSA operations to function using a single BN_CTX.
10436 * Preliminary support for certificate policy evaluation and checking. This
10437 is initially intended to pass the tests outlined in "Conformance Testing
10438 of Relying Party Client Certificate Path Processing Logic" v1.07.
10442 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10443 remained unused and not that useful. A variety of other little bignum
10444 tweaks and fixes have also been made continuing on from the audit (see
10449 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10450 associated ASN1, EVP and SSL functions and old ASN1 macros.
10454 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10455 and this should never fail. So the return value from the use of
10456 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10457 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10461 * BN_CTX_get() should return zero-valued bignums, providing the same
10462 initialised value as BN_new().
10464 *Geoff Thorpe, suggested by Ulf Möller*
10466 * Support for inhibitAnyPolicy certificate extension.
10470 * An audit of the BIGNUM code is underway, for which debugging code is
10471 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10472 is considered valid when processing BIGNUMs, and causes execution to
10473 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10474 further steps are taken to deliberately pollute unused data in BIGNUM
10475 structures to try and expose faulty code further on. For now, openssl will
10476 (in its default mode of operation) continue to tolerate the inconsistent
10477 forms that it has tolerated in the past, but authors and packagers should
10478 consider trying openssl and their own applications when compiled with
10479 these debugging symbols defined. It will help highlight potential bugs in
10480 their own code, and will improve the test coverage for OpenSSL itself. At
10481 some point, these tighter rules will become openssl's default to improve
10482 maintainability, though the assert()s and other overheads will remain only
10483 in debugging configurations. See bn.h for more details.
10485 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10487 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10488 that can only be obtained through BN_CTX_new() (which implicitly
10489 initialises it). The presence of this function only made it possible
10490 to overwrite an existing structure (and cause memory leaks).
10494 * Because of the callback-based approach for implementing LHASH as a
10495 template type, lh_insert() adds opaque objects to hash-tables and
10496 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10497 to clean up those corresponding objects before destroying the hash table
10498 (and losing the object pointers). So some over-zealous constifications in
10499 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10500 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10501 prototyped to have "const" restrictions on the object pointers they are
10502 given (and so aren't required to cast them away any more).
10506 * The tmdiff.h API was so ugly and minimal that our own timing utility
10507 (speed) prefers to use its own implementation. The two implementations
10508 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10509 its object type properly exposed (MS_TM) instead of casting to/from
10510 `char *`. This may still change yet if someone realises MS_TM and
10512 aren't necessarily the greatest nomenclatures - but this is what was used
10513 internally to the implementation so I've used that for now.
10517 * Ensure that deprecated functions do not get compiled when
10518 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10519 the self-tests were still using deprecated key-generation functions so
10520 these have been updated also.
10524 * Reorganise PKCS#7 code to separate the digest location functionality
10525 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10526 New function PKCS7_set_digest() to set the digest type for PKCS#7
10527 digestedData type. Add additional code to correctly generate the
10528 digestedData type and add support for this type in PKCS7 initialization
10533 * New function PKCS7_set0_type_other() this initializes a PKCS7
10534 structure of type "other".
10538 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10539 sure the loop does correctly stop and breaking ("division by zero")
10540 modulus operations are not performed. The (pre-generated) prime
10541 table crypto/bn/bn_prime.h was already correct, but it could not be
10542 re-generated on some platforms because of the "division by zero"
10543 situation in the script.
10545 *Ralf S. Engelschall*
10547 * Update support for ECC-based TLS ciphersuites according to
10548 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10549 SHA-1 now is only used for "small" curves (where the
10550 representation of a field element takes up to 24 bytes); for
10551 larger curves, the field element resulting from ECDH is directly
10552 used as premaster secret.
10554 *Douglas Stebila (Sun Microsystems Laboratories)*
10556 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10557 curve secp160r1 to the tests.
10559 *Douglas Stebila (Sun Microsystems Laboratories)*
10561 * Add the possibility to load symbols globally with DSO.
10563 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10565 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10566 control of the error stack.
10570 * Add support for STORE in ENGINE.
10574 * Add the STORE type. The intention is to provide a common interface
10575 to certificate and key stores, be they simple file-based stores, or
10576 HSM-type store, or LDAP stores, or...
10577 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10581 * Add a generic structure called OPENSSL_ITEM. This can be used to
10582 pass a list of arguments to any function as well as provide a way
10583 for a function to pass data back to the caller.
10587 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10588 works like BUF_strdup() but can be used to duplicate a portion of
10589 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10594 * Add the function sk_find_ex() which works like sk_find(), but will
10595 return an index to an element even if an exact match couldn't be
10596 found. The index is guaranteed to point at the element where the
10597 searched-for key would be inserted to preserve sorting order.
10601 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10602 takes an extra flags argument for optional functionality. Currently,
10603 the following flags are defined:
10605 OBJ_BSEARCH_VALUE_ON_NOMATCH
10606 This one gets OBJ_bsearch_ex() to return a pointer to the first
10607 element where the comparing function returns a negative or zero
10610 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10611 This one gets OBJ_bsearch_ex() to return a pointer to the first
10612 element where the comparing function returns zero. This is useful
10613 if there are more than one element where the comparing function
10618 * Make it possible to create self-signed certificates with 'openssl ca'
10619 in such a way that the self-signed certificate becomes part of the
10620 CA database and uses the same mechanisms for serial number generation
10621 as all other certificate signing. The new flag '-selfsign' enables
10622 this functionality. Adapt CA.sh and CA.pl.in.
10626 * Add functionality to check the public key of a certificate request
10627 against a given private. This is useful to check that a certificate
10628 request can be signed by that key (self-signing).
10632 * Make it possible to have multiple active certificates with the same
10633 subject in the CA index file. This is done only if the keyword
10634 'unique_subject' is set to 'no' in the main CA section (default
10635 if 'CA_default') of the configuration file. The value is saved
10636 with the database itself in a separate index attribute file,
10637 named like the index file with '.attr' appended to the name.
10641 * Generate multi-valued AVAs using '+' notation in config files for
10646 * Support for nameConstraints certificate extension.
10650 * Support for policyConstraints certificate extension.
10654 * Support for policyMappings certificate extension.
10658 * Make sure the default DSA_METHOD implementation only uses its
10659 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10660 and change its own handlers to be NULL so as to remove unnecessary
10661 indirection. This lets alternative implementations fallback to the
10662 default implementation more easily.
10666 * Support for directoryName in GeneralName related extensions
10671 * Make it possible to link applications using Makefile.shared.
10672 Make that possible even when linking against static libraries!
10676 * Support for single pass processing for S/MIME signing. This now
10677 means that S/MIME signing can be done from a pipe, in addition
10678 cleartext signing (multipart/signed type) is effectively streaming
10679 and the signed data does not need to be all held in memory.
10681 This is done with a new flag PKCS7_STREAM. When this flag is set
10682 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10683 is done after the data is output (and digests calculated) in
10684 SMIME_write_PKCS7().
10688 * Add full support for -rpath/-R, both in shared libraries and
10689 applications, at least on the platforms where it's known how
10694 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10695 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10696 will now compute a table of multiples of the generator that
10697 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10698 faster (notably in the case of a single point multiplication,
10699 scalar * generator).
10701 *Nils Larsch, Bodo Moeller*
10703 * IPv6 support for certificate extensions. The various extensions
10704 which use the IP:a.b.c.d can now take IPv6 addresses using the
10705 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10710 * Added an ENGINE that implements RSA by performing private key
10711 exponentiations with the GMP library. The conversions to and from
10712 GMP's mpz_t format aren't optimised nor are any montgomery forms
10713 cached, and on x86 it appears OpenSSL's own performance has caught up.
10714 However there are likely to be other architectures where GMP could
10715 provide a boost. This ENGINE is not built in by default, but it can be
10716 specified at Configure time and should be accompanied by the necessary
10717 linker additions, eg;
10718 ./config -DOPENSSL_USE_GMP -lgmp
10722 * "openssl engine" will not display ENGINE/DSO load failure errors when
10723 testing availability of engines with "-t" - the old behaviour is
10724 produced by increasing the feature's verbosity with "-tt".
10728 * ECDSA routines: under certain error conditions uninitialized BN objects
10729 could be freed. Solution: make sure initialization is performed early
10730 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10735 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10736 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10737 software implementations. For DSA and DH, parameter generation can
10738 also be overridden by providing the appropriate method callbacks.
10742 * Change the "progress" mechanism used in key-generation and
10743 primality testing to functions that take a new BN_GENCB pointer in
10744 place of callback/argument pairs. The new API functions have `_ex`
10745 postfixes and the older functions are reimplemented as wrappers for
10746 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10747 declarations of the old functions to help (graceful) attempts to
10748 migrate to the new functions. Also, the new key-generation API
10749 functions operate on a caller-supplied key-structure and return
10750 success/failure rather than returning a key or NULL - this is to
10751 help make "keygen" another member function of RSA_METHOD etc.
10753 Example for using the new callback interface:
10755 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10756 void *my_arg = ...;
10759 BN_GENCB_set(&my_cb, my_callback, my_arg);
10761 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10762 /* For the meaning of a, b in calls to my_callback(), see the
10763 * documentation of the function that calls the callback.
10764 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10765 * my_callback should return 1 if it wants BN_is_prime_ex()
10766 * to continue, or 0 to stop.
10771 * Change the ZLIB compression method to be stateful, and make it
10772 available to TLS with the number defined in
10773 draft-ietf-tls-compression-04.txt.
10777 * Add the ASN.1 structures and functions for CertificatePair, which
10778 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10780 CertificatePair ::= SEQUENCE {
10781 forward [0] Certificate OPTIONAL,
10782 reverse [1] Certificate OPTIONAL,
10783 -- at least one of the pair shall be present -- }
10785 Also implement the PEM functions to read and write certificate
10786 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10788 This needed to be defined, mostly for the sake of the LDAP
10789 attribute crossCertificatePair, but may prove useful elsewhere as
10794 * Make it possible to inhibit symlinking of shared libraries in
10795 Makefile.shared, for Cygwin's sake.
10799 * Extend the BIGNUM API by creating a function
10800 void BN_set_negative(BIGNUM *a, int neg);
10801 and a macro that behave like
10802 int BN_is_negative(const BIGNUM *a);
10804 to avoid the need to access 'a->neg' directly in applications.
10808 * Implement fast modular reduction for pseudo-Mersenne primes
10809 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10810 EC_GROUP_new_curve_GFp() will now automatically use this
10813 *Nils Larsch <nla@trustcenter.de>*
10815 * Add new lock type (CRYPTO_LOCK_BN).
10819 * Change the ENGINE framework to automatically load engines
10820 dynamically from specific directories unless they could be
10821 found to already be built in or loaded. Move all the
10822 current engines except for the cryptodev one to a new
10823 directory engines/.
10824 The engines in engines/ are built as shared libraries if
10825 the "shared" options was given to ./Configure or ./config.
10826 Otherwise, they are inserted in libcrypto.a.
10827 /usr/local/ssl/engines is the default directory for dynamic
10828 engines, but that can be overridden at configure time through
10829 the usual use of --prefix and/or --openssldir, and at run
10830 time with the environment variable OPENSSL_ENGINES.
10832 *Geoff Thorpe and Richard Levitte*
10834 * Add Makefile.shared, a helper makefile to build shared
10835 libraries. Adapt Makefile.org.
10839 * Add version info to Win32 DLLs.
10841 *Peter 'Luna' Runestig" <peter@runestig.com>*
10843 * Add new 'medium level' PKCS#12 API. Certificates and keys
10844 can be added using this API to created arbitrary PKCS#12
10845 files while avoiding the low level API.
10847 New options to PKCS12_create(), key or cert can be NULL and
10848 will then be omitted from the output file. The encryption
10849 algorithm NIDs can be set to -1 for no encryption, the mac
10850 iteration count can be set to 0 to omit the mac.
10852 Enhance pkcs12 utility by making the -nokeys and -nocerts
10853 options work when creating a PKCS#12 file. New option -nomac
10854 to omit the mac, NONE can be set for an encryption algorithm.
10855 New code is modified to use the enhanced PKCS12_create()
10856 instead of the low level API.
10860 * Extend ASN1 encoder to support indefinite length constructed
10861 encoding. This can output sequences tags and octet strings in
10862 this form. Modify pk7_asn1.c to support indefinite length
10863 encoding. This is experimental and needs additional code to
10864 be useful, such as an ASN1 bio and some enhanced streaming
10867 Extend template encode functionality so that tagging is passed
10868 down to the template encoder.
10872 * Let 'openssl req' fail if an argument to '-newkey' is not
10873 recognized instead of using RSA as a default.
10877 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10878 As these are not official, they are not included in "ALL";
10879 the "ECCdraft" ciphersuite group alias can be used to select them.
10881 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10883 * Add ECDH engine support.
10885 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10887 * Add ECDH in new directory crypto/ecdh/.
10889 *Douglas Stebila (Sun Microsystems Laboratories)*
10891 * Let BN_rand_range() abort with an error after 100 iterations
10892 without success (which indicates a broken PRNG).
10896 * Change BN_mod_sqrt() so that it verifies that the input value
10897 is really the square of the return value. (Previously,
10898 BN_mod_sqrt would show GIGO behaviour.)
10902 * Add named elliptic curves over binary fields from X9.62, SECG,
10903 and WAP/WTLS; add OIDs that were still missing.
10905 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10907 * Extend the EC library for elliptic curves over binary fields
10908 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10911 EC_GF2m_simple_method
10915 EC_GROUP_new_curve_GF2m
10916 EC_GROUP_set_curve_GF2m
10917 EC_GROUP_get_curve_GF2m
10918 EC_POINT_set_affine_coordinates_GF2m
10919 EC_POINT_get_affine_coordinates_GF2m
10920 EC_POINT_set_compressed_coordinates_GF2m
10922 Point compression for binary fields is disabled by default for
10923 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10926 As binary polynomials are represented as BIGNUMs, various members
10927 of the EC_GROUP and EC_POINT data structures can be shared
10928 between the implementations for prime fields and binary fields;
10929 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10930 are essentially identical to their `..._GFp` counterparts.
10931 (For simplicity, the `..._GFp` prefix has been dropped from
10932 various internal method names.)
10934 An internal 'field_div' method (similar to 'field_mul' and
10935 'field_sqr') has been added; this is used only for binary fields.
10937 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10939 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10940 through methods ('mul', 'precompute_mult').
10942 The generic implementations (now internally called 'ec_wNAF_mul'
10943 and 'ec_wNAF_precomputed_mult') remain the default if these
10944 methods are undefined.
10946 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10948 * New function EC_GROUP_get_degree, which is defined through
10949 EC_METHOD. For curves over prime fields, this returns the bit
10950 length of the modulus.
10952 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10954 * New functions EC_GROUP_dup, EC_POINT_dup.
10955 (These simply call ..._new and ..._copy).
10957 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10959 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10960 Polynomials are represented as BIGNUMs (where the sign bit is not
10961 used) in the following functions [macros]:
10964 BN_GF2m_sub [= BN_GF2m_add]
10965 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10966 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10967 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10969 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10970 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10971 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10972 BN_GF2m_cmp [= BN_ucmp]
10974 (Note that only the 'mod' functions are actually for fields GF(2^m).
10975 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10977 For some functions, an the irreducible polynomial defining a
10978 field can be given as an 'unsigned int[]' with strictly
10979 decreasing elements giving the indices of those bits that are set;
10980 i.e., p[] represents the polynomial
10981 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10983 p[0] > p[1] > ... > p[k] = 0.
10984 This applies to the following functions:
10987 BN_GF2m_mod_mul_arr
10988 BN_GF2m_mod_sqr_arr
10989 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10990 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10991 BN_GF2m_mod_exp_arr
10992 BN_GF2m_mod_sqrt_arr
10993 BN_GF2m_mod_solve_quad_arr
10997 Conversion can be performed by the following functions:
11002 bntest.c has additional tests for binary polynomial arithmetic.
11004 Two implementations for BN_GF2m_mod_div() are available.
11005 The default algorithm simply uses BN_GF2m_mod_inv() and
11006 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11007 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11008 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11010 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11012 * Add new error code 'ERR_R_DISABLED' that can be used when some
11013 functionality is disabled at compile-time.
11015 *Douglas Stebila <douglas.stebila@sun.com>*
11017 * Change default behaviour of 'openssl asn1parse' so that more
11018 information is visible when viewing, e.g., a certificate:
11020 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11021 mode the content of non-printable OCTET STRINGs is output in a
11022 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11023 avoid the appearance of a printable string.
11025 *Nils Larsch <nla@trustcenter.de>*
11027 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11029 EC_GROUP_set_asn1_flag()
11030 EC_GROUP_get_asn1_flag()
11031 EC_GROUP_set_point_conversion_form()
11032 EC_GROUP_get_point_conversion_form()
11033 These control ASN1 encoding details:
11034 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11035 has been set to OPENSSL_EC_NAMED_CURVE.
11036 - Points are encoded in uncompressed form by default; options for
11037 asn1_for are as for point2oct, namely
11038 POINT_CONVERSION_COMPRESSED
11039 POINT_CONVERSION_UNCOMPRESSED
11040 POINT_CONVERSION_HYBRID
11042 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11044 EC_GROUP_set_seed()
11045 EC_GROUP_get0_seed()
11046 EC_GROUP_get_seed_len()
11047 This is used only for ASN1 purposes (so far).
11049 *Nils Larsch <nla@trustcenter.de>*
11051 * Add 'field_type' member to EC_METHOD, which holds the NID
11052 of the appropriate field type OID. The new function
11053 EC_METHOD_get_field_type() returns this value.
11055 *Nils Larsch <nla@trustcenter.de>*
11058 EC_POINT_point2bn()
11059 EC_POINT_bn2point()
11060 EC_POINT_point2hex()
11061 EC_POINT_hex2point()
11062 providing useful interfaces to EC_POINT_point2oct() and
11063 EC_POINT_oct2point().
11065 *Nils Larsch <nla@trustcenter.de>*
11067 * Change internals of the EC library so that the functions
11068 EC_GROUP_set_generator()
11069 EC_GROUP_get_generator()
11070 EC_GROUP_get_order()
11071 EC_GROUP_get_cofactor()
11072 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11073 to methods, which would lead to unnecessary code duplication when
11074 adding different types of curves.
11076 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11078 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11079 arithmetic, and such that modified wNAFs are generated
11080 (which avoid length expansion in many cases).
11084 * Add a function EC_GROUP_check_discriminant() (defined via
11085 EC_METHOD) that verifies that the curve discriminant is non-zero.
11087 Add a function EC_GROUP_check() that makes some sanity tests
11088 on a EC_GROUP, its generator and order. This includes
11089 EC_GROUP_check_discriminant().
11091 *Nils Larsch <nla@trustcenter.de>*
11093 * Add ECDSA in new directory crypto/ecdsa/.
11095 Add applications 'openssl ecparam' and 'openssl ecdsa'
11096 (these are based on 'openssl dsaparam' and 'openssl dsa').
11098 ECDSA support is also included in various other files across the
11099 library. Most notably,
11100 - 'openssl req' now has a '-newkey ecdsa:file' option;
11101 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11102 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11103 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11104 them suitable for ECDSA where domain parameters must be
11105 extracted before the specific public key;
11106 - ECDSA engine support has been added.
11108 *Nils Larsch <nla@trustcenter.de>*
11110 * Include some named elliptic curves, and add OIDs from X9.62,
11111 SECG, and WAP/WTLS. Each curve can be obtained from the new
11113 EC_GROUP_new_by_curve_name(),
11114 and the list of available named curves can be obtained with
11115 EC_get_builtin_curves().
11116 Also add a 'curve_name' member to EC_GROUP objects, which can be
11118 EC_GROUP_set_curve_name()
11119 EC_GROUP_get_curve_name()
11121 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11123 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11124 was actually never needed) and in BN_mul(). The removal in BN_mul()
11125 required a small change in bn_mul_part_recursive() and the addition
11126 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11127 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11128 bn_sub_words() and bn_add_words() except they take arrays with
11133 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11135 * Cleanse PEM buffers before freeing them since they may contain
11138 *Benjamin Bennett <ben@psc.edu>*
11140 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11141 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11142 authentication-only ciphersuites.
11146 * Since AES128 and AES256 share a single mask bit in the logic of
11147 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11148 kludge to work properly if AES128 is available and AES256 isn't.
11152 * Expand security boundary to match 1.1.1 module.
11156 * Remove redundant features: hash file source, editing of test vectors
11157 modify fipsld to use external fips_premain.c signature.
11161 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11162 run algorithm test programs.
11166 * Make algorithm test programs more tolerant of whitespace.
11170 * Have SSL/TLS server implementation tolerate "mismatched" record
11171 protocol version while receiving ClientHello even if the
11172 ClientHello is fragmented. (The server can't insist on the
11173 particular protocol version it has chosen before the ServerHello
11174 message has informed the client about his choice.)
11178 * Load error codes if they are not already present instead of using a
11179 static variable. This allows them to be cleanly unloaded and reloaded.
11183 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11185 * Introduce limits to prevent malicious keys being able to
11186 cause a denial of service. ([CVE-2006-2940])
11188 *Steve Henson, Bodo Moeller*
11190 * Fix ASN.1 parsing of certain invalid structures that can result
11191 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11193 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11194 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11196 * Fix SSL client code which could crash if connecting to a
11197 malicious SSLv2 server. ([CVE-2006-4343])
11199 *Tavis Ormandy and Will Drewry, Google Security Team*
11201 * Change ciphersuite string processing so that an explicit
11202 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11203 will no longer include "AES128-SHA"), and any other similar
11204 ciphersuite (same bitmap) from *other* protocol versions (so that
11205 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11206 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11207 changes from 0.9.8b and 0.9.8d.
11211 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11213 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11214 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11216 * Change the Unix randomness entropy gathering to use poll() when
11217 possible instead of select(), since the latter has some
11218 undesirable limitations.
11220 *Darryl Miles via Richard Levitte and Bodo Moeller*
11222 * Disable rogue ciphersuites:
11224 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11225 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11226 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11228 The latter two were purportedly from
11229 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11232 Also deactivate the remaining ciphersuites from
11233 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11234 unofficial, and the ID has long expired.
11238 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11239 dual-core machines) and other potential thread-safety issues.
11243 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11245 * Adapt fipsld and the build system to link against the validated FIPS
11246 module in FIPS mode.
11250 * Fixes for VC++ 2005 build under Windows.
11254 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11255 from a Windows bash shell such as MSYS. It is autodetected from the
11256 "config" script when run from a VC++ environment. Modify standard VC++
11257 build to use fipscanister.o from the GNU make build.
11261 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11263 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11264 The value now differs depending on if you build for FIPS or not.
11265 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11266 safely run with a non-FIPSed libcrypto, as it may crash because of
11267 the difference induced by this change.
11271 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11273 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11274 (part of SSL_OP_ALL). This option used to disable the
11275 countermeasure against man-in-the-middle protocol-version
11276 rollback in the SSL 2.0 server implementation, which is a bad
11277 idea. ([CVE-2005-2969])
11279 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11280 for Information Security, National Institute of Advanced Industrial
11281 Science and Technology [AIST, Japan)]*
11283 * Minimal support for X9.31 signatures and PSS padding modes. This is
11284 mainly for FIPS compliance and not fully integrated at this stage.
11288 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11289 the exponentiation using a fixed-length exponent. (Otherwise,
11290 the information leaked through timing could expose the secret key
11291 after many signatures; cf. Bleichenbacher's attack on DSA with
11296 * Make a new fixed-window mod_exp implementation the default for
11297 RSA, DSA, and DH private-key operations so that the sequence of
11298 squares and multiplies and the memory access pattern are
11299 independent of the particular secret key. This will mitigate
11300 cache-timing and potential related attacks.
11302 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11303 and this is automatically used by BN_mod_exp_mont() if the new flag
11304 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11305 will use this BN flag for private exponents unless the flag
11306 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11307 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11309 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11311 * Change the client implementation for SSLv23_method() and
11312 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11313 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11314 (Previously, the SSL 2.0 backwards compatible Client Hello
11315 message format would be used even with SSL_OP_NO_SSLv2.)
11319 * Add support for smime-type MIME parameter in S/MIME messages which some
11324 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11325 a threadsafe manner. Modify rsa code to use new function and add calls
11326 to dsa and dh code (which had race conditions before).
11330 * Include the fixed error library code in the C error file definitions
11331 instead of fixing them up at runtime. This keeps the error code
11332 structures constant.
11336 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11338 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11341 * Fixes for newer kerberos headers. NB: the casts are needed because
11342 the 'length' field is signed on one version and unsigned on another
11343 with no (?) obvious way to tell the difference, without these VC++
11344 complains. Also the "definition" of FAR (blank) is no longer included
11345 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11346 some needed definitions.
11350 * Undo Cygwin change.
11354 * Added support for proxy certificates according to RFC 3820.
11355 Because they may be a security thread to unaware applications,
11356 they must be explicitly allowed in run-time. See
11357 docs/HOWTO/proxy_certificates.txt for further information.
11361 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11363 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11364 server and client random values. Previously
11365 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11366 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11368 This change has negligible security impact because:
11370 1. Server and client random values still have 24 bytes of pseudo random
11373 2. Server and client random values are sent in the clear in the initial
11376 3. The master secret is derived using the premaster secret (48 bytes in
11377 size for static RSA ciphersuites) as well as client server and random
11380 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11383 *Stephen Henson, reported by UK NISCC*
11385 * Use Windows randomness collection on Cygwin.
11389 * Fix hang in EGD/PRNGD query when communication socket is closed
11390 prematurely by EGD/PRNGD.
11392 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11394 * Prompt for pass phrases when appropriate for PKCS12 input format.
11398 * Back-port of selected performance improvements from development
11399 branch, as well as improved support for PowerPC platforms.
11403 * Add lots of checks for memory allocation failure, error codes to indicate
11404 failure and freeing up memory if a failure occurs.
11406 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11408 * Add new -passin argument to dgst.
11412 * Perform some character comparisons of different types in X509_NAME_cmp:
11413 this is needed for some certificates that re-encode DNs into UTF8Strings
11414 (in violation of RFC3280) and can't or won't issue name rollover
11419 * Make an explicit check during certificate validation to see that
11420 the CA setting in each certificate on the chain is correct. As a
11421 side effect always do the following basic checks on extensions,
11422 not just when there's an associated purpose to the check:
11424 - if there is an unhandled critical extension (unless the user
11425 has chosen to ignore this fault)
11426 - if the path length has been exceeded (if one is set at all)
11427 - that certain extensions fit the associated purpose (if one has
11432 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11434 * Avoid a race condition when CRLs are checked in a multi threaded
11435 environment. This would happen due to the reordering of the revoked
11436 entries during signature checking and serial number lookup. Now the
11437 encoding is cached and the serial number sort performed under a lock.
11438 Add new STACK function sk_is_sorted().
11442 * Add Delta CRL to the extension code.
11446 * Various fixes to s3_pkt.c so alerts are sent properly.
11448 *David Holmes <d.holmes@f5.com>*
11450 * Reduce the chances of duplicate issuer name and serial numbers (in
11451 violation of RFC3280) using the OpenSSL certificate creation utilities.
11452 This is done by creating a random 64 bit value for the initial serial
11453 number when a serial number file is created or when a self signed
11454 certificate is created using 'openssl req -x509'. The initial serial
11455 number file is created using 'openssl x509 -next_serial' in CA.pl
11456 rather than being initialized to 1.
11460 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11462 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11463 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11465 *Joe Orton, Steve Henson*
11467 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11470 *Joe Orton, Steve Henson*
11472 * Make it possible to have multiple active certificates with the same
11473 subject in the CA index file. This is done only if the keyword
11474 'unique_subject' is set to 'no' in the main CA section (default
11475 if 'CA_default') of the configuration file. The value is saved
11476 with the database itself in a separate index attribute file,
11477 named like the index file with '.attr' appended to the name.
11481 * X509 verify fixes. Disable broken certificate workarounds when
11482 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11483 keyUsage extension present. Don't accept CRLs with unhandled critical
11484 extensions: since verify currently doesn't process CRL extensions this
11485 rejects a CRL with *any* critical extensions. Add new verify error codes
11490 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11491 A clarification of RFC2560 will require the use of OCTET STRINGs and
11492 some implementations cannot handle the current raw format. Since OpenSSL
11493 copies and compares OCSP nonces as opaque blobs without any attempt at
11494 parsing them this should not create any compatibility issues.
11498 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11499 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11500 this HMAC (and other) operations are several times slower than OpenSSL
11505 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11507 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11509 * Use the correct content when signing type "other".
11513 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11515 * Fix various bugs revealed by running the NISCC test suite:
11517 Stop out of bounds reads in the ASN1 code when presented with
11518 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11520 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11522 If verify callback ignores invalid public key errors don't try to check
11523 certificate signature with the NULL public key.
11527 * New -ignore_err option in ocsp application to stop the server
11528 exiting on the first error in a request.
11532 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11533 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11538 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11539 extra data after the compression methods not only for TLS 1.0
11540 but also for SSL 3.0 (as required by the specification).
11542 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11544 * Change X509_certificate_type() to mark the key as exported/exportable
11545 when it's 512 *bits* long, not 512 bytes.
11549 * Change AES_cbc_encrypt() so it outputs exact multiple of
11550 blocks during encryption.
11554 * Various fixes to base64 BIO and non blocking I/O. On write
11555 flushes were not handled properly if the BIO retried. On read
11556 data was not being buffered properly and had various logic bugs.
11557 This also affects blocking I/O when the data being decoded is a
11562 * Various S/MIME bugfixes and compatibility changes:
11563 output correct application/pkcs7 MIME type if
11564 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11565 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11566 of files as .eml work). Correctly handle very long lines in MIME
11571 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11573 * Countermeasure against the Klima-Pokorny-Rosa extension of
11574 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11575 a protocol version number mismatch like a decryption error
11576 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11580 * Turn on RSA blinding by default in the default implementation
11581 to avoid a timing attack. Applications that don't want it can call
11582 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11583 They would be ill-advised to do so in most cases.
11585 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11587 * Change RSA blinding code so that it works when the PRNG is not
11588 seeded (in this case, the secret RSA exponent is abused as
11589 an unpredictable seed -- if it is not unpredictable, there
11590 is no point in blinding anyway). Make RSA blinding thread-safe
11591 by remembering the creator's thread ID in rsa->blinding and
11592 having all other threads use local one-time blinding factors
11593 (this requires more computation than sharing rsa->blinding, but
11594 avoids excessive locking; and if an RSA object is not shared
11595 between threads, blinding will still be very fast).
11599 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11600 ENGINE as defaults for all supported algorithms irrespective of
11601 the 'flags' parameter. 'flags' is now honoured, so applications
11602 should make sure they are passing it correctly.
11606 * Target "mingw" now allows native Windows code to be generated in
11607 the Cygwin environment as well as with the MinGW compiler.
11611 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11613 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11614 via timing by performing a MAC computation even if incorrect
11615 block cipher padding has been found. This is a countermeasure
11616 against active attacks where the attacker has to distinguish
11617 between bad padding and a MAC verification error. ([CVE-2003-0078])
11619 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11620 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11621 Martin Vuagnoux (EPFL, Ilion)*
11623 * Make the no-err option work as intended. The intention with no-err
11624 is not to have the whole error stack handling routines removed from
11625 libcrypto, it's only intended to remove all the function name and
11626 reason texts, thereby removing some of the footprint that may not
11627 be interesting if those errors aren't displayed anyway.
11629 NOTE: it's still possible for any application or module to have its
11630 own set of error texts inserted. The routines are there, just not
11631 used by default when no-err is given.
11635 * Add support for FreeBSD on IA64.
11637 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11639 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11640 Kerberos function mit_des_cbc_cksum(). Before this change,
11641 the value returned by DES_cbc_cksum() was like the one from
11642 mit_des_cbc_cksum(), except the bytes were swapped.
11644 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11646 * Allow an application to disable the automatic SSL chain building.
11647 Before this a rather primitive chain build was always performed in
11648 ssl3_output_cert_chain(): an application had no way to send the
11649 correct chain if the automatic operation produced an incorrect result.
11651 Now the chain builder is disabled if either:
11653 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11655 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11657 The reasoning behind this is that an application would not want the
11658 auto chain building to take place if extra chain certificates are
11659 present and it might also want a means of sending no additional
11660 certificates (for example the chain has two certificates and the
11665 * Add the possibility to build without the ENGINE framework.
11667 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11669 * Under Win32 gmtime() can return NULL: check return value in
11670 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11674 * DSA routines: under certain error conditions uninitialized BN objects
11675 could be freed. Solution: make sure initialization is performed early
11676 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11677 Nils Larsch <nla@trustcenter.de> via PR#459)
11681 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11682 checked on reconnect on the client side, therefore session resumption
11683 could still fail with a "ssl session id is different" error. This
11684 behaviour is masked when SSL_OP_ALL is used due to
11685 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11686 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11687 followup to PR #377.
11691 * IA-32 assembler support enhancements: unified ELF targets, support
11692 for SCO/Caldera platforms, fix for Cygwin shared build.
11696 * Add support for FreeBSD on sparc64. As a consequence, support for
11697 FreeBSD on non-x86 processors is separate from x86 processors on
11698 the config script, much like the NetBSD support.
11700 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11702 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11704 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11707 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11708 code (06) was taken as the first octet of the session ID and the last
11709 octet was ignored consequently. As a result SSLv2 client side session
11710 caching could not have worked due to the session ID mismatch between
11712 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11717 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11718 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11723 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11724 seems that in spite of existing for more than a year, many application
11725 author have done nothing to provide the necessary callbacks, which
11726 means that this particular engine will not work properly anywhere.
11727 This is a very unfortunate situation which forces us, in the name
11728 of usability, to give the hw_ncipher.c a static lock, which is part
11730 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11731 appear in 0.9.8 or later. We EXPECT application authors to have
11732 dealt properly with this when 0.9.8 is released (unless we actually
11733 make such changes in the libcrypto locking code that changes will
11734 have to be made anyway).
11738 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11739 octets have been read, EOF or an error occurs. Without this change
11740 some truncated ASN1 structures will not produce an error.
11744 * Disable Heimdal support, since it hasn't been fully implemented.
11745 Still give the possibility to force the use of Heimdal, but with
11746 warnings and a request that patches get sent to openssl-dev.
11750 * Add the VC-CE target, introduce the WINCE sysname, and add
11751 INSTALL.WCE and appropriate conditionals to make it build.
11753 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11755 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11756 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11757 edit numbers of the version.
11759 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11761 * Introduce safe string copy and catenation functions
11762 (BUF_strlcpy() and BUF_strlcat()).
11764 *Ben Laurie (CHATS) and Richard Levitte*
11766 * Avoid using fixed-size buffers for one-line DNs.
11768 *Ben Laurie (CHATS)*
11770 * Add BUF_MEM_grow_clean() to avoid information leakage when
11771 resizing buffers containing secrets, and use where appropriate.
11773 *Ben Laurie (CHATS)*
11775 * Avoid using fixed size buffers for configuration file location.
11777 *Ben Laurie (CHATS)*
11779 * Avoid filename truncation for various CA files.
11781 *Ben Laurie (CHATS)*
11783 * Use sizeof in preference to magic numbers.
11785 *Ben Laurie (CHATS)*
11787 * Avoid filename truncation in cert requests.
11789 *Ben Laurie (CHATS)*
11791 * Add assertions to check for (supposedly impossible) buffer
11794 *Ben Laurie (CHATS)*
11796 * Don't cache truncated DNS entries in the local cache (this could
11797 potentially lead to a spoofing attack).
11799 *Ben Laurie (CHATS)*
11801 * Fix various buffers to be large enough for hex/decimal
11802 representations in a platform independent manner.
11804 *Ben Laurie (CHATS)*
11806 * Add CRYPTO_realloc_clean() to avoid information leakage when
11807 resizing buffers containing secrets, and use where appropriate.
11809 *Ben Laurie (CHATS)*
11811 * Add BIO_indent() to avoid much slightly worrying code to do
11814 *Ben Laurie (CHATS)*
11816 * Convert sprintf()/BIO_puts() to BIO_printf().
11818 *Ben Laurie (CHATS)*
11820 * buffer_gets() could terminate with the buffer only half
11823 *Ben Laurie (CHATS)*
11825 * Add assertions to prevent user-supplied crypto functions from
11826 overflowing internal buffers by having large block sizes, etc.
11828 *Ben Laurie (CHATS)*
11830 * New OPENSSL_assert() macro (similar to assert(), but enabled
11833 *Ben Laurie (CHATS)*
11835 * Eliminate unused copy of key in RC4.
11837 *Ben Laurie (CHATS)*
11839 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11841 *Ben Laurie (CHATS)*
11843 * Fix off-by-one error in EGD path.
11845 *Ben Laurie (CHATS)*
11847 * If RANDFILE path is too long, ignore instead of truncating.
11849 *Ben Laurie (CHATS)*
11851 * Eliminate unused and incorrectly sized X.509 structure
11854 *Ben Laurie (CHATS)*
11856 * Eliminate unused and dangerous function knumber().
11858 *Ben Laurie (CHATS)*
11860 * Eliminate unused and dangerous structure, KSSL_ERR.
11862 *Ben Laurie (CHATS)*
11864 * Protect against overlong session ID context length in an encoded
11865 session object. Since these are local, this does not appear to be
11868 *Ben Laurie (CHATS)*
11870 * Change from security patch (see 0.9.6e below) that did not affect
11871 the 0.9.6 release series:
11873 Remote buffer overflow in SSL3 protocol - an attacker could
11874 supply an oversized master key in Kerberos-enabled versions.
11877 *Ben Laurie (CHATS)*
11879 * Change the SSL kerb5 codes to match RFC 2712.
11883 * Make -nameopt work fully for req and add -reqopt switch.
11885 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11887 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11889 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11891 * Make sure tests can be performed even if the corresponding algorithms
11892 have been removed entirely. This was also the last step to make
11893 OpenSSL compilable with DJGPP under all reasonable conditions.
11895 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11897 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11898 to allow version independent disabling of normally unselected ciphers,
11899 which may be activated as a side-effect of selecting a single cipher.
11901 (E.g., cipher list string "RSA" enables ciphersuites that are left
11902 out of "ALL" because they do not provide symmetric encryption.
11903 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11905 *Lutz Jaenicke, Bodo Moeller*
11907 * Add appropriate support for separate platform-dependent build
11908 directories. The recommended way to make a platform-dependent
11909 build directory is the following (tested on Linux), maybe with
11912 # Place yourself outside of the OpenSSL source tree. In
11913 # this example, the environment variable OPENSSL_SOURCE
11914 # is assumed to contain the absolute OpenSSL source directory.
11915 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11916 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11917 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11918 mkdir -p `dirname $F`
11919 ln -s $OPENSSL_SOURCE/$F $F
11922 To be absolutely sure not to disturb the source tree, a "make clean"
11923 is a good thing. If it isn't successful, don't worry about it,
11924 it probably means the source directory is very clean.
11928 * Make sure any ENGINE control commands make local copies of string
11929 pointers passed to them whenever necessary. Otherwise it is possible
11930 the caller may have overwritten (or deallocated) the original string
11931 data when a later ENGINE operation tries to use the stored values.
11933 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11935 * Improve diagnostics in file reading and command-line digests.
11937 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11939 * Add AES modes CFB and OFB to the object database. Correct an
11940 error in AES-CFB decryption.
11944 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11945 allows existing EVP_CIPHER_CTX structures to be reused after
11946 calling `EVP_*Final()`. This behaviour is used by encryption
11947 BIOs and some applications. This has the side effect that
11948 applications must explicitly clean up cipher contexts with
11949 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11953 * Check the values of dna and dnb in bn_mul_recursive before calling
11954 bn_mul_comba (a non zero value means the a or b arrays do not contain
11955 n2 elements) and fallback to bn_mul_normal if either is not zero.
11959 * Fix escaping of non-ASCII characters when using the -subj option
11960 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11964 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11965 form for "surname", serialNumber has no short form.
11966 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11967 therefore remove "mail" short name for "internet 7".
11968 The OID for unique identifiers in X509 certificates is
11969 x500UniqueIdentifier, not uniqueIdentifier.
11970 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11974 * Add an "init" command to the ENGINE config module and auto initialize
11975 ENGINEs. Without any "init" command the ENGINE will be initialized
11976 after all ctrl commands have been executed on it. If init=1 the
11977 ENGINE is initialized at that point (ctrls before that point are run
11978 on the uninitialized ENGINE and after on the initialized one). If
11979 init=0 then the ENGINE will not be initialized at all.
11983 * Fix the 'app_verify_callback' interface so that the user-defined
11984 argument is actually passed to the callback: In the
11985 SSL_CTX_set_cert_verify_callback() prototype, the callback
11986 declaration has been changed from
11989 int (*cb)(X509_STORE_CTX *,void *);
11990 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11991 i=s->ctx->app_verify_callback(&ctx)
11992 has been changed into
11993 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11995 To update applications using SSL_CTX_set_cert_verify_callback(),
11996 a dummy argument can be added to their callback functions.
11998 *D. K. Smetters <smetters@parc.xerox.com>*
12000 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12002 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12004 * Add and OPENSSL_LOAD_CONF define which will cause
12005 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12006 This allows older applications to transparently support certain
12007 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12008 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12009 load the config file and OPENSSL_add_all_algorithms_conf() which will
12010 always load it have also been added.
12014 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12015 Adjust NIDs and EVP layer.
12017 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12019 * Config modules support in openssl utility.
12021 Most commands now load modules from the config file,
12022 though in a few (such as version) this isn't done
12023 because it couldn't be used for anything.
12025 In the case of ca and req the config file used is
12026 the same as the utility itself: that is the -config
12027 command line option can be used to specify an
12032 * Move default behaviour from OPENSSL_config(). If appname is NULL
12033 use "openssl_conf" if filename is NULL use default openssl config file.
12037 * Add an argument to OPENSSL_config() to allow the use of an alternative
12038 config section name. Add a new flag to tolerate a missing config file
12039 and move code to CONF_modules_load_file().
12043 * Support for crypto accelerator cards from Accelerated Encryption
12044 Processing, www.aep.ie. (Use engine 'aep')
12045 The support was copied from 0.9.6c [engine] and adapted/corrected
12046 to work with the new engine framework.
12048 *AEP Inc. and Richard Levitte*
12050 * Support for SureWare crypto accelerator cards from Baltimore
12051 Technologies. (Use engine 'sureware')
12052 The support was copied from 0.9.6c [engine] and adapted
12053 to work with the new engine framework.
12057 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12058 make the newer ENGINE framework commands for the CHIL engine work.
12060 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12062 * Make it possible to produce shared libraries on ReliantUNIX.
12064 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12066 * Add the configuration target debug-linux-ppro.
12067 Make 'openssl rsa' use the general key loading routines
12068 implemented in `apps.c`, and make those routines able to
12069 handle the key format FORMAT_NETSCAPE and the variant
12072 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12074 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12076 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12078 * Add -keyform to rsautl, and document -engine.
12080 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12082 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12083 BIO_R_NO_SUCH_FILE error code rather than the generic
12084 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12088 * Add new functions
12089 ERR_peek_last_error
12090 ERR_peek_last_error_line
12091 ERR_peek_last_error_line_data.
12092 These are similar to
12094 ERR_peek_error_line
12095 ERR_peek_error_line_data,
12096 but report on the latest error recorded rather than the first one
12097 still in the error queue.
12099 *Ben Laurie, Bodo Moeller*
12101 * default_algorithms option in ENGINE config module. This allows things
12103 default_algorithms = ALL
12104 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12108 * Preliminary ENGINE config module.
12112 * New experimental application configuration code.
12116 * Change the AES code to follow the same name structure as all other
12117 symmetric ciphers, and behave the same way. Move everything to
12118 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12120 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12122 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12124 *Ben Laurie and Theo de Raadt*
12126 * Add option to output public keys in req command.
12128 *Massimiliano Pala madwolf@openca.org*
12130 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12131 (up to about 10% better than before for P-192 and P-224).
12135 * New functions/macros
12137 SSL_CTX_set_msg_callback(ctx, cb)
12138 SSL_CTX_set_msg_callback_arg(ctx, arg)
12139 SSL_set_msg_callback(ssl, cb)
12140 SSL_set_msg_callback_arg(ssl, arg)
12142 to request calling a callback function
12144 void cb(int write_p, int version, int content_type,
12145 const void *buf, size_t len, SSL *ssl, void *arg)
12147 whenever a protocol message has been completely received
12148 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12149 protocol version according to which the SSL library interprets
12150 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12151 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12152 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12153 specification (change_cipher_spec(20), alert(21), handshake(22)).
12154 'buf' and 'len' point to the actual message, 'ssl' to the
12155 SSL object, and 'arg' is the application-defined value set by
12156 SSL[_CTX]_set_msg_callback_arg().
12158 'openssl s_client' and 'openssl s_server' have new '-msg' options
12159 to enable a callback that displays all protocol messages.
12163 * Change the shared library support so shared libraries are built as
12164 soon as the corresponding static library is finished, and thereby get
12165 openssl and the test programs linked against the shared library.
12166 This still only happens when the keyword "shard" has been given to
12167 the configuration scripts.
12169 NOTE: shared library support is still an experimental thing, and
12170 backward binary compatibility is still not guaranteed.
12172 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12174 * Add support for Subject Information Access extension.
12176 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12178 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12179 additional bytes when new memory had to be allocated, not just
12180 when reusing an existing buffer.
12184 * New command line and configuration option 'utf8' for the req command.
12185 This allows field values to be specified as UTF8 strings.
12189 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12190 runs for the former and machine-readable output for the latter.
12194 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12195 of the e-mail address in the DN (i.e., it will go into a certificate
12196 extension only). The new configuration file option 'email_in_dn = no'
12197 has the same effect.
12199 *Massimiliano Pala madwolf@openca.org*
12201 * Change all functions with names starting with `des_` to be starting
12202 with `DES_` instead. Add wrappers that are compatible with libdes,
12203 but are named `_ossl_old_des_*`. Finally, add macros that map the
12204 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12205 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12206 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12209 Since we provide two compatibility mappings, the user needs to
12210 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12211 compatibility is desired. The default (i.e., when that macro
12212 isn't defined) is OpenSSL 0.9.6c compatibility.
12214 There are also macros that enable and disable the support of old
12215 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12216 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12217 are defined, the default will apply: to support the old des routines.
12219 In either case, one must include openssl/des.h to get the correct
12220 definitions. Do not try to just include openssl/des_old.h, that
12223 NOTE: This is a major break of an old API into a new one. Software
12224 authors are encouraged to switch to the `DES_` style functions. Some
12225 time in the future, des_old.h and the libdes compatibility functions
12226 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12227 default), and then completely removed.
12231 * Test for certificates which contain unsupported critical extensions.
12232 If such a certificate is found during a verify operation it is
12233 rejected by default: this behaviour can be overridden by either
12234 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12235 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12236 X509_supported_extension() has also been added which returns 1 if a
12237 particular extension is supported.
12241 * Modify the behaviour of EVP cipher functions in similar way to digests
12242 to retain compatibility with existing code.
12246 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12247 compatibility with existing code. In particular the 'ctx' parameter does
12248 not have to be to be initialized before the call to EVP_DigestInit() and
12249 it is tidied up after a call to EVP_DigestFinal(). New function
12250 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12251 EVP_MD_CTX_copy() changed to not require the destination to be
12252 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12253 requires the destination to be valid.
12255 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12256 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12260 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12261 so that complete 'Handshake' protocol structures are kept in memory
12262 instead of overwriting 'msg_type' and 'length' with 'body' data.
12266 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12268 *Massimo Santin via Richard Levitte*
12270 * Major restructuring to the underlying ENGINE code. This includes
12271 reduction of linker bloat, separation of pure "ENGINE" manipulation
12272 (initialisation, etc) from functionality dealing with implementations
12273 of specific crypto interfaces. This change also introduces integrated
12274 support for symmetric ciphers and digest implementations - so ENGINEs
12275 can now accelerate these by providing EVP_CIPHER and EVP_MD
12276 implementations of their own. This is detailed in
12277 [crypto/engine/README.md](crypto/engine/README.md)
12278 as it couldn't be adequately described here. However, there are a few
12279 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12280 were changed in the original introduction of ENGINE code have now
12281 reverted back - the hooking from this code to ENGINE is now a good
12282 deal more passive and at run-time, operations deal directly with
12283 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12284 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12285 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12286 they were not being used by the framework as there is no concept of a
12287 BIGNUM_METHOD and they could not be generalised to the new
12288 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12289 ENGINE_cpy() has been removed as it cannot be consistently defined in
12294 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12298 * Change mkdef.pl to sort symbols that get the same entry number,
12299 and make sure the automatically generated functions `ERR_load_*`
12300 become part of libeay.num as well.
12304 * New function SSL_renegotiate_pending(). This returns true once
12305 renegotiation has been requested (either SSL_renegotiate() call
12306 or HelloRequest/ClientHello received from the peer) and becomes
12307 false once a handshake has been completed.
12308 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12309 sends a HelloRequest, but does not ensure that a handshake takes
12310 place. SSL_renegotiate_pending() is useful for checking if the
12311 client has followed the request.)
12315 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12316 By default, clients may request session resumption even during
12317 renegotiation (if session ID contexts permit); with this option,
12318 session resumption is possible only in the first handshake.
12320 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12321 more bits available for options that should not be part of
12322 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12326 * Add some demos for certificate and certificate request creation.
12330 * Make maximum certificate chain size accepted from the peer application
12331 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12332 "Douglas E. Engert" <deengert@anl.gov>.
12336 * Add support for shared libraries for Unixware-7
12337 (Boyd Lynn Gerber <gerberb@zenez.com>).
12341 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12342 be done prior to destruction. Use this to unload error strings from
12343 ENGINEs that load their own error strings. NB: This adds two new API
12344 functions to "get" and "set" this destroy handler in an ENGINE.
12348 * Alter all existing ENGINE implementations (except "openssl" and
12349 "openbsd") to dynamically instantiate their own error strings. This
12350 makes them more flexible to be built both as statically-linked ENGINEs
12351 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12352 Also, add stub code to each that makes building them as self-contained
12353 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12357 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12358 implementations into applications that are completely implemented in
12359 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12360 commands that can be used to configure what shared-library to load and
12361 to control aspects of the way it is handled. Also, made an update to
12362 the [README-Engine.md](README-Engine.md) file
12363 that brings its information up-to-date and
12364 provides some information and instructions on the "dynamic" ENGINE
12365 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12369 * Make it possible to unload ranges of ERR strings with a new
12370 "ERR_unload_strings" function.
12374 * Add a copy() function to EVP_MD.
12378 * Make EVP_MD routines take a context pointer instead of just the
12379 md_data void pointer.
12383 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12384 that the digest can only process a single chunk of data
12385 (typically because it is provided by a piece of
12386 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12387 is only going to provide a single chunk of data, and hence the
12388 framework needn't accumulate the data for oneshot drivers.
12392 * As with "ERR", make it possible to replace the underlying "ex_data"
12393 functions. This change also alters the storage and management of global
12394 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12395 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12396 index counters. The API functions that use this state have been changed
12397 to take a "class_index" rather than pointers to the class's local STACK
12398 and counter, and there is now an API function to dynamically create new
12399 classes. This centralisation allows us to (a) plug a lot of the
12400 thread-safety problems that existed, and (b) makes it possible to clean
12401 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12402 such data would previously have always leaked in application code and
12403 workarounds were in place to make the memory debugging turn a blind eye
12404 to it. Application code that doesn't use this new function will still
12405 leak as before, but their memory debugging output will announce it now
12406 rather than letting it slide.
12408 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12409 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12410 has a return value to indicate success or failure.
12414 * Make it possible to replace the underlying "ERR" functions such that the
12415 global state (2 LHASH tables and 2 locks) is only used by the "default"
12416 implementation. This change also adds two functions to "get" and "set"
12417 the implementation prior to it being automatically set the first time
12418 any other ERR function takes place. Ie. an application can call "get",
12419 pass the return value to a module it has just loaded, and that module
12420 can call its own "set" function using that value. This means the
12421 module's "ERR" operations will use (and modify) the error state in the
12422 application and not in its own statically linked copy of OpenSSL code.
12426 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12427 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12428 the operation, and provides a more encapsulated way for external code
12429 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12430 to use these functions rather than manually incrementing the counts.
12432 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12436 * Add EVP test program.
12440 * Add symmetric cipher support to ENGINE. Expect the API to change!
12444 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12445 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12446 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12447 These allow a CRL to be built without having to access X509_CRL fields
12448 directly. Modify 'ca' application to use new functions.
12452 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12453 bug workarounds. Rollback attack detection is a security feature.
12454 The problem will only arise on OpenSSL servers when TLSv1 is not
12455 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12456 Software authors not wanting to support TLSv1 will have special reasons
12457 for their choice and can explicitly enable this option.
12459 *Bodo Moeller, Lutz Jaenicke*
12461 * Rationalise EVP so it can be extended: don't include a union of
12462 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12463 (similar to those existing for EVP_CIPHER_CTX).
12468 EVP_MD_CTX_init(&md); /* new function call */
12469 EVP_DigestInit(&md, EVP_sha1());
12470 EVP_DigestUpdate(&md, in, len);
12471 EVP_DigestFinal(&md, out, NULL);
12472 EVP_MD_CTX_cleanup(&md); /* new function call */
12476 * Make DES key schedule conform to the usual scheme, as well as
12477 correcting its structure. This means that calls to DES functions
12478 now have to pass a pointer to a des_key_schedule instead of a
12479 plain des_key_schedule (which was actually always a pointer
12482 des_key_schedule ks;
12484 des_set_key_checked(..., &ks);
12485 des_ncbc_encrypt(..., &ks, ...);
12487 (Note that a later change renames 'des_...' into 'DES_...'.)
12491 * Initial reduction of linker bloat: the use of some functions, such as
12492 PEM causes large amounts of unused functions to be linked in due to
12493 poor organisation. For example pem_all.c contains every PEM function
12494 which has a knock on effect of linking in large amounts of (unused)
12495 ASN1 code. Grouping together similar functions and splitting unrelated
12496 functions prevents this.
12500 * Cleanup of EVP macros.
12504 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12505 correct `_ecb suffix`.
12509 * Add initial OCSP responder support to ocsp application. The
12510 revocation information is handled using the text based index
12511 use by the ca application. The responder can either handle
12512 requests generated internally, supplied in files (for example
12513 via a CGI script) or using an internal minimal server.
12517 * Add configuration choices to get zlib compression for TLS.
12521 * Changes to Kerberos SSL for RFC 2712 compliance:
12522 1. Implemented real KerberosWrapper, instead of just using
12523 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12524 2. Implemented optional authenticator field of KerberosWrapper.
12526 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12527 and authenticator structs; see crypto/krb5/.
12529 Generalized Kerberos calls to support multiple Kerberos libraries.
12530 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12531 via Richard Levitte*
12533 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12534 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12535 values for each of the key sizes rather than having just
12536 parameters (and 'speed' generating keys each time).
12540 * Speed up EVP routines.
12543 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12544 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12545 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12546 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12548 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12549 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12550 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12553 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12555 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12559 * Added the OS2-EMX target.
12561 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12563 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12564 New functions to support `NCONF `routines in extension code.
12565 New function `CONF_set_nconf()`
12566 to allow functions which take an `NCONF` to also handle the old `LHASH`
12567 structure: this means that the old `CONF` compatible routines can be
12568 retained (in particular w.rt. extensions) without having to duplicate the
12569 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12573 * Enhance the general user interface with mechanisms for inner control
12574 and with possibilities to have yes/no kind of prompts.
12578 * Change all calls to low level digest routines in the library and
12579 applications to use EVP. Add missing calls to HMAC_cleanup() and
12580 don't assume HMAC_CTX can be copied using memcpy().
12582 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12584 * Add the possibility to control engines through control names but with
12585 arbitrary arguments instead of just a string.
12586 Change the key loaders to take a UI_METHOD instead of a callback
12587 function pointer. NOTE: this breaks binary compatibility with earlier
12588 versions of OpenSSL [engine].
12589 Adapt the nCipher code for these new conditions and add a card insertion
12594 * Enhance the general user interface with mechanisms to better support
12595 dialog box interfaces, application-defined prompts, the possibility
12596 to use defaults (for example default passwords from somewhere else)
12597 and interrupts/cancellations.
12601 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12602 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12606 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12607 tidy up some unnecessarily weird code in 'sk_new()').
12609 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12611 * Change the key loading routines for ENGINEs to use the same kind
12612 callback (pem_password_cb) as all other routines that need this
12617 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12618 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12619 than this minimum value is recommended.
12623 * New random seeder for OpenVMS, using the system process statistics
12624 that are easily reachable.
12628 * Windows apparently can't transparently handle global
12629 variables defined in DLLs. Initialisations such as:
12631 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12633 won't compile. This is used by the any applications that need to
12634 declare their own ASN1 modules. This was fixed by adding the option
12635 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12636 needed for static libraries under Win32.
12640 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12641 setting of purpose and trust fields. New X509_STORE trust and
12642 purpose functions and tidy up setting in other SSL functions.
12646 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12647 structure. These are inherited by X509_STORE_CTX when it is
12648 initialised. This allows various defaults to be set in the
12649 X509_STORE structure (such as flags for CRL checking and custom
12650 purpose or trust settings) for functions which only use X509_STORE_CTX
12651 internally such as S/MIME.
12653 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12654 trust settings if they are not set in X509_STORE. This allows X509_STORE
12655 purposes and trust (in S/MIME for example) to override any set by default.
12657 Add command line options for CRL checking to smime, s_client and s_server
12662 * Initial CRL based revocation checking. If the CRL checking flag(s)
12663 are set then the CRL is looked up in the X509_STORE structure and
12664 its validity and signature checked, then if the certificate is found
12665 in the CRL the verify fails with a revoked error.
12667 Various new CRL related callbacks added to X509_STORE_CTX structure.
12669 Command line options added to 'verify' application to support this.
12671 This needs some additional work, such as being able to handle multiple
12672 CRLs with different times, extension based lookup (rather than just
12673 by subject name) and ultimately more complete V2 CRL extension
12678 * Add a general user interface API (crypto/ui/). This is designed
12679 to replace things like des_read_password and friends (backward
12680 compatibility functions using this new API are provided).
12681 The purpose is to remove prompting functions from the DES code
12682 section as well as provide for prompting through dialog boxes in
12683 a window system and the like.
12687 * Add "ex_data" support to ENGINE so implementations can add state at a
12688 per-structure level rather than having to store it globally.
12692 * Make it possible for ENGINE structures to be copied when retrieved by
12693 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12694 This causes the "original" ENGINE structure to act like a template,
12695 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12696 operational state can be localised to each ENGINE structure, despite the
12697 fact they all share the same "methods". New ENGINE structures returned in
12698 this case have no functional references and the return value is the single
12699 structural reference. This matches the single structural reference returned
12700 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12705 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12706 needs to match any other type at all we need to manually clear the
12711 * Changes to the "openssl engine" utility to include;
12712 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12713 about an ENGINE's available control commands.
12714 - executing control commands from command line arguments using the
12715 '-pre' and '-post' switches. '-post' is only used if '-t' is
12716 specified and the ENGINE is successfully initialised. The syntax for
12717 the individual commands are colon-separated, for example;
12718 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12722 * New dynamic control command support for ENGINEs. ENGINEs can now
12723 declare their own commands (numbers), names (strings), descriptions,
12724 and input types for run-time discovery by calling applications. A
12725 subset of these commands are implicitly classed as "executable"
12726 depending on their input type, and only these can be invoked through
12727 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12728 can be based on user input, config files, etc). The distinction is
12729 that "executable" commands cannot return anything other than a boolean
12730 result and can only support numeric or string input, whereas some
12731 discoverable commands may only be for direct use through
12732 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12733 pointers, or other custom uses. The "executable" commands are to
12734 support parameterisations of ENGINE behaviour that can be
12735 unambiguously defined by ENGINEs and used consistently across any
12736 OpenSSL-based application. Commands have been added to all the
12737 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12738 control over shared-library paths without source code alterations.
12742 * Changed all ENGINE implementations to dynamically allocate their
12743 ENGINEs rather than declaring them statically. Apart from this being
12744 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12745 this also allows the implementations to compile without using the
12746 internal engine_int.h header.
12750 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12751 'const' value. Any code that should be able to modify a RAND_METHOD
12752 should already have non-const pointers to it (ie. they should only
12753 modify their own ones).
12757 * Made a variety of little tweaks to the ENGINE code.
12758 - "atalla" and "ubsec" string definitions were moved from header files
12759 to C code. "nuron" string definitions were placed in variables
12760 rather than hard-coded - allowing parameterisation of these values
12761 later on via ctrl() commands.
12762 - Removed unused "#if 0"'d code.
12763 - Fixed engine list iteration code so it uses ENGINE_free() to release
12764 structural references.
12765 - Constified the RAND_METHOD element of ENGINE structures.
12766 - Constified various get/set functions as appropriate and added
12767 missing functions (including a catch-all ENGINE_cpy that duplicates
12768 all ENGINE values onto a new ENGINE except reference counts/state).
12769 - Removed NULL parameter checks in get/set functions. Setting a method
12770 or function to NULL is a way of cancelling out a previously set
12771 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12772 and doesn't justify the extra error symbols and code.
12773 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12774 flags from engine_int.h to engine.h.
12775 - Changed prototypes for ENGINE handler functions (init(), finish(),
12776 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12780 * Implement binary inversion algorithm for BN_mod_inverse in addition
12781 to the algorithm using long division. The binary algorithm can be
12782 used only if the modulus is odd. On 32-bit systems, it is faster
12783 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12784 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12785 up to 450 bits. In 64-bit environments, the binary algorithm
12786 appears to be advantageous for much longer moduli; here we use it
12787 for moduli up to 2048 bits.
12791 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12792 could not support the combine flag in choice fields.
12796 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12797 extensions from a certificate request to the certificate.
12801 * Allow multiple 'certopt' and 'nameopt' options to be separated
12802 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12803 file: this allows the display of the certificate about to be
12804 signed to be customised, to allow certain fields to be included
12805 or excluded and extension details. The old system didn't display
12806 multicharacter strings properly, omitted fields not in the policy
12807 and couldn't display additional details such as extensions.
12811 * Function EC_POINTs_mul for multiple scalar multiplication
12812 of an arbitrary number of elliptic curve points
12813 \sum scalars[i]*points[i],
12814 optionally including the generator defined for the EC_GROUP:
12815 scalar*generator + \sum scalars[i]*points[i].
12817 EC_POINT_mul is a simple wrapper function for the typical case
12818 that the point list has just one item (besides the optional
12823 * First EC_METHODs for curves over GF(p):
12825 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12826 operations and provides various method functions that can also
12827 operate with faster implementations of modular arithmetic.
12829 EC_GFp_mont_method() reuses most functions that are part of
12830 EC_GFp_simple_method, but uses Montgomery arithmetic.
12832 *Bodo Moeller; point addition and point doubling
12833 implementation directly derived from source code provided by
12834 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12836 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12837 crypto/ec/ec_lib.c):
12839 Curves are EC_GROUP objects (with an optional group generator)
12840 based on EC_METHODs that are built into the library.
12842 Points are EC_POINT objects based on EC_GROUP objects.
12844 Most of the framework would be able to handle curves over arbitrary
12845 finite fields, but as there are no obvious types for fields other
12846 than GF(p), some functions are limited to that for now.
12850 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12851 that the file contains a complete HTTP response.
12855 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12856 change the def and num file printf format specifier from "%-40sXXX"
12857 to "%-39s XXX". The latter will always guarantee a space after the
12858 field while the former will cause them to run together if the field
12859 is 40 of more characters long.
12863 * Constify the cipher and digest 'method' functions and structures
12864 and modify related functions to take constant EVP_MD and EVP_CIPHER
12869 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12870 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12874 * Modify `EVP_Digest*()` routines so they now return values. Although the
12875 internal software routines can never fail additional hardware versions
12880 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12882 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12883 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12886 ERR_R_NESTED_ASN1_ERROR
12888 ERR_R_MISSING_ASN1_EOS
12889 were 4 .. 9, conflicting with
12890 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12892 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12893 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12895 Add new error code 'ERR_R_INTERNAL_ERROR'.
12899 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12904 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12905 sets the subject name for a new request or supersedes the
12906 subject name in a given request. Formats that can be parsed are
12907 'CN=Some Name, OU=myOU, C=IT'
12909 'CN=Some Name/OU=myOU/C=IT'.
12911 Add options '-batch' and '-verbose' to 'openssl req'.
12913 *Massimiliano Pala <madwolf@hackmasters.net>*
12915 * Introduce the possibility to access global variables through
12916 functions on platform were that's the best way to handle exporting
12917 global variables in shared libraries. To enable this functionality,
12918 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12919 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12920 is normally done by Configure or something similar).
12922 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12923 in the source file (foo.c) like this:
12925 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12926 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12928 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12929 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12931 OPENSSL_DECLARE_GLOBAL(int,foo);
12932 #define foo OPENSSL_GLOBAL_REF(foo)
12933 OPENSSL_DECLARE_GLOBAL(double,bar);
12934 #define bar OPENSSL_GLOBAL_REF(bar)
12936 The #defines are very important, and therefore so is including the
12937 header file everywhere where the defined globals are used.
12939 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12940 of ASN.1 items, but that structure is a bit different.
12942 The largest change is in util/mkdef.pl which has been enhanced with
12943 better and easier to understand logic to choose which symbols should
12944 go into the Windows .def files as well as a number of fixes and code
12945 cleanup (among others, algorithm keywords are now sorted
12946 lexicographically to avoid constant rewrites).
12950 * In BN_div() keep a copy of the sign of 'num' before writing the
12951 result to 'rm' because if rm==num the value will be overwritten
12952 and produce the wrong result if 'num' is negative: this caused
12953 problems with BN_mod() and BN_nnmod().
12957 * Function OCSP_request_verify(). This checks the signature on an
12958 OCSP request and verifies the signer certificate. The signer
12959 certificate is just checked for a generic purpose and OCSP request
12964 * Add OCSP_check_validity() function to check the validity of OCSP
12965 responses. OCSP responses are prepared in real time and may only
12966 be a few seconds old. Simply checking that the current time lies
12967 between thisUpdate and nextUpdate max reject otherwise valid responses
12968 caused by either OCSP responder or client clock inaccuracy. Instead
12969 we allow thisUpdate and nextUpdate to fall within a certain period of
12970 the current time. The age of the response can also optionally be
12971 checked. Two new options -validity_period and -status_age added to
12976 * If signature or public key algorithm is unrecognized print out its
12977 OID rather that just UNKNOWN.
12981 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12982 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12983 ID to be generated from the issuer certificate alone which can then be
12984 passed to OCSP_id_issuer_cmp().
12988 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12989 ASN1 modules to export functions returning ASN1_ITEM pointers
12990 instead of the ASN1_ITEM structures themselves. This adds several
12991 new macros which allow the underlying ASN1 function/structure to
12992 be accessed transparently. As a result code should not use ASN1_ITEM
12993 references directly (such as &X509_it) but instead use the relevant
12994 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12995 use of the new ASN1 code on platforms where exporting structures
12996 is problematical (for example in shared libraries) but exporting
12997 functions returning pointers to structures is not.
13001 * Add support for overriding the generation of SSL/TLS session IDs.
13002 These callbacks can be registered either in an SSL_CTX or per SSL.
13003 The purpose of this is to allow applications to control, if they wish,
13004 the arbitrary values chosen for use as session IDs, particularly as it
13005 can be useful for session caching in multiple-server environments. A
13006 command-line switch for testing this (and any client code that wishes
13007 to use such a feature) has been added to "s_server".
13009 *Geoff Thorpe, Lutz Jaenicke*
13011 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13012 of the form `#if defined(...) || defined(...) || ...` and
13013 `#if !defined(...) && !defined(...) && ...`. This also avoids
13014 the growing number of special cases it was previously handling.
13018 * Make all configuration macros available for application by making
13019 sure they are available in opensslconf.h, by giving them names starting
13020 with `OPENSSL_` to avoid conflicts with other packages and by making
13021 sure e_os2.h will cover all platform-specific cases together with
13023 Additionally, it is now possible to define configuration/platform-
13024 specific names (called "system identities"). In the C code, these
13025 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13026 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13027 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13032 * New option -set_serial to 'req' and 'x509' this allows the serial
13033 number to use to be specified on the command line. Previously self
13034 signed certificates were hard coded with serial number 0 and the
13035 CA options of 'x509' had to use a serial number in a file which was
13040 * New options to 'ca' utility to support V2 CRL entry extensions.
13041 Currently CRL reason, invalidity date and hold instruction are
13042 supported. Add new CRL extensions to V3 code and some new objects.
13046 * New function EVP_CIPHER_CTX_set_padding() this is used to
13047 disable standard block padding (aka PKCS#5 padding) in the EVP
13048 API, which was previously mandatory. This means that the data is
13049 not padded in any way and so the total length much be a multiple
13050 of the block size, otherwise an error occurs.
13054 * Initial (incomplete) OCSP SSL support.
13058 * New function OCSP_parse_url(). This splits up a URL into its host,
13059 port and path components: primarily to parse OCSP URLs. New -url
13060 option to ocsp utility.
13064 * New nonce behavior. The return value of OCSP_check_nonce() now
13065 reflects the various checks performed. Applications can decide
13066 whether to tolerate certain situations such as an absent nonce
13067 in a response when one was present in a request: the ocsp application
13068 just prints out a warning. New function OCSP_add1_basic_nonce()
13069 this is to allow responders to include a nonce in a response even if
13070 the request is nonce-less.
13074 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13075 skipped when using openssl x509 multiple times on a single input file,
13076 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13080 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13081 set string type: to handle setting ASN1_TIME structures. Fix ca
13082 utility to correctly initialize revocation date of CRLs.
13086 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13087 the clients preferred ciphersuites and rather use its own preferences.
13088 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13089 Internet Explorer by ensuring unchanged hash method during stepup.
13090 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13094 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13095 to aes and add a new 'exist' option to print out symbols that don't
13100 * Additional options to ocsp utility to allow flags to be set and
13101 additional certificates supplied.
13105 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13106 OCSP client a number of certificate to only verify the response
13111 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13112 handle the new API. Currently only ECB, CBC modes supported. Add new
13115 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13116 Encryption Standard (AES) Ciphersuites for Transport Layer
13117 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13118 not enabled by default and were not part of the "ALL" ciphersuite
13119 alias because they were not yet official; they could be
13120 explicitly requested by specifying the "AESdraft" ciphersuite
13121 group alias. In the final release of OpenSSL 0.9.7, the group
13122 alias is called "AES" and is part of "ALL".)
13124 *Ben Laurie, Steve Henson, Bodo Moeller*
13126 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13127 request to response.
13131 * Functions for OCSP responders. OCSP_request_onereq_count(),
13132 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13133 extract information from a certificate request. OCSP_response_create()
13134 creates a response and optionally adds a basic response structure.
13135 OCSP_basic_add1_status() adds a complete single response to a basic
13136 response and returns the OCSP_SINGLERESP structure just added (to allow
13137 extensions to be included for example). OCSP_basic_add1_cert() adds a
13138 certificate to a basic response and OCSP_basic_sign() signs a basic
13139 response with various flags. New helper functions ASN1_TIME_check()
13140 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13141 (converts ASN1_TIME to GeneralizedTime).
13145 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13146 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13147 structure from a certificate. X509_pubkey_digest() digests the public_key
13148 contents: this is used in various key identifiers.
13152 * Make sk_sort() tolerate a NULL argument.
13154 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13156 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13157 passed by the function are trusted implicitly. If any of them signed the
13158 response then it is assumed to be valid and is not verified.
13162 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13163 to data. This was previously part of the PKCS7 ASN1 code. This
13164 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13165 *Steve Henson, reported by Kenneth R. Robinette
13166 <support@securenetterm.com>*
13168 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13169 routines: without these tracing memory leaks is very painful.
13170 Fix leaks in PKCS12 and PKCS7 routines.
13174 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13175 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13176 effectively meant GeneralizedTime would never be used. Now it
13177 is initialised to -1 but X509_time_adj() now has to check the value
13178 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13179 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13180 *Steve Henson, reported by Kenneth R. Robinette
13181 <support@securenetterm.com>*
13183 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13184 result in a zero length in the ASN1_INTEGER structure which was
13185 not consistent with the structure when d2i_ASN1_INTEGER() was used
13186 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13187 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13188 where it did not print out a minus for negative ASN1_INTEGER.
13192 * Add summary printout to ocsp utility. The various functions which
13193 convert status values to strings have been renamed to:
13194 OCSP_response_status_str(), OCSP_cert_status_str() and
13195 OCSP_crl_reason_str() and are no longer static. New options
13196 to verify nonce values and to disable verification. OCSP response
13197 printout format cleaned up.
13201 * Add additional OCSP certificate checks. These are those specified
13202 in RFC2560. This consists of two separate checks: the CA of the
13203 certificate being checked must either be the OCSP signer certificate
13204 or the issuer of the OCSP signer certificate. In the latter case the
13205 OCSP signer certificate must contain the OCSP signing extended key
13206 usage. This check is performed by attempting to match the OCSP
13207 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13208 in the OCSP_CERTID structures of the response.
13212 * Initial OCSP certificate verification added to OCSP_basic_verify()
13213 and related routines. This uses the standard OpenSSL certificate
13214 verify routines to perform initial checks (just CA validity) and
13215 to obtain the certificate chain. Then additional checks will be
13216 performed on the chain. Currently the root CA is checked to see
13217 if it is explicitly trusted for OCSP signing. This is used to set
13218 a root CA as a global signing root: that is any certificate that
13219 chains to that CA is an acceptable OCSP signing certificate.
13223 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13224 extensions from a separate configuration file.
13225 As when reading extensions from the main configuration file,
13226 the '-extensions ...' option may be used for specifying the
13229 *Massimiliano Pala <madwolf@comune.modena.it>*
13231 * New OCSP utility. Allows OCSP requests to be generated or
13232 read. The request can be sent to a responder and the output
13233 parsed, outputted or printed in text form. Not complete yet:
13234 still needs to check the OCSP response validity.
13238 * New subcommands for 'openssl ca':
13239 `openssl ca -status <serial>` prints the status of the cert with
13240 the given serial number (according to the index file).
13241 `openssl ca -updatedb` updates the expiry status of certificates
13244 *Massimiliano Pala <madwolf@comune.modena.it>*
13246 * New '-newreq-nodes' command option to CA.pl. This is like
13247 '-newreq', but calls 'openssl req' with the '-nodes' option
13248 so that the resulting key is not encrypted.
13250 *Damien Miller <djm@mindrot.org>*
13252 * New configuration for the GNU Hurd.
13254 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13256 * Initial code to implement OCSP basic response verify. This
13257 is currently incomplete. Currently just finds the signer's
13258 certificate and verifies the signature on the response.
13262 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13263 value of OPENSSLDIR. This is available via the new '-d' option
13264 to 'openssl version', and is also included in 'openssl version -a'.
13268 * Allowing defining memory allocation callbacks that will be given
13269 file name and line number information in additional arguments
13270 (a `const char*` and an int). The basic functionality remains, as
13271 well as the original possibility to just replace malloc(),
13272 realloc() and free() by functions that do not know about these
13273 additional arguments. To register and find out the current
13274 settings for extended allocation functions, the following
13275 functions are provided:
13277 CRYPTO_set_mem_ex_functions
13278 CRYPTO_set_locked_mem_ex_functions
13279 CRYPTO_get_mem_ex_functions
13280 CRYPTO_get_locked_mem_ex_functions
13282 These work the same way as CRYPTO_set_mem_functions and friends.
13283 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13284 extended allocation function is enabled.
13285 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13286 a conventional allocation function is enabled.
13288 *Richard Levitte, Bodo Moeller*
13290 * Finish off removing the remaining LHASH function pointer casts.
13291 There should no longer be any prototype-casting required when using
13292 the LHASH abstraction, and any casts that remain are "bugs". See
13293 the callback types and macros at the head of lhash.h for details
13294 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13298 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13299 If /dev/[u]random devices are not available or do not return enough
13300 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13302 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13303 /etc/entropy will be queried once each in this sequence, querying stops
13304 when enough entropy was collected without querying more sockets.
13308 * Change the Unix RAND_poll() variant to be able to poll several
13309 random devices, as specified by DEVRANDOM, until a sufficient amount
13310 of data has been collected. We spend at most 10 ms on each file
13311 (select timeout) and read in non-blocking mode. DEVRANDOM now
13312 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13313 (previously it was just the string "/dev/urandom"), so on typical
13314 platforms the 10 ms delay will never occur.
13315 Also separate out the Unix variant to its own file, rand_unix.c.
13316 For VMS, there's a currently-empty rand_vms.c.
13320 * Move OCSP client related routines to ocsp_cl.c. These
13321 provide utility functions which an application needing
13322 to issue a request to an OCSP responder and analyse the
13323 response will typically need: as opposed to those which an
13324 OCSP responder itself would need which will be added later.
13326 OCSP_request_sign() signs an OCSP request with an API similar
13327 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13328 response. OCSP_response_get1_basic() extracts basic response
13329 from response. OCSP_resp_find_status(): finds and extracts status
13330 information from an OCSP_CERTID structure (which will be created
13331 when the request structure is built). These are built from lower
13332 level functions which work on OCSP_SINGLERESP structures but
13333 won't normally be used unless the application wishes to examine
13334 extensions in the OCSP response for example.
13336 Replace nonce routines with a pair of functions.
13337 OCSP_request_add1_nonce() adds a nonce value and optionally
13338 generates a random value. OCSP_check_nonce() checks the
13339 validity of the nonce in an OCSP response.
13343 * Change function OCSP_request_add() to OCSP_request_add0_id().
13344 This doesn't copy the supplied OCSP_CERTID and avoids the
13345 need to free up the newly created id. Change return type
13346 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13347 This can then be used to add extensions to the request.
13348 Deleted OCSP_request_new(), since most of its functionality
13349 is now in OCSP_REQUEST_new() (and the case insensitive name
13350 clash) apart from the ability to set the request name which
13351 will be added elsewhere.
13355 * Update OCSP API. Remove obsolete extensions argument from
13356 various functions. Extensions are now handled using the new
13357 OCSP extension code. New simple OCSP HTTP function which
13358 can be used to send requests and parse the response.
13362 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13363 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13364 uses the special reorder version of SET OF to sort the attributes
13365 and reorder them to match the encoded order. This resolves a long
13366 standing problem: a verify on a PKCS7 structure just after signing
13367 it used to fail because the attribute order did not match the
13368 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13369 it uses the received order. This is necessary to tolerate some broken
13370 software that does not order SET OF. This is handled by encoding
13371 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13372 to produce the required SET OF.
13376 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13377 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13378 files to get correct declarations of the ASN.1 item variables.
13382 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13383 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13384 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13385 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13386 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13387 ASN1_ITEM and no wrapper functions.
13391 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13392 replace the old function pointer based I/O routines. Change most of
13393 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13397 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13398 lines, recognize more "algorithms" that can be deselected, and make
13399 it complain about algorithm deselection that isn't recognised.
13403 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13404 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13405 to use new functions. Add NO_ASN1_OLD which can be set to remove
13406 some old style ASN1 functions: this can be used to determine if old
13407 code will still work when these eventually go away.
13411 * New extension functions for OCSP structures, these follow the
13412 same conventions as certificates and CRLs.
13416 * New function X509V3_add1_i2d(). This automatically encodes and
13417 adds an extension. Its behaviour can be customised with various
13418 flags to append, replace or delete. Various wrappers added for
13419 certificates and CRLs.
13423 * Fix to avoid calling the underlying ASN1 print routine when
13424 an extension cannot be parsed. Correct a typo in the
13425 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13429 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13430 entries for variables.
13434 * Add functionality to `apps/openssl.c` for detecting locking
13435 problems: As the program is single-threaded, all we have
13436 to do is register a locking callback using an array for
13437 storing which locks are currently held by the program.
13441 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13442 SSL_get_ex_data_X509_STORE_idx(), which is used in
13443 ssl_verify_cert_chain() and thus can be called at any time
13444 during TLS/SSL handshakes so that thread-safety is essential.
13445 Unfortunately, the ex_data design is not at all suited
13446 for multi-threaded use, so it probably should be abolished.
13450 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13452 *Broadcom, tweaked and integrated by Geoff Thorpe*
13454 * Move common extension printing code to new function
13455 X509V3_print_extensions(). Reorganise OCSP print routines and
13456 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13460 * New function X509_signature_print() to remove duplication in some
13465 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13466 set (this was treated exactly the same as SET OF previously). This
13467 is used to reorder the STACK representing the structure to match the
13468 encoding. This will be used to get round a problem where a PKCS7
13469 structure which was signed could not be verified because the STACK
13470 order did not reflect the encoded order.
13474 * Reimplement the OCSP ASN1 module using the new code.
13478 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13479 for its ASN1 operations. The old style function pointers still exist
13480 for now but they will eventually go away.
13484 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13485 completely replaces the old ASN1 functionality with a table driven
13486 encoder and decoder which interprets an ASN1_ITEM structure describing
13487 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13488 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13489 has also been converted to the new form.
13493 * Change BN_mod_exp_recp so that negative moduli are tolerated
13494 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13495 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13496 for negative moduli.
13500 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13501 of not touching the result's sign bit.
13505 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13510 * Changed the LHASH code to use prototypes for callbacks, and created
13511 macros to declare and implement thin (optionally static) functions
13512 that provide type-safety and avoid function pointer casting for the
13513 type-specific callbacks.
13517 * Added Kerberos Cipher Suites to be used with TLS, as written in
13519 *Veers Staats <staatsvr@asc.hpc.mil>,
13520 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13522 * Reformat the FAQ so the different questions and answers can be divided
13523 in sections depending on the subject.
13527 * Have the zlib compression code load ZLIB.DLL dynamically under
13532 * New function BN_mod_sqrt for computing square roots modulo a prime
13533 (using the probabilistic Tonelli-Shanks algorithm unless
13534 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13535 be handled deterministically).
13537 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13539 * Make BN_mod_inverse faster by explicitly handling small quotients
13540 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13541 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13545 * New function BN_kronecker.
13549 * Fix BN_gcd so that it works on negative inputs; the result is
13550 positive unless both parameters are zero.
13551 Previously something reasonably close to an infinite loop was
13552 possible because numbers could be growing instead of shrinking
13553 in the implementation of Euclid's algorithm.
13557 * Fix BN_is_word() and BN_is_one() macros to take into account the
13558 sign of the number in question.
13560 Fix BN_is_word(a,w) to work correctly for w == 0.
13562 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13563 because its test if the absolute value of 'a' equals 'w'.
13564 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13565 it exists mostly for use in the implementations of BN_is_zero(),
13566 BN_is_one(), and BN_is_word().
13570 * New function BN_swap.
13574 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13575 the exponentiation functions are more likely to produce reasonable
13576 results on negative inputs.
13580 * Change BN_mod_mul so that the result is always non-negative.
13581 Previously, it could be negative if one of the factors was negative;
13582 I don't think anyone really wanted that behaviour.
13586 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13587 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13588 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13589 and add new functions:
13598 BN_mod_lshift1_quick
13600 BN_mod_lshift_quick
13602 These functions always generate non-negative results.
13604 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13605 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13607 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13608 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13609 be reduced modulo `m`.
13611 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13614 The following entry accidentally appeared in the CHANGES file
13615 distributed with OpenSSL 0.9.7. The modifications described in
13616 it do *not* apply to OpenSSL 0.9.7.
13618 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13619 was actually never needed) and in BN_mul(). The removal in BN_mul()
13620 required a small change in bn_mul_part_recursive() and the addition
13621 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13622 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13623 bn_sub_words() and bn_add_words() except they take arrays with
13629 * In 'openssl passwd', verify passwords read from the terminal
13630 unless the '-salt' option is used (which usually means that
13631 verification would just waste user's time since the resulting
13632 hash is going to be compared with some given password hash)
13633 or the new '-noverify' option is used.
13635 This is an incompatible change, but it does not affect
13636 non-interactive use of 'openssl passwd' (passwords on the command
13637 line, '-stdin' option, '-in ...' option) and thus should not
13638 cause any problems.
13642 * Remove all references to RSAref, since there's no more need for it.
13646 * Make DSO load along a path given through an environment variable
13647 (SHLIB_PATH) with shl_load().
13651 * Constify the ENGINE code as a result of BIGNUM constification.
13652 Also constify the RSA code and most things related to it. In a
13653 few places, most notable in the depth of the ASN.1 code, ugly
13654 casts back to non-const were required (to be solved at a later
13659 * Make it so the openssl application has all engines loaded by default.
13663 * Constify the BIGNUM routines a little more.
13667 * Add the following functions:
13669 ENGINE_load_cswift()
13671 ENGINE_load_atalla()
13672 ENGINE_load_nuron()
13673 ENGINE_load_builtin_engines()
13675 That way, an application can itself choose if external engines that
13676 are built-in in OpenSSL shall ever be used or not. The benefit is
13677 that applications won't have to be linked with libdl or other dso
13678 libraries unless it's really needed.
13680 Changed 'openssl engine' to load all engines on demand.
13681 Changed the engine header files to avoid the duplication of some
13682 declarations (they differed!).
13686 * 'openssl engine' can now list capabilities.
13690 * Better error reporting in 'openssl engine'.
13694 * Never call load_dh_param(NULL) in s_server.
13698 * Add engine application. It can currently list engines by name and
13699 identity, and test if they are actually available.
13703 * Improve RPM specification file by forcing symbolic linking and making
13704 sure the installed documentation is also owned by root.root.
13706 *Damien Miller <djm@mindrot.org>*
13708 * Give the OpenSSL applications more possibilities to make use of
13709 keys (public as well as private) handled by engines.
13713 * Add OCSP code that comes from CertCo.
13717 * Add VMS support for the Rijndael code.
13721 * Added untested support for Nuron crypto accelerator.
13725 * Add support for external cryptographic devices. This code was
13726 previously distributed separately as the "engine" branch.
13728 *Geoff Thorpe, Richard Levitte*
13730 * Rework the filename-translation in the DSO code. It is now possible to
13731 have far greater control over how a "name" is turned into a filename
13732 depending on the operating environment and any oddities about the
13733 different shared library filenames on each system.
13737 * Support threads on FreeBSD-elf in Configure.
13741 * Fix for SHA1 assembly problem with MASM: it produces
13742 warnings about corrupt line number information when assembling
13743 with debugging information. This is caused by the overlapping
13746 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13749 NCONF_get_number() has no error checking at all. As a replacement,
13750 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13751 promoted strongly. The old NCONF_get_number is kept around for
13752 binary backward compatibility.
13753 Make it possible for methods to load from something other than a BIO,
13754 by providing a function pointer that is given a name instead of a BIO.
13755 For example, this could be used to load configuration data from an
13760 * Fix for non blocking accept BIOs. Added new I/O special reason
13761 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13762 with non blocking I/O was not possible because no retry code was
13763 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13768 * Added the beginnings of Rijndael support.
13772 * Fix for bug in DirectoryString mask setting. Add support for
13773 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13774 to allow certificate printing to more controllable, additional
13775 'certopt' option to 'x509' to allow new printing options to be
13780 * Clean old EAY MD5 hack from e_os.h.
13784 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13786 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13787 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13789 *Joe Orton, Steve Henson*
13791 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13793 * Fix additional bug revealed by the NISCC test suite:
13795 Stop bug triggering large recursion when presented with
13796 certain ASN.1 tags ([CVE-2003-0851])
13800 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13802 * Fix various bugs revealed by running the NISCC test suite:
13804 Stop out of bounds reads in the ASN1 code when presented with
13805 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13807 If verify callback ignores invalid public key errors don't try to check
13808 certificate signature with the NULL public key.
13812 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13813 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13818 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13819 extra data after the compression methods not only for TLS 1.0
13820 but also for SSL 3.0 (as required by the specification).
13822 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13824 * Change X509_certificate_type() to mark the key as exported/exportable
13825 when it's 512 *bits* long, not 512 bytes.
13829 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13831 * Countermeasure against the Klima-Pokorny-Rosa extension of
13832 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13833 a protocol version number mismatch like a decryption error
13834 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13838 * Turn on RSA blinding by default in the default implementation
13839 to avoid a timing attack. Applications that don't want it can call
13840 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13841 They would be ill-advised to do so in most cases.
13843 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13845 * Change RSA blinding code so that it works when the PRNG is not
13846 seeded (in this case, the secret RSA exponent is abused as
13847 an unpredictable seed -- if it is not unpredictable, there
13848 is no point in blinding anyway). Make RSA blinding thread-safe
13849 by remembering the creator's thread ID in rsa->blinding and
13850 having all other threads use local one-time blinding factors
13851 (this requires more computation than sharing rsa->blinding, but
13852 avoids excessive locking; and if an RSA object is not shared
13853 between threads, blinding will still be very fast).
13857 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13859 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13860 via timing by performing a MAC computation even if incorrect
13861 block cipher padding has been found. This is a countermeasure
13862 against active attacks where the attacker has to distinguish
13863 between bad padding and a MAC verification error. ([CVE-2003-0078])
13865 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13866 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13867 Martin Vuagnoux (EPFL, Ilion)*
13869 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13871 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13872 memory from its contents. This is done with a counter that will
13873 place alternating values in each byte. This can be used to solve
13874 two issues: 1) the removal of calls to memset() by highly optimizing
13875 compilers, and 2) cleansing with other values than 0, since those can
13876 be read through on certain media, for example a swap space on disk.
13880 * Bugfix: client side session caching did not work with external caching,
13881 because the session->cipher setting was not restored when reloading
13882 from the external cache. This problem was masked, when
13883 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13884 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13888 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13889 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13891 *Zeev Lieber <zeev-l@yahoo.com>*
13893 * Undo an undocumented change introduced in 0.9.6e which caused
13894 repeated calls to OpenSSL_add_all_ciphers() and
13895 OpenSSL_add_all_digests() to be ignored, even after calling
13900 * Change the default configuration reader to deal with last line not
13901 being properly terminated.
13905 * Change X509_NAME_cmp() so it applies the special rules on handling
13906 DN values that are of type PrintableString, as well as RDNs of type
13907 emailAddress where the value has the type ia5String.
13909 *stefank@valicert.com via Richard Levitte*
13911 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13912 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13913 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13914 the bitwise-OR of the two for use by the majority of applications
13915 wanting this behaviour, and update the docs. The documented
13916 behaviour and actual behaviour were inconsistent and had been
13917 changing anyway, so this is more a bug-fix than a behavioural
13920 *Geoff Thorpe, diagnosed by Nadav Har'El*
13922 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13923 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13927 * Fix initialization code race conditions in
13928 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13929 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13930 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13931 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13932 ssl2_get_cipher_by_char(),
13933 ssl3_get_cipher_by_char().
13935 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13937 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13938 the cached sessions are flushed, as the remove_cb() might use ex_data
13939 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13940 (see [openssl.org #212]).
13942 *Geoff Thorpe, Lutz Jaenicke*
13944 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13945 length, instead of the encoding length to d2i_ASN1_OBJECT.
13949 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13951 * [In 0.9.6g-engine release:]
13952 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13954 *Lynn Gazis <lgazis@rainbow.com>*
13956 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13958 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13959 and get fix the header length calculation.
13960 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13961 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13963 * Use proper error handling instead of 'assertions' in buffer
13964 overflow checks added in 0.9.6e. This prevents DoS (the
13965 assertions could call abort()).
13967 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13969 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13971 * Add various sanity checks to asn1_get_length() to reject
13972 the ASN1 length bytes if they exceed sizeof(long), will appear
13973 negative or the content length exceeds the length of the
13976 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13978 * Fix cipher selection routines: ciphers without encryption had no flags
13979 for the cipher strength set and where therefore not handled correctly
13980 by the selection routines (PR #130).
13984 * Fix EVP_dsa_sha macro.
13989 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13990 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13991 that was added in OpenSSL 0.9.6d.
13993 As the countermeasure turned out to be incompatible with some
13994 broken SSL implementations, the new option is part of SSL_OP_ALL.
13995 SSL_OP_ALL is usually employed when compatibility with weird SSL
13996 implementations is desired (e.g. '-bugs' option to 's_client' and
13997 's_server'), so the new option is automatically set in many
14002 * Changes in security patch:
14004 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14005 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14006 Air Force Materiel Command, USAF, under agreement number
14009 * Add various sanity checks to asn1_get_length() to reject
14010 the ASN1 length bytes if they exceed sizeof(long), will appear
14011 negative or the content length exceeds the length of the
14012 supplied buffer. ([CVE-2002-0659])
14014 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14016 * Assertions for various potential buffer overflows, not known to
14017 happen in practice.
14019 *Ben Laurie (CHATS)*
14021 * Various temporary buffers to hold ASCII versions of integers were
14022 too small for 64 bit platforms. ([CVE-2002-0655])
14023 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14025 * Remote buffer overflow in SSL3 protocol - an attacker could
14026 supply an oversized session ID to a client. ([CVE-2002-0656])
14028 *Ben Laurie (CHATS)*
14030 * Remote buffer overflow in SSL2 protocol - an attacker could
14031 supply an oversized client master key. ([CVE-2002-0656])
14033 *Ben Laurie (CHATS)*
14035 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14037 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14038 encoded as NULL) with id-dsa-with-sha1.
14040 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14042 * Check various `X509_...()` return values in `apps/req.c`.
14044 *Nils Larsch <nla@trustcenter.de>*
14046 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14047 an end-of-file condition would erroneously be flagged, when the CRLF
14048 was just at the end of a processed block. The bug was discovered when
14049 processing data through a buffering memory BIO handing the data to a
14050 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14051 <ptsekov@syntrex.com> and Nedelcho Stanev.
14055 * Implement a countermeasure against a vulnerability recently found
14056 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14057 before application data chunks to avoid the use of known IVs
14058 with data potentially chosen by the attacker.
14062 * Fix length checks in ssl3_get_client_hello().
14066 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14067 to prevent ssl3_read_internal() from incorrectly assuming that
14068 ssl3_read_bytes() found application data while handshake
14069 processing was enabled when in fact s->s3->in_read_app_data was
14070 merely automatically cleared during the initial handshake.
14072 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14074 * Fix object definitions for Private and Enterprise: they were not
14075 recognized in their shortname (=lowercase) representation. Extend
14076 obj_dat.pl to issue an error when using undefined keywords instead
14077 of silently ignoring the problem (Svenning Sorensen
14078 <sss@sss.dnsalias.net>).
14082 * Fix DH_generate_parameters() so that it works for 'non-standard'
14083 generators, i.e. generators other than 2 and 5. (Previously, the
14084 code did not properly initialise the 'add' and 'rem' values to
14085 BN_generate_prime().)
14087 In the new general case, we do not insist that 'generator' is
14088 actually a primitive root: This requirement is rather pointless;
14089 a generator of the order-q subgroup is just as good, if not
14094 * Map new X509 verification errors to alerts. Discovered and submitted by
14095 Tom Wu <tom@arcot.com>.
14099 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14100 returning non-zero before the data has been completely received
14101 when using non-blocking I/O.
14103 *Bodo Moeller; problem pointed out by John Hughes*
14105 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14107 *Ben Laurie, Lutz Jaenicke*
14109 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14110 Yoram Zahavi <YoramZ@gilian.com>).
14114 * Add information about CygWin 1.3 and on, and preserve proper
14115 configuration for the versions before that.
14117 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14119 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14120 check whether we deal with a copy of a session and do not delete from
14121 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14122 <izhar@checkpoint.com>.
14126 * Do not store session data into the internal session cache, if it
14127 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14128 flag is set). Proposed by Aslam <aslam@funk.com>.
14132 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14137 * [In 0.9.6d-engine release:]
14138 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14140 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14142 * Add the configuration target linux-s390x.
14144 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14146 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14147 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14148 variable as an indication that a ClientHello message has been
14149 received. As the flag value will be lost between multiple
14150 invocations of ssl3_accept when using non-blocking I/O, the
14151 function may not be aware that a handshake has actually taken
14152 place, thus preventing a new session from being added to the
14155 To avoid this problem, we now set s->new_session to 2 instead of
14156 using a local variable.
14158 *Lutz Jaenicke, Bodo Moeller*
14160 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14161 if the SSL_R_LENGTH_MISMATCH error is detected.
14163 *Geoff Thorpe, Bodo Moeller*
14165 * New 'shared_ldflag' column in Configure platform table.
14169 * Fix EVP_CIPHER_mode macro.
14171 *"Dan S. Camper" <dan@bti.net>*
14173 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14174 type, we must throw them away by setting rr->length to 0.
14176 *D P Chang <dpc@qualys.com>*
14178 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14180 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14181 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14182 worked incorrectly for those cases where range = `10..._2` and
14183 `3*range` is two bits longer than range.)
14187 * Only add signing time to PKCS7 structures if it is not already
14192 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14193 OBJ_ld_ce should be OBJ_id_ce.
14194 Also some ip-pda OIDs in crypto/objects/objects.txt were
14195 incorrect (cf. RFC 3039).
14197 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14199 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14200 returns early because it has nothing to do.
14202 *Andy Schneider <andy.schneider@bjss.co.uk>*
14204 * [In 0.9.6c-engine release:]
14205 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14207 *Andy Schneider <andy.schneider@bjss.co.uk>*
14209 * [In 0.9.6c-engine release:]
14210 Add support for Cryptographic Appliance's keyserver technology.
14211 (Use engine 'keyclient')
14213 *Cryptographic Appliances and Geoff Thorpe*
14215 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14216 is called via tools/c89.sh because arguments have to be
14217 rearranged (all '-L' options must appear before the first object
14220 *Richard Shapiro <rshapiro@abinitio.com>*
14222 * [In 0.9.6c-engine release:]
14223 Add support for Broadcom crypto accelerator cards, backported
14226 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14228 * [In 0.9.6c-engine release:]
14229 Add support for SureWare crypto accelerator cards from
14230 Baltimore Technologies. (Use engine 'sureware')
14232 *Baltimore Technologies and Mark Cox*
14234 * [In 0.9.6c-engine release:]
14235 Add support for crypto accelerator cards from Accelerated
14236 Encryption Processing, www.aep.ie. (Use engine 'aep')
14238 *AEP Inc. and Mark Cox*
14240 * Add a configuration entry for gcc on UnixWare.
14242 *Gary Benson <gbenson@redhat.com>*
14244 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14245 messages are stored in a single piece (fixed-length part and
14246 variable-length part combined) and fix various bugs found on the way.
14250 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14251 instead. BIO_gethostbyname() does not know what timeouts are
14252 appropriate, so entries would stay in cache even when they have
14254 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14256 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14257 faced with a pathologically small ClientHello fragment that does
14258 not contain client_version: Instead of aborting with an error,
14259 simply choose the highest available protocol version (i.e.,
14260 TLS 1.0 unless it is disabled). In practice, ClientHello
14261 messages are never sent like this, but this change gives us
14262 strictly correct behaviour at least for TLS.
14266 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14267 never resets s->method to s->ctx->method when called from within
14268 one of the SSL handshake functions.
14270 *Bodo Moeller; problem pointed out by Niko Baric*
14272 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14273 (sent using the client's version number) if client_version is
14274 smaller than the protocol version in use. Also change
14275 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14276 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14277 the client will at least see that alert.
14281 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14286 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14287 client receives HelloRequest while in a handshake.
14289 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14291 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14292 should end in 'break', not 'goto end' which circumvents various
14293 cleanups done in state SSL_ST_OK. But session related stuff
14294 must be disabled for SSL_ST_OK in the case that we just sent a
14297 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14298 before just sending a HelloRequest.
14300 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14302 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14303 reveal whether illegal block cipher padding was found or a MAC
14304 verification error occurred. (Neither SSLerr() codes nor alerts
14305 are directly visible to potential attackers, but the information
14306 may leak via logfiles.)
14308 Similar changes are not required for the SSL 2.0 implementation
14309 because the number of padding bytes is sent in clear for SSL 2.0,
14310 and the extra bytes are just ignored. However ssl/s2_pkt.c
14311 failed to verify that the purported number of padding bytes is in
14316 * Add OpenUNIX-8 support including shared libraries
14317 (Boyd Lynn Gerber <gerberb@zenez.com>).
14321 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14322 'wristwatch attack' using huge encoding parameters (cf.
14323 James H. Manger's CRYPTO 2001 paper). Note that the
14324 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14325 encoding parameters and hence was not vulnerable.
14329 * BN_sqr() bug fix.
14331 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14333 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14334 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14335 followed by modular reduction.
14337 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14339 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14340 equivalent based on BN_pseudo_rand() instead of BN_rand().
14344 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14345 This function was broken, as the check for a new client hello message
14346 to handle SGC did not allow these large messages.
14347 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14351 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14355 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14356 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14360 * Rework the configuration and shared library support for Tru64 Unix.
14361 The configuration part makes use of modern compiler features and
14362 still retains old compiler behavior for those that run older versions
14363 of the OS. The shared library support part includes a variant that
14364 uses the RPATH feature, and is available through the special
14365 configuration target "alpha-cc-rpath", which will never be selected
14368 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14370 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14371 with the same message size as in ssl3_get_certificate_request().
14372 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14373 messages might inadvertently be reject as too long.
14375 *Petr Lampa <lampa@fee.vutbr.cz>*
14377 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14381 * Modified SSL library such that the verify_callback that has been set
14382 specifically for an SSL object with SSL_set_verify() is actually being
14383 used. Before the change, a verify_callback set with this function was
14384 ignored and the verify_callback() set in the SSL_CTX at the time of
14385 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14386 to allow the necessary settings.
14390 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14391 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14392 done automatically (in contradiction to the requirements of the C
14393 standard). This made problems when used from OpenSSH.
14397 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14398 dh->length and always used
14400 BN_rand_range(priv_key, dh->p).
14402 BN_rand_range() is not necessary for Diffie-Hellman, and this
14403 specific range makes Diffie-Hellman unnecessarily inefficient if
14404 dh->length (recommended exponent length) is much smaller than the
14405 length of dh->p. We could use BN_rand_range() if the order of
14406 the subgroup was stored in the DH structure, but we only have
14411 BN_rand(priv_key, l, ...)
14413 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14420 RSA_eay_public_encrypt
14421 RSA_eay_private_decrypt
14422 RSA_eay_private_encrypt (signing)
14423 RSA_eay_public_decrypt (signature verification)
14425 (default implementations for RSA_public_encrypt,
14426 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14427 always reject numbers >= n.
14431 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14432 to synchronize access to 'locking_thread'. This is necessary on
14433 systems where access to 'locking_thread' (an 'unsigned long'
14434 variable) is not atomic.
14438 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14439 *before* setting the 'crypto_lock_rand' flag. The previous code had
14440 a race condition if 0 is a valid thread ID.
14442 *Travis Vitek <vitek@roguewave.com>*
14444 * Add support for shared libraries under Irix.
14446 *Albert Chin-A-Young <china@thewrittenword.com>*
14448 * Add configuration option to build on Linux on both big-endian and
14449 little-endian MIPS.
14451 *Ralf Baechle <ralf@uni-koblenz.de>*
14453 * Add the possibility to create shared libraries on HP-UX.
14457 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14459 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14460 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14461 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14462 PRNG state recovery was possible based on the output of
14463 one PRNG request appropriately sized to gain knowledge on
14464 'md' followed by enough consecutive 1-byte PRNG requests
14465 to traverse all of 'state'.
14467 1. When updating 'md_local' (the current thread's copy of 'md')
14468 during PRNG output generation, hash all of the previous
14469 'md_local' value, not just the half used for PRNG output.
14471 2. Make the number of bytes from 'state' included into the hash
14472 independent from the number of PRNG bytes requested.
14474 The first measure alone would be sufficient to avoid
14475 Markku-Juhani's attack. (Actually it had never occurred
14476 to me that the half of 'md_local' used for chaining was the
14477 half from which PRNG output bytes were taken -- I had always
14478 assumed that the secret half would be used.) The second
14479 measure makes sure that additional data from 'state' is never
14480 mixed into 'md_local' in small portions; this heuristically
14481 further strengthens the PRNG.
14485 * Fix crypto/bn/asm/mips3.s.
14489 * When only the key is given to "enc", the IV is undefined. Print out
14490 an error message in this case.
14494 * Handle special case when X509_NAME is empty in X509 printing routines.
14498 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14499 positive and less than q.
14503 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14504 used: it isn't thread safe and the add_lock_callback should handle
14507 *Paul Rose <Paul.Rose@bridge.com>*
14509 * Verify that incoming data obeys the block size in
14510 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14516 *Ulf Möller, Bodo Möller*
14518 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14519 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14520 when fixing the server behaviour for backwards-compatible 'client
14521 hello' messages. (Note that the attack is impractical against
14522 SSL 3.0 and TLS 1.0 anyway because length and version checking
14523 means that the probability of guessing a valid ciphertext is
14524 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14527 Before 0.9.5, the countermeasure (hide the error by generating a
14528 random 'decryption result') did not work properly because
14529 ERR_clear_error() was missing, meaning that SSL_get_error() would
14530 detect the supposedly ignored error.
14532 Both problems are now fixed.
14536 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14537 (previously it was 1024).
14541 * Fix for compatibility mode trust settings: ignore trust settings
14542 unless some valid trust or reject settings are present.
14546 * Fix for blowfish EVP: its a variable length cipher.
14550 * Fix various bugs related to DSA S/MIME verification. Handle missing
14551 parameters in DSA public key structures and return an error in the
14552 DSA routines if parameters are absent.
14556 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14557 in the current directory if neither $RANDFILE nor $HOME was set.
14558 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14559 caused some confusion to Windows users who haven't defined $HOME.
14560 Thus RAND_file_name() is changed again: e_os.h can define a
14561 DEFAULT_HOME, which will be used if $HOME is not set.
14562 For Windows, we use "C:"; on other platforms, we still require
14563 environment variables.
14565 * Move 'if (!initialized) RAND_poll()' into regions protected by
14566 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14567 having multiple threads call RAND_poll() concurrently.
14571 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14572 combination of a flag and a thread ID variable.
14573 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14574 flag), *other* threads can enter ssleay_add_bytes without obeying
14575 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14576 that they do not hold after the first thread unsets add_do_not_lock).
14580 * Change bctest again: '-x' expressions are not available in all
14581 versions of 'test'.
14585 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14587 * Fix a couple of memory leaks in PKCS7_dataDecode()
14589 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14591 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14592 the default extension for executables, if any. Also, make the perl
14593 scripts that use symlink() to test if it really exists and use "cp"
14594 if it doesn't. All this made OpenSSL compilable and installable in
14599 * Fix for asn1_GetSequence() for indefinite length constructed data.
14600 If SEQUENCE is length is indefinite just set c->slen to the total
14601 amount of data available.
14603 *Steve Henson, reported by shige@FreeBSD.org*
14605 *This change does not apply to 0.9.7.*
14607 * Change bctest to avoid here-documents inside command substitution
14608 (workaround for FreeBSD /bin/sh bug).
14609 For compatibility with Ultrix, avoid shell functions (introduced
14610 in the bctest version that searches along $PATH).
14614 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14615 with des_encrypt() defined on some operating systems, like Solaris
14620 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14621 On the Importance of Eliminating Errors in Cryptographic
14622 Computations, J. Cryptology 14 (2001) 2, 101-119,
14623 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14627 * MIPS assembler BIGNUM division bug fix.
14631 * Disabled incorrect Alpha assembler code.
14635 * Fix PKCS#7 decode routines so they correctly update the length
14636 after reading an EOC for the EXPLICIT tag.
14640 *This change does not apply to 0.9.7.*
14642 * Fix bug in PKCS#12 key generation routines. This was triggered
14643 if a 3DES key was generated with a 0 initial byte. Include
14644 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14645 (but broken) behaviour.
14649 * Enhance bctest to search for a working bc along $PATH and print
14652 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14654 * Fix memory leaks in err.c: free err_data string if necessary;
14655 don't write to the wrong index in ERR_set_error_data.
14659 * Implement ssl23_peek (analogous to ssl23_read), which previously
14664 * Replace rdtsc with `_emit` statements for VC++ version 5.
14666 *Jeremy Cooper <jeremy@baymoo.org>*
14668 * Make it possible to reuse SSLv2 sessions.
14672 * In copy_email() check for >= 0 as a return value for
14673 X509_NAME_get_index_by_NID() since 0 is a valid index.
14675 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14677 * Avoid coredump with unsupported or invalid public keys by checking if
14678 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14679 PKCS7_verify() fails with non detached data.
14683 * Don't use getenv in library functions when run as setuid/setgid.
14684 New function OPENSSL_issetugid().
14688 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14689 due to incorrect handling of multi-threading:
14691 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14693 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14695 3. Count how many times MemCheck_off() has been called so that
14696 nested use can be treated correctly. This also avoids
14697 inband-signalling in the previous code (which relied on the
14698 assumption that thread ID 0 is impossible).
14702 * Add "-rand" option also to s_client and s_server.
14706 * Fix CPU detection on Irix 6.x.
14707 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14708 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14710 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14715 *This change does not apply to 0.9.7.*
14717 * Use the cached encoding of an X509_NAME structure rather than
14718 copying it. This is apparently the reason for the libsafe "errors"
14719 but the code is actually correct.
14723 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14724 Bleichenbacher's DSA attack.
14725 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14726 to be set and top=0 forces the highest bit to be set; top=-1 is new
14727 and leaves the highest bit random.
14729 *Ulf Moeller, Bodo Moeller*
14731 * In the `NCONF_...`-based implementations for `CONF_...` queries
14732 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14733 a temporary CONF structure with the data component set to NULL
14734 (which gives segmentation faults in lh_retrieve).
14735 Instead, use NULL for the CONF pointer in CONF_get_string and
14736 CONF_get_number (which may use environment variables) and directly
14737 return NULL from CONF_get_section.
14741 * Fix potential buffer overrun for EBCDIC.
14745 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14746 keyUsage if basicConstraints absent for a CA.
14750 * Make SMIME_write_PKCS7() write mail header values with a format that
14751 is more generally accepted (no spaces before the semicolon), since
14752 some programs can't parse those values properly otherwise. Also make
14753 sure BIO's that break lines after each write do not create invalid
14758 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14759 macros previously used would not encode an empty SEQUENCE OF
14760 and break the signature.
14764 *This change does not apply to 0.9.7.*
14766 * Zero the premaster secret after deriving the master secret in
14771 * Add some EVP_add_digest_alias registrations (as found in
14772 OpenSSL_add_all_digests()) to SSL_library_init()
14773 aka OpenSSL_add_ssl_algorithms(). This provides improved
14774 compatibility with peers using X.509 certificates
14775 with unconventional AlgorithmIdentifier OIDs.
14779 * Fix for Irix with NO_ASM.
14781 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14783 * ./config script fixes.
14785 *Ulf Moeller, Richard Levitte*
14787 * Fix 'openssl passwd -1'.
14791 * Change PKCS12_key_gen_asc() so it can cope with non null
14792 terminated strings whose length is passed in the passlen
14793 parameter, for example from PEM callbacks. This was done
14794 by adding an extra length parameter to asc2uni().
14796 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14798 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14799 call failed, free the DSA structure.
14803 * Fix to uni2asc() to cope with zero length Unicode strings.
14804 These are present in some PKCS#12 files.
14808 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14809 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14810 when writing a 32767 byte record.
14812 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14814 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14815 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14817 (RSA objects have a reference count access to which is protected
14818 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14819 so they are meant to be shared between threads.)
14820 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14821 "Reddie, Steven" <Steven.Reddie@ca.com>*
14823 * Fix a deadlock in CRYPTO_mem_leaks().
14827 * Use better test patterns in bntest.
14831 * rand_win.c fix for Borland C.
14835 * BN_rshift bugfix for n == 0.
14839 * Add a 'bctest' script that checks for some known 'bc' bugs
14840 so that 'make test' does not abort just because 'bc' is broken.
14844 * Store verify_result within SSL_SESSION also for client side to
14845 avoid potential security hole. (Re-used sessions on the client side
14846 always resulted in verify_result==X509_V_OK, not using the original
14847 result of the server certificate verification.)
14851 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14852 SSL3_RT_APPLICATION_DATA, return 0.
14853 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14858 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14859 releases, have been re-implemented by renaming the previous
14860 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14861 and ssl3_read_internal, respectively, and adding 'peek' parameters
14862 to them. The new ssl[23]_{read,peek} functions are calls to
14863 ssl[23]_read_internal with the 'peek' flag set appropriately.
14864 A 'peek' parameter has also been added to ssl3_read_bytes, which
14865 does the actual work for ssl3_read_internal.
14869 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14870 the method-specific "init()" handler. Also clean up ex_data after
14871 calling the method-specific "finish()" handler. Previously, this was
14872 happening the other way round.
14876 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14877 The previous value, 12, was not always sufficient for BN_mod_exp().
14881 * Make sure that shared libraries get the internal name engine with
14882 the full version number and not just 0. This should mark the
14883 shared libraries as not backward compatible. Of course, this should
14884 be changed again when we can guarantee backward binary compatibility.
14888 * Fix typo in get_cert_by_subject() in by_dir.c
14890 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14892 * Rework the system to generate shared libraries:
14894 - Make note of the expected extension for the shared libraries and
14895 if there is a need for symbolic links from for example libcrypto.so.0
14896 to libcrypto.so.0.9.7. There is extended info in Configure for
14899 - Make as few rebuilds of the shared libraries as possible.
14901 - Still avoid linking the OpenSSL programs with the shared libraries.
14903 - When installing, install the shared libraries separately from the
14908 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14910 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14911 and not in SSL_clear because the latter is also used by the
14912 accept/connect functions; previously, the settings made by
14913 SSL_set_read_ahead would be lost during the handshake.
14915 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14917 * Correct util/mkdef.pl to be selective about disabled algorithms.
14918 Previously, it would create entries for disabled algorithms no
14923 * Added several new manual pages for SSL_* function.
14927 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14929 * In ssl23_get_client_hello, generate an error message when faced
14930 with an initial SSL 3.0/TLS record that is too small to contain the
14931 first two bytes of the ClientHello message, i.e. client_version.
14932 (Note that this is a pathologic case that probably has never happened
14933 in real life.) The previous approach was to use the version number
14934 from the record header as a substitute; but our protocol choice
14935 should not depend on that one because it is not authenticated
14936 by the Finished messages.
14940 * More robust randomness gathering functions for Windows.
14942 *Jeffrey Altman <jaltman@columbia.edu>*
14944 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14945 not set then we don't setup the error code for issuer check errors
14946 to avoid possibly overwriting other errors which the callback does
14947 handle. If an application does set the flag then we assume it knows
14948 what it is doing and can handle the new informational codes
14953 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14954 a general "ANY" type, as such it should be able to decode anything
14955 including tagged types. However it didn't check the class so it would
14956 wrongly interpret tagged types in the same way as their universal
14957 counterpart and unknown types were just rejected. Changed so that the
14958 tagged and unknown types are handled in the same way as a SEQUENCE:
14959 that is the encoding is stored intact. There is also a new type
14960 "V_ASN1_OTHER" which is used when the class is not universal, in this
14961 case we have no idea what the actual type is so we just lump them all
14966 * On VMS, stdout may very well lead to a file that is written to
14967 in a record-oriented fashion. That means that every write() will
14968 write a separate record, which will be read separately by the
14969 programs trying to read from it. This can be very confusing.
14971 The solution is to put a BIO filter in the way that will buffer
14972 text until a linefeed is reached, and then write everything a
14973 line at a time, so every record written will be an actual line,
14974 not chunks of lines and not (usually doesn't happen, but I've
14975 seen it once) several lines in one record. BIO_f_linebuffer() is
14978 Currently, it's a VMS-only method, because that's where it has
14979 been tested well enough.
14983 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14984 it can return incorrect results.
14985 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14986 but it was in 0.9.6-beta[12].)
14990 * Disable the check for content being present when verifying detached
14991 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14992 include zero length content when signing messages.
14996 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14997 BIO_ctrl (for BIO pairs).
15001 * Add DSO method for VMS.
15005 * Bug fix: Montgomery multiplication could produce results with the
15010 * Add RPM specification openssl.spec and modify it to build three
15011 packages. The default package contains applications, application
15012 documentation and run-time libraries. The devel package contains
15013 include files, static libraries and function documentation. The
15014 doc package contains the contents of the doc directory. The original
15015 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15019 * Add a large number of documentation files for many SSL routines.
15021 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15023 * Add a configuration entry for Sony News 4.
15025 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15027 * Don't set the two most significant bits to one when generating a
15028 random number < q in the DSA library.
15032 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15033 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15034 the underlying transport is blocking) if a handshake took place.
15035 (The default behaviour is needed by applications such as s_client
15036 and s_server that use select() to determine when to use SSL_read;
15037 but for applications that know in advance when to expect data, it
15038 just makes things more complicated.)
15042 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15047 * Add a few more EBCDIC conditionals that make `req` and `x509`
15048 work better on such systems.
15050 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15052 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15053 Update PKCS12_parse() so it copies the friendlyName and the
15054 keyid to the certificates aux info.
15058 * Fix bug in PKCS7_verify() which caused an infinite loop
15059 if there was more than one signature.
15061 *Sven Uszpelkat <su@celocom.de>*
15063 * Major change in util/mkdef.pl to include extra information
15064 about each symbol, as well as presenting variables as well
15065 as functions. This change means that there's n more need
15066 to rebuild the .num files when some algorithms are excluded.
15070 * Allow the verify time to be set by an application,
15071 rather than always using the current time.
15075 * Phase 2 verify code reorganisation. The certificate
15076 verify code now looks up an issuer certificate by a
15077 number of criteria: subject name, authority key id
15078 and key usage. It also verifies self signed certificates
15079 by the same criteria. The main comparison function is
15080 X509_check_issued() which performs these checks.
15082 Lot of changes were necessary in order to support this
15083 without completely rewriting the lookup code.
15085 Authority and subject key identifier are now cached.
15087 The LHASH 'certs' is X509_STORE has now been replaced
15088 by a STACK_OF(X509_OBJECT). This is mainly because an
15089 LHASH can't store or retrieve multiple objects with
15090 the same hash value.
15092 As a result various functions (which were all internal
15093 use only) have changed to handle the new X509_STORE
15094 structure. This will break anything that messed round
15095 with X509_STORE internally.
15097 The functions X509_STORE_add_cert() now checks for an
15098 exact match, rather than just subject name.
15100 The X509_STORE API doesn't directly support the retrieval
15101 of multiple certificates matching a given criteria, however
15102 this can be worked round by performing a lookup first
15103 (which will fill the cache with candidate certificates)
15104 and then examining the cache for matches. This is probably
15105 the best we can do without throwing out X509_LOOKUP
15106 entirely (maybe later...).
15108 The X509_VERIFY_CTX structure has been enhanced considerably.
15110 All certificate lookup operations now go via a get_issuer()
15111 callback. Although this currently uses an X509_STORE it
15112 can be replaced by custom lookups. This is a simple way
15113 to bypass the X509_STORE hackery necessary to make this
15114 work and makes it possible to use more efficient techniques
15115 in future. A very simple version which uses a simple
15116 STACK for its trusted certificate store is also provided
15117 using X509_STORE_CTX_trusted_stack().
15119 The verify_cb() and verify() callbacks now have equivalents
15120 in the X509_STORE_CTX structure.
15122 X509_STORE_CTX also has a 'flags' field which can be used
15123 to customise the verify behaviour.
15127 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15128 excludes S/MIME capabilities.
15132 * When a certificate request is read in keep a copy of the
15133 original encoding of the signed data and use it when outputting
15134 again. Signatures then use the original encoding rather than
15135 a decoded, encoded version which may cause problems if the
15136 request is improperly encoded.
15140 * For consistency with other BIO_puts implementations, call
15141 buffer_write(b, ...) directly in buffer_puts instead of calling
15144 In BIO_puts, increment b->num_write as in BIO_write.
15146 *Peter.Sylvester@EdelWeb.fr*
15148 * Fix BN_mul_word for the case where the word is 0. (We have to use
15149 BN_zero, we may not return a BIGNUM with an array consisting of
15150 words set to zero.)
15154 * Avoid calling abort() from within the library when problems are
15155 detected, except if preprocessor symbols have been defined
15156 (such as REF_CHECK, BN_DEBUG etc.).
15160 * New openssl application 'rsautl'. This utility can be
15161 used for low level RSA operations. DER public key
15162 BIO/fp routines also added.
15166 * New Configure entry and patches for compiling on QNX 4.
15168 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15170 * A demo state-machine implementation was sponsored by
15171 Nuron (<http://www.nuron.com/>) and is now available in
15172 demos/state_machine.
15176 * New options added to the 'dgst' utility for signature
15177 generation and verification.
15181 * Unrecognized PKCS#7 content types are now handled via a
15182 catch all ASN1_TYPE structure. This allows unsupported
15183 types to be stored as a "blob" and an application can
15184 encode and decode it manually.
15188 * Fix various signed/unsigned issues to make a_strex.c
15189 compile under VC++.
15191 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15193 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15194 length if passed a buffer. ASN1_INTEGER_to_BN failed
15195 if passed a NULL BN and its argument was negative.
15197 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15199 * Modification to PKCS#7 encoding routines to output definite
15200 length encoding. Since currently the whole structures are in
15201 memory there's not real point in using indefinite length
15202 constructed encoding. However if OpenSSL is compiled with
15203 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15207 * Added BIO_vprintf() and BIO_vsnprintf().
15211 * Added more prefixes to parse for in the strings written
15212 through a logging bio, to cover all the levels that are available
15213 through syslog. The prefixes are now:
15215 PANIC, EMERG, EMR => LOG_EMERG
15216 ALERT, ALR => LOG_ALERT
15217 CRIT, CRI => LOG_CRIT
15218 ERROR, ERR => LOG_ERR
15219 WARNING, WARN, WAR => LOG_WARNING
15220 NOTICE, NOTE, NOT => LOG_NOTICE
15221 INFO, INF => LOG_INFO
15222 DEBUG, DBG => LOG_DEBUG
15224 and as before, if none of those prefixes are present at the
15225 beginning of the string, LOG_ERR is chosen.
15227 On Win32, the `LOG_*` levels are mapped according to this:
15229 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15230 LOG_WARNING => EVENTLOG_WARNING_TYPE
15231 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15235 * Made it possible to reconfigure with just the configuration
15236 argument "reconf" or "reconfigure". The command line arguments
15237 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15238 and are retrieved from there when reconfiguring.
15244 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15246 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15250 * The obj_dat.pl script was messing up the sorting of object
15251 names. The reason was that it compared the quoted version
15252 of strings as a result "OCSP" > "OCSP Signing" because
15253 " > SPACE. Changed script to store unquoted versions of
15254 names and add quotes on output. It was also omitting some
15255 names from the lookup table if they were given a default
15256 value (that is if SN is missing it is given the same
15257 value as LN and vice versa), these are now added on the
15258 grounds that if an object has a name we should be able to
15259 look it up. Finally added warning output when duplicate
15260 short or long names are found.
15264 * Changes needed for Tandem NSK.
15266 *Scott Uroff <scott@xypro.com>*
15268 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15269 RSA_padding_check_SSLv23(), special padding was never detected
15270 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15271 version rollback attacks was not effective.
15273 In s23_clnt.c, don't use special rollback-attack detection padding
15274 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15275 client; similarly, in s23_srvr.c, don't do the rollback check if
15276 SSL 2.0 is the only protocol enabled in the server.
15280 * Make it possible to get hexdumps of unprintable data with 'openssl
15281 asn1parse'. By implication, the functions ASN1_parse_dump() and
15282 BIO_dump_indent() are added.
15286 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15287 these print out strings and name structures based on various
15288 flags including RFC2253 support and proper handling of
15289 multibyte characters. Added options to the 'x509' utility
15290 to allow the various flags to be set.
15294 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15295 Also change the functions X509_cmp_current_time() and
15296 X509_gmtime_adj() work with an ASN1_TIME structure,
15297 this will enable certificates using GeneralizedTime in validity
15298 dates to be checked.
15302 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15303 negative public key encodings) on by default,
15304 NO_NEG_PUBKEY_BUG can be set to disable it.
15308 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15309 content octets. An i2c_ASN1_OBJECT is unnecessary because
15310 the encoding can be trivially obtained from the structure.
15314 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15315 not read locks (`CRYPTO_r_[un]lock`).
15319 * A first attempt at creating official support for shared
15320 libraries through configuration. I've kept it so the
15321 default is static libraries only, and the OpenSSL programs
15322 are always statically linked for now, but there are
15323 preparations for dynamic linking in place.
15324 This has been tested on Linux and Tru64.
15328 * Randomness polling function for Win9x, as described in:
15329 Peter Gutmann, Software Generation of Practically Strong
15334 * Fix so PRNG is seeded in req if using an already existing
15339 * New options to smime application. -inform and -outform
15340 allow alternative formats for the S/MIME message including
15341 PEM and DER. The -content option allows the content to be
15342 specified separately. This should allow things like Netscape
15343 form signing output easier to verify.
15347 * Fix the ASN1 encoding of tags using the 'long form'.
15351 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15352 STRING types. These convert content octets to and from the
15353 underlying type. The actual tag and length octets are
15354 already assumed to have been read in and checked. These
15355 are needed because all other string types have virtually
15356 identical handling apart from the tag. By having versions
15357 of the ASN1 functions that just operate on content octets
15358 IMPLICIT tagging can be handled properly. It also allows
15359 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15360 and ASN1_INTEGER are identical apart from the tag.
15364 * Change the handling of OID objects as follows:
15366 - New object identifiers are inserted in objects.txt, following
15367 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15368 - objects.pl is used to process obj_mac.num and create a new
15370 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15373 This is currently kind of a hack, and the perl code in objects.pl
15374 isn't very elegant, but it works as I intended. The simplest way
15375 to check that it worked correctly is to look in obj_dat.h and
15376 check the array nid_objs and make sure the objects haven't moved
15377 around (this is important!). Additions are OK, as well as
15378 consistent name changes.
15382 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15386 * Addition of the command line parameter '-rand file' to 'openssl req'.
15387 The given file adds to whatever has already been seeded into the
15388 random pool through the RANDFILE configuration file option or
15389 environment variable, or the default random state file.
15393 * mkstack.pl now sorts each macro group into lexical order.
15394 Previously the output order depended on the order the files
15395 appeared in the directory, resulting in needless rewriting
15400 * Patches to make OpenSSL compile under Win32 again. Mostly
15401 work arounds for the VC++ problem that it treats func() as
15402 func(void). Also stripped out the parts of mkdef.pl that
15403 added extra typesafe functions: these no longer exist.
15407 * Reorganisation of the stack code. The macros are now all
15408 collected in safestack.h . Each macro is defined in terms of
15409 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15410 DEBUG_SAFESTACK is now handled in terms of function casts,
15411 this has the advantage of retaining type safety without the
15412 use of additional functions. If DEBUG_SAFESTACK is not defined
15413 then the non typesafe macros are used instead. Also modified the
15414 mkstack.pl script to handle the new form. Needs testing to see
15415 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15416 the default if no major problems. Similar behaviour for ASN1_SET_OF
15417 and PKCS12_STACK_OF.
15421 * When some versions of IIS use the 'NET' form of private key the
15422 key derivation algorithm is different. Normally MD5(password) is
15423 used as a 128 bit RC4 key. In the modified case
15424 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15425 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15426 as the old Netscape_RSA functions except they have an additional
15427 'sgckey' parameter which uses the modified algorithm. Also added
15428 an -sgckey command line option to the rsa utility. Thanks to
15429 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15430 algorithm to openssl-dev.
15434 * The evp_local.h macros were using 'c.##kname' which resulted in
15435 invalid expansion on some systems (SCO 5.0.5 for example).
15436 Corrected to 'c.kname'.
15438 *Phillip Porch <root@theporch.com>*
15440 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15441 a STACK of email addresses from a certificate or request, these look
15442 in the subject name and the subject alternative name extensions and
15443 omit any duplicate addresses.
15447 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15448 This makes DSA verification about 2 % faster.
15452 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15453 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15454 plus overhead for 1024 bit moduli).
15455 This makes exponentiations about 0.5 % faster for 1024 bit
15456 exponents (as measured by "openssl speed rsa2048").
15460 * Rename memory handling macros to avoid conflicts with other
15462 Malloc => OPENSSL_malloc
15463 Malloc_locked => OPENSSL_malloc_locked
15464 Realloc => OPENSSL_realloc
15465 Free => OPENSSL_free
15469 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15470 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15474 * CygWin32 support.
15476 *John Jarvie <jjarvie@newsguy.com>*
15478 * The type-safe stack code has been rejigged. It is now only compiled
15479 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15480 by default all type-specific stack functions are "#define"d back to
15481 standard stack functions. This results in more streamlined output
15482 but retains the type-safety checking possibilities of the original
15487 * The STACK code has been cleaned up, and certain type declarations
15488 that didn't make a lot of sense have been brought in line. This has
15489 also involved a cleanup of sorts in safestack.h to more correctly
15490 map type-safe stack functions onto their plain stack counterparts.
15491 This work has also resulted in a variety of "const"ifications of
15492 lots of the code, especially `_cmp` operations which should normally
15493 be prototyped with "const" parameters anyway.
15497 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15498 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15499 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15500 where all of 'md' is used each time the PRNG is used, but 'state'
15501 is used only indexed by a cyclic counter. As entropy may not be
15502 well distributed from the beginning, 'md' is important as a
15503 chaining variable. However, the output function chains only half
15504 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15505 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15506 in all of 'state' being rewritten, with the new values depending
15507 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15511 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15512 the handshake is continued after ssl_verify_cert_chain();
15513 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15514 can lead to 'unexplainable' connection aborts later.
15516 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15518 * Major EVP API cipher revision.
15519 Add hooks for extra EVP features. This allows various cipher
15520 parameters to be set in the EVP interface. Support added for variable
15521 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15522 setting of RC2 and RC5 parameters.
15524 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15527 Remove lots of duplicated code from the EVP library. For example *every*
15528 cipher init() function handles the 'iv' in the same way according to the
15529 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15530 for CFB and OFB modes they zero ctx->num.
15532 New functionality allows removal of S/MIME code RC2 hack.
15534 Most of the routines have the same form and so can be declared in terms
15537 By shifting this to the top level EVP_CipherInit() it can be removed from
15538 all individual ciphers. If the cipher wants to handle IVs or keys
15539 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15542 Change lots of functions like EVP_EncryptUpdate() to now return a
15543 value: although software versions of the algorithms cannot fail
15544 any installed hardware versions can.
15548 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15549 this option is set, tolerate broken clients that send the negotiated
15550 protocol version number instead of the requested protocol version
15555 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15556 i.e. non-zero for export ciphersuites, zero otherwise.
15557 Previous versions had this flag inverted, inconsistent with
15558 rsa_tmp_cb (..._TMP_RSA_CB).
15560 *Bodo Moeller; problem reported by Amit Chopra*
15562 * Add missing DSA library text string. Work around for some IIS
15563 key files with invalid SEQUENCE encoding.
15567 * Add a document (doc/standards.txt) that list all kinds of standards
15568 and so on that are implemented in OpenSSL.
15572 * Enhance c_rehash script. Old version would mishandle certificates
15573 with the same subject name hash and wouldn't handle CRLs at all.
15574 Added -fingerprint option to crl utility, to support new c_rehash
15579 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15583 * Fix for SSL server purpose checking. Server checking was
15584 rejecting certificates which had extended key usage present
15585 but no ssl client purpose.
15587 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15589 * Make PKCS#12 code work with no password. The PKCS#12 spec
15590 is a little unclear about how a blank password is handled.
15591 Since the password in encoded as a BMPString with terminating
15592 double NULL a zero length password would end up as just the
15593 double NULL. However no password at all is different and is
15594 handled differently in the PKCS#12 key generation code. NS
15595 treats a blank password as zero length. MSIE treats it as no
15596 password on export: but it will try both on import. We now do
15597 the same: PKCS12_parse() tries zero length and no password if
15598 the password is set to "" or NULL (NULL is now a valid password:
15599 it wasn't before) as does the pkcs12 application.
15603 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15604 perror when PEM_read_bio_X509_REQ fails, the error message must
15605 be obtained from the error queue.
15609 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15610 it in ERR_remove_state if appropriate, and change ERR_get_state
15611 accordingly to avoid race conditions (this is necessary because
15612 thread_hash is no longer constant once set).
15616 * Bugfix for linux-elf makefile.one.
15620 * RSA_get_default_method() will now cause a default
15621 RSA_METHOD to be chosen if one doesn't exist already.
15622 Previously this was only set during a call to RSA_new()
15623 or RSA_new_method(NULL) meaning it was possible for
15624 RSA_get_default_method() to return NULL.
15628 * Added native name translation to the existing DSO code
15629 that will convert (if the flag to do so is set) filenames
15630 that are sufficiently small and have no path information
15631 into a canonical native form. Eg. "blah" converted to
15632 "libblah.so" or "blah.dll" etc.
15636 * New function ERR_error_string_n(e, buf, len) which is like
15637 ERR_error_string(e, buf), but writes at most 'len' bytes
15638 including the 0 terminator. For ERR_error_string_n, 'buf'
15641 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15643 * CONF library reworked to become more general. A new CONF
15644 configuration file reader "class" is implemented as well as a
15645 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15646 old `CONF_*` functions are still there, but are reimplemented to
15647 work in terms of the new functions. Also, a set of functions
15648 to handle the internal storage of the configuration data is
15649 provided to make it easier to write new configuration file
15650 reader "classes" (I can definitely see something reading a
15651 configuration file in XML format, for example), called `_CONF_*`,
15652 or "the configuration storage API"...
15654 The new configuration file reading functions are:
15656 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15657 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15659 NCONF_default, NCONF_WIN32
15661 NCONF_dump_fp, NCONF_dump_bio
15663 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15664 NCONF_new creates a new CONF object. This works in the same way
15665 as other interfaces in OpenSSL, like the BIO interface.
15666 `NCONF_dump_*` dump the internal storage of the configuration file,
15667 which is useful for debugging. All other functions take the same
15668 arguments as the old `CONF_*` functions with the exception of the
15669 first that must be a `CONF *` instead of a `LHASH *`.
15671 To make it easier to use the new classes with the old `CONF_*` functions,
15672 the function CONF_set_default_method is provided.
15676 * Add '-tls1' option to 'openssl ciphers', which was already
15677 mentioned in the documentation but had not been implemented.
15678 (This option is not yet really useful because even the additional
15679 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15683 * Initial DSO code added into libcrypto for letting OpenSSL (and
15684 OpenSSL-based applications) load shared libraries and bind to
15685 them in a portable way.
15687 *Geoff Thorpe, with contributions from Richard Levitte*
15689 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15691 * Make sure _lrotl and _lrotr are only used with MSVC.
15693 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15694 (the default implementation of RAND_status).
15696 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15697 to '-clrext' (= clear extensions), as intended and documented.
15698 *Bodo Moeller; inconsistency pointed out by Michael Attili
15699 <attili@amaxo.com>*
15701 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15702 was larger than the MD block size.
15704 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15706 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15707 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15708 using the passed key: if the passed key was a private key the result
15709 of X509_print(), for example, would be to print out all the private key
15714 * des_quad_cksum() byte order bug fix.
15715 *Ulf Möller, using the problem description in krb4-0.9.7, where
15716 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15718 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15721 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15723 * For easily testing in shell scripts whether some command
15724 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15725 returns with exit code 0 iff no command of the given name is available.
15726 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15727 the output goes to stdout and nothing is printed to stderr.
15728 Additional arguments are always ignored.
15730 Since for each cipher there is a command of the same name,
15731 the 'no-cipher' compilation switches can be tested this way.
15733 ('openssl no-XXX' is not able to detect pseudo-commands such
15734 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15738 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15742 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15743 is set; it will be thrown away anyway because each handshake creates
15745 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15746 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15747 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15748 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15752 * New s_client option -ign_eof: EOF at stdin is ignored, and
15753 'Q' and 'R' lose their special meanings (quit/renegotiate).
15754 This is part of what -quiet does; unlike -quiet, -ign_eof
15755 does not suppress any output.
15759 * Add compatibility options to the purpose and trust code. The
15760 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15761 accepts a certificate or CA, this was the previous behaviour,
15762 with all the associated security issues.
15764 X509_TRUST_COMPAT is the old trust behaviour: only and
15765 automatically trust self signed roots in certificate store. A
15766 new trust setting X509_TRUST_DEFAULT is used to specify that
15767 a purpose has no associated trust setting and it should instead
15768 use the value in the default purpose.
15772 * Fix the PKCS#8 DSA private key code so it decodes keys again
15773 and fix a memory leak.
15777 * In util/mkerr.pl (which implements 'make errors'), preserve
15778 reason strings from the previous version of the .c file, as
15779 the default to have only downcase letters (and digits) in
15780 automatically generated reasons codes is not always appropriate.
15784 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15785 using strerror. Previously, ERR_reason_error_string() returned
15786 library names as reason strings for SYSerr; but SYSerr is a special
15787 case where small numbers are errno values, not library numbers.
15791 * Add '-dsaparam' option to 'openssl dhparam' application. This
15792 converts DSA parameters into DH parameters. (When creating parameters,
15793 DSA_generate_parameters is used.)
15797 * Include 'length' (recommended exponent length) in C code generated
15798 by 'openssl dhparam -C'.
15802 * The second argument to set_label in perlasm was already being used
15803 so couldn't be used as a "file scope" flag. Moved to third argument
15808 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15809 instead of RAND_bytes for encryption IVs and salts.
15813 * Include RAND_status() into RAND_METHOD instead of implementing
15814 it only for md_rand.c Otherwise replacing the PRNG by calling
15815 RAND_set_rand_method would be impossible.
15819 * Don't let DSA_generate_key() enter an infinite loop if the random
15820 number generation fails.
15824 * New 'rand' application for creating pseudo-random output.
15828 * Added configuration support for Linux/IA64
15830 *Rolf Haberrecker <rolf@suse.de>*
15832 * Assembler module support for Mingw32.
15836 * Shared library support for HPUX (in shlib/).
15838 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15840 * Shared library support for Solaris gcc.
15842 *Lutz Behnke <behnke@trustcenter.de>*
15844 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15846 * PKCS7_encrypt() was adding text MIME headers twice because they
15847 were added manually and by SMIME_crlf_copy().
15851 * In bntest.c don't call BN_rand with zero bits argument.
15853 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15855 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15856 case was implemented. This caused BN_div_recp() to fail occasionally.
15860 * Add an optional second argument to the set_label() in the perl
15861 assembly language builder. If this argument exists and is set
15862 to 1 it signals that the assembler should use a symbol whose
15863 scope is the entire file, not just the current function. This
15864 is needed with MASM which uses the format label:: for this scope.
15866 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15868 * Change the ASN1 types so they are typedefs by default. Before
15869 almost all types were #define'd to ASN1_STRING which was causing
15870 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15875 * Change names of new functions to the new get1/get0 naming
15876 convention: After 'get1', the caller owns a reference count
15877 and has to call `..._free`; 'get0' returns a pointer to some
15878 data structure without incrementing reference counters.
15879 (Some of the existing 'get' functions increment a reference
15880 counter, some don't.)
15881 Similarly, 'set1' and 'add1' functions increase reference
15882 counters or duplicate objects.
15886 * Allow for the possibility of temp RSA key generation failure:
15887 the code used to assume it always worked and crashed on failure.
15891 * Fix potential buffer overrun problem in BIO_printf().
15892 *Ulf Möller, using public domain code by Patrick Powell; problem
15893 pointed out by David Sacerdote <das33@cornell.edu>*
15895 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15896 RAND_egd() and RAND_status(). In the command line application,
15897 the EGD socket can be specified like a seed file using RANDFILE
15902 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15903 Some CAs (e.g. Verisign) distribute certificates in this form.
15907 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15908 list to exclude them. This means that no special compilation option
15909 is needed to use anonymous DH: it just needs to be included in the
15914 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15915 EVP_MD_type. The old functionality is available in a new macro called
15916 EVP_MD_md(). Change code that uses it and update docs.
15920 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15921 where the `void *` argument is replaced by a function pointer argument.
15922 Previously `void *` was abused to point to functions, which works on
15923 many platforms, but is not correct. As these functions are usually
15924 called by macros defined in OpenSSL header files, most source code
15925 should work without changes.
15929 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15930 sections with information on -D... compiler switches used for
15931 compiling the library so that applications can see them. To enable
15932 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15933 must be defined. E.g.,
15934 #define OPENSSL_ALGORITHM_DEFINES
15935 #include <openssl/opensslconf.h>
15936 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15938 *Richard Levitte, Ulf and Bodo Möller*
15940 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15945 * Change the 'other' type in certificate aux info to a STACK_OF
15946 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15947 the required ASN1 format: arbitrary types determined by an OID.
15951 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15952 argument to 'req'. This is not because the function is newer or
15953 better than others it just uses the work 'NEW' in the certificate
15954 request header lines. Some software needs this.
15958 * Reorganise password command line arguments: now passwords can be
15959 obtained from various sources. Delete the PEM_cb function and make
15960 it the default behaviour: i.e. if the callback is NULL and the
15961 usrdata argument is not NULL interpret it as a null terminated pass
15962 phrase. If usrdata and the callback are NULL then the pass phrase
15963 is prompted for as usual.
15967 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15968 the support is automatically enabled. The resulting binaries will
15969 autodetect the card and use it if present.
15971 *Ben Laurie and Compaq Inc.*
15973 * Work around for Netscape hang bug. This sends certificate request
15974 and server done in one record. Since this is perfectly legal in the
15975 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15976 the bugs/SSLv3 entry for more info.
15980 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15984 * Add -rand argument to smime and pkcs12 applications and read/write
15989 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15993 * Add command line password options to the remaining applications.
15997 * Bug fix for BN_div_recp() for numerators with an even number of
16002 * More tests in bntest.c, and changed test_bn output.
16006 * ./config recognizes MacOS X now.
16010 * Bug fix for BN_div() when the first words of num and divisor are
16011 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16015 * Add support for various broken PKCS#8 formats, and command line
16016 options to produce them.
16020 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16021 get temporary BIGNUMs from a BN_CTX.
16025 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16030 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16031 include a #define from the old name to the new. The original intent
16032 was that statically linked binaries could for example just call
16033 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16034 link with digests. This never worked because SSLeay_add_all_digests()
16035 and SSLeay_add_all_ciphers() were in the same source file so calling
16036 one would link with the other. They are now in separate source files.
16040 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16044 * Use a less unusual form of the Miller-Rabin primality test (it used
16045 a binary algorithm for exponentiation integrated into the Miller-Rabin
16046 loop, our standard modexp algorithms are faster).
16050 * Support for the EBCDIC character set completed.
16052 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16054 * Source code cleanups: use const where appropriate, eliminate casts,
16055 use `void *` instead of `char *` in lhash.
16059 * Bugfix: ssl3_send_server_key_exchange was not restartable
16060 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16061 this the server could overwrite ephemeral keys that the client
16066 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16067 using 50 iterations of the Rabin-Miller test.
16069 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16070 iterations of the Rabin-Miller test as required by the appendix
16071 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16072 As BN_is_prime_fasttest includes trial division, DSA parameter
16073 generation becomes much faster.
16075 This implies a change for the callback functions in DSA_is_prime
16076 and DSA_generate_parameters: The callback function is called once
16077 for each positive witness in the Rabin-Miller test, not just
16078 occasionally in the inner loop; and the parameters to the
16079 callback function now provide an iteration count for the outer
16080 loop rather than for the current invocation of the inner loop.
16081 DSA_generate_parameters additionally can call the callback
16082 function with an 'iteration count' of -1, meaning that a
16083 candidate has passed the trial division test (when q is generated
16084 from an application-provided seed, trial division is skipped).
16088 * New function BN_is_prime_fasttest that optionally does trial
16089 division before starting the Rabin-Miller test and has
16090 an additional BN_CTX * argument (whereas BN_is_prime always
16091 has to allocate at least one BN_CTX).
16092 'callback(1, -1, cb_arg)' is called when a number has passed the
16093 trial division stage.
16097 * Fix for bug in CRL encoding. The validity dates weren't being handled
16102 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16106 * New function BN_pseudo_rand().
16110 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16111 bignum version of BN_from_montgomery() with the working code from
16112 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16117 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16118 made it impossible to use the same SSL_SESSION data structure in
16119 SSL2 clients in multiple threads.
16123 * The return value of RAND_load_file() no longer counts bytes obtained
16124 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16125 to seed the PRNG (previously an explicit byte count was required).
16127 *Ulf Möller, Bodo Möller*
16129 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16130 used `char *` instead of `void *` and had casts all over the place.
16134 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16138 * Retain source code compatibility for BN_prime_checks macro:
16139 BN_is_prime(..., BN_prime_checks, ...) now uses
16140 BN_prime_checks_for_size to determine the appropriate number of
16141 Rabin-Miller iterations.
16145 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16146 DH_CHECK_P_NOT_SAFE_PRIME.
16147 (Check if this is true? OpenPGP calls them "strong".)
16151 * Merge the functionality of "dh" and "gendh" programs into a new program
16152 "dhparam". The old programs are retained for now but will handle DH keys
16153 (instead of parameters) in future.
16157 * Make the ciphers, s_server and s_client programs check the return values
16158 when a new cipher list is set.
16162 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16163 ciphers. Before when the 56bit ciphers were enabled the sorting was
16166 The syntax for the cipher sorting has been extended to support sorting by
16167 cipher-strength (using the strength_bits hard coded in the tables).
16168 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16170 Fix a bug in the cipher-command parser: when supplying a cipher command
16171 string with an "undefined" symbol (neither command nor alphanumeric
16172 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16173 an error is flagged.
16175 Due to the strength-sorting extension, the code of the
16176 ssl_create_cipher_list() function was completely rearranged. I hope that
16177 the readability was also increased :-)
16179 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16181 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16182 for the first serial number and places 2 in the serial number file. This
16183 avoids problems when the root CA is created with serial number zero and
16184 the first user certificate has the same issuer name and serial number
16189 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16190 the new code. Add documentation for this stuff.
16194 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16195 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16196 structures and behave in an analogous way to the X509v3 functions:
16197 they shouldn't be called directly but wrapper functions should be used
16200 So we also now have some wrapper functions that call the X509at functions
16201 when passed certificate requests. (TO DO: similar things can be done with
16202 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16203 things. Some of these need some d2i or i2d and print functionality
16204 because they handle more complex structures.)
16208 * Add missing #ifndefs that caused missing symbols when building libssl
16209 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16210 NO_RSA in `ssl/s2*.c`.
16212 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16214 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16215 has a return value which indicates the quality of the random data
16216 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16217 error queue. New function RAND_pseudo_bytes() generates output that is
16218 guaranteed to be unique but not unpredictable. RAND_add is like
16219 RAND_seed, but takes an extra argument for an entropy estimate
16220 (RAND_seed always assumes full entropy).
16224 * Do more iterations of Rabin-Miller probable prime test (specifically,
16225 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16226 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16227 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16228 false-positive rate of at most 2^-80 for random input.
16232 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16236 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16237 in the 0.9.5 release), this returns the chain
16238 from an X509_CTX structure with a dup of the stack and all
16239 the X509 reference counts upped: so the stack will exist
16240 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16243 Also make SSL_SESSION_print() print out the verify return
16248 * Add manpage for the pkcs12 command. Also change the default
16249 behaviour so MAC iteration counts are used unless the new
16250 -nomaciter option is used. This improves file security and
16251 only older versions of MSIE (4.0 for example) need it.
16255 * Honor the no-xxx Configure options when creating .DEF files.
16259 * Add PKCS#10 attributes to field table: challengePassword,
16260 unstructuredName and unstructuredAddress. These are taken from
16261 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16262 international characters are used.
16264 More changes to X509_ATTRIBUTE code: allow the setting of types
16265 based on strings. Remove the 'loc' parameter when adding
16266 attributes because these will be a SET OF encoding which is sorted
16271 * Initial changes to the 'req' utility to allow request generation
16272 automation. This will allow an application to just generate a template
16273 file containing all the field values and have req construct the
16276 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16277 used all over the place including certificate requests and PKCS#7
16278 structures. They are currently handled manually where necessary with
16279 some primitive wrappers for PKCS#7. The new functions behave in a
16280 manner analogous to the X509 extension functions: they allow
16281 attributes to be looked up by NID and added.
16283 Later something similar to the X509V3 code would be desirable to
16284 automatically handle the encoding, decoding and printing of the
16285 more complex types. The string types like challengePassword can
16286 be handled by the string table functions.
16288 Also modified the multi byte string table handling. Now there is
16289 a 'global mask' which masks out certain types. The table itself
16290 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16291 is useful when for example there is only one permissible type
16292 (as in countryName) and using the mask might result in no valid
16297 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16298 SSL_get_peer_finished to allow applications to obtain the latest
16299 Finished messages sent to the peer or expected from the peer,
16300 respectively. (SSL_get_peer_finished is usually the Finished message
16301 actually received from the peer, otherwise the protocol will be aborted.)
16303 As the Finished message are message digests of the complete handshake
16304 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16305 be used for external authentication procedures when the authentication
16306 provided by SSL/TLS is not desired or is not enough.
16310 * Enhanced support for Alpha Linux is added. Now ./config checks if
16311 the host supports BWX extension and if Compaq C is present on the
16312 $PATH. Just exploiting of the BWX extension results in 20-30%
16313 performance kick for some algorithms, e.g. DES and RC4 to mention
16314 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16319 * Add support for MS "fast SGC". This is arguably a violation of the
16320 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16321 weak crypto and after checking the certificate is SGC a second one
16322 with strong crypto. MS SGC stops the first handshake after receiving
16323 the server certificate message and sends a second client hello. Since
16324 a server will typically do all the time consuming operations before
16325 expecting any further messages from the client (server key exchange
16326 is the most expensive) there is little difference between the two.
16328 To get OpenSSL to support MS SGC we have to permit a second client
16329 hello message after we have sent server done. In addition we have to
16330 reset the MAC if we do get this second client hello.
16334 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16335 if a DER encoded private key is RSA or DSA traditional format. Changed
16336 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16337 format DER encoded private key. Newer code should use PKCS#8 format which
16338 has the key type encoded in the ASN1 structure. Added DER private key
16339 support to pkcs8 application.
16343 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16344 ciphersuites has been selected (as required by the SSL 3/TLS 1
16345 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16346 is set, we interpret this as a request to violate the specification
16347 (the worst that can happen is a handshake failure, and 'correct'
16348 behaviour would result in a handshake failure anyway).
16352 * In SSL_CTX_add_session, take into account that there might be multiple
16353 SSL_SESSION structures with the same session ID (e.g. when two threads
16354 concurrently obtain them from an external cache).
16355 The internal cache can handle only one SSL_SESSION with a given ID,
16356 so if there's a conflict, we now throw out the old one to achieve
16361 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16362 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16363 some routines that use cipher OIDs: some ciphers do not have OIDs
16364 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16369 * Simplify the trust setting structure and code. Now we just have
16370 two sequences of OIDs for trusted and rejected settings. These will
16371 typically have values the same as the extended key usage extension
16372 and any application specific purposes.
16374 The trust checking code now has a default behaviour: it will just
16375 check for an object with the same NID as the passed id. Functions can
16376 be provided to override either the default behaviour or the behaviour
16377 for a given id. SSL client, server and email already have functions
16378 in place for compatibility: they check the NID and also return "trusted"
16379 if the certificate is self signed.
16383 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16384 traditional format into an EVP_PKEY structure.
16388 * Add a password callback function PEM_cb() which either prompts for
16389 a password if usr_data is NULL or otherwise assumes it is a null
16390 terminated password. Allow passwords to be passed on command line
16391 environment or config files in a few more utilities.
16395 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16396 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16397 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16398 Update documentation.
16402 * Support for ASN1 "NULL" type. This could be handled before by using
16403 ASN1_TYPE but there wasn't any function that would try to read a NULL
16404 and produce an error if it couldn't. For compatibility we also have
16405 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16406 don't allocate anything because they don't need to.
16410 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16413 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16415 * Rebuild of the memory allocation routines used by OpenSSL code and
16416 possibly others as well. The purpose is to make an interface that
16417 provide hooks so anyone can build a separate set of allocation and
16418 deallocation routines to be used by OpenSSL, for example memory
16419 pool implementations, or something else, which was previously hard
16420 since Malloc(), Realloc() and Free() were defined as macros having
16421 the values malloc, realloc and free, respectively (except for Win32
16422 compilations). The same is provided for memory debugging code.
16423 OpenSSL already comes with functionality to find memory leaks, but
16424 this gives people a chance to debug other memory problems.
16426 With these changes, a new set of functions and macros have appeared:
16428 CRYPTO_set_mem_debug_functions() [F]
16429 CRYPTO_get_mem_debug_functions() [F]
16430 CRYPTO_dbg_set_options() [F]
16431 CRYPTO_dbg_get_options() [F]
16432 CRYPTO_malloc_debug_init() [M]
16434 The memory debug functions are NULL by default, unless the library
16435 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16436 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16437 gives the standard debugging functions that come with OpenSSL) or
16438 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16439 provided by the library user) must be used. When the standard
16440 debugging functions are used, CRYPTO_dbg_set_options can be used to
16441 request additional information:
16442 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16443 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16445 Also, things like CRYPTO_set_mem_functions will always give the
16446 expected result (the new set of functions is used for allocation
16447 and deallocation) at all times, regardless of platform and compiler
16450 To finish it up, some functions that were never use in any other
16451 way than through macros have a new API and new semantic:
16453 CRYPTO_dbg_malloc()
16454 CRYPTO_dbg_realloc()
16457 All macros of value have retained their old syntax.
16459 *Richard Levitte and Bodo Moeller*
16461 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16462 ordering of SMIMECapabilities wasn't in "strength order" and there
16463 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16468 * Some ASN1 types with illegal zero length encoding (INTEGER,
16469 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16471 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16473 * Merge in my S/MIME library for OpenSSL. This provides a simple
16474 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16475 functionality to handle multipart/signed properly) and a utility
16476 called 'smime' to call all this stuff. This is based on code I
16477 originally wrote for Celo who have kindly allowed it to be
16478 included in OpenSSL.
16482 * Add variants des_set_key_checked and des_set_key_unchecked of
16483 des_set_key (aka des_key_sched). Global variable des_check_key
16484 decides which of these is called by des_set_key; this way
16485 des_check_key behaves as it always did, but applications and
16486 the library itself, which was buggy for des_check_key == 1,
16487 have a cleaner way to pick the version they need.
16491 * New function PKCS12_newpass() which changes the password of a
16496 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16497 dynamic mix. In both cases the ids can be used as an index into the
16498 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16499 functions so they accept a list of the field values and the
16500 application doesn't need to directly manipulate the X509_TRUST
16505 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16510 * Modify the way the V3 extension code looks up extensions. This now
16511 works in a similar way to the object code: we have some "standard"
16512 extensions in a static table which is searched with OBJ_bsearch()
16513 and the application can add dynamic ones if needed. The file
16514 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16515 updated whenever a new extension is added to the core code and kept
16516 in ext_nid order. There is a simple program 'tabtest.c' which checks
16517 this. New extensions are not added too often so this file can readily
16518 be maintained manually.
16520 There are two big advantages in doing things this way. The extensions
16521 can be looked up immediately and no longer need to be "added" using
16522 X509V3_add_standard_extensions(): this function now does nothing.
16523 Side note: I get *lots* of email saying the extension code doesn't
16524 work because people forget to call this function.
16525 Also no dynamic allocation is done unless new extensions are added:
16526 so if we don't add custom extensions there is no need to call
16527 X509V3_EXT_cleanup().
16531 * Modify enc utility's salting as follows: make salting the default. Add a
16532 magic header, so unsalted files fail gracefully instead of just decrypting
16533 to garbage. This is because not salting is a big security hole, so people
16534 should be discouraged from doing it.
16538 * Fixes and enhancements to the 'x509' utility. It allowed a message
16539 digest to be passed on the command line but it only used this
16540 parameter when signing a certificate. Modified so all relevant
16541 operations are affected by the digest parameter including the
16542 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16543 DSA key was used because it didn't fix the digest.
16547 * Initial certificate chain verify code. Currently tests the untrusted
16548 certificates for consistency with the verify purpose (which is set
16549 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16551 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16552 this is because it will reject chains with invalid extensions whereas
16553 every previous version of OpenSSL and SSLeay made no checks at all.
16555 Trust code: checks the root CA for the relevant trust settings. Trust
16556 settings have an initial value consistent with the verify purpose: e.g.
16557 if the verify purpose is for SSL client use it expects the CA to be
16558 trusted for SSL client use. However the default value can be changed to
16559 permit custom trust settings: one example of this would be to only trust
16560 certificates from a specific "secure" set of CAs.
16562 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16563 which should be used for version portability: especially since the
16564 verify structure is likely to change more often now.
16566 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16567 to set them. If not set then assume SSL clients will verify SSL servers
16570 Two new options to the verify program: -untrusted allows a set of
16571 untrusted certificates to be passed in and -purpose which sets the
16572 intended purpose of the certificate. If a purpose is set then the
16573 new chain verify code is used to check extension consistency.
16577 * Support for the authority information access extension.
16581 * Modify RSA and DSA PEM read routines to transparently handle
16582 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16583 public keys in a format compatible with certificate
16584 SubjectPublicKeyInfo structures. Unfortunately there were already
16585 functions called *_PublicKey_* which used various odd formats so
16586 these are retained for compatibility: however the DSA variants were
16587 never in a public release so they have been deleted. Changed dsa/rsa
16588 utilities to handle the new format: note no releases ever handled public
16589 keys so we should be OK.
16591 The primary motivation for this change is to avoid the same fiasco
16592 that dogs private keys: there are several incompatible private key
16593 formats some of which are standard and some OpenSSL specific and
16594 require various evil hacks to allow partial transparent handling and
16595 even then it doesn't work with DER formats. Given the option anything
16596 other than PKCS#8 should be dumped: but the other formats have to
16597 stay in the name of compatibility.
16599 With public keys and the benefit of hindsight one standard format
16600 is used which works with EVP_PKEY, RSA or DSA structures: though
16601 it clearly returns an error if you try to read the wrong kind of key.
16603 Added a -pubkey option to the 'x509' utility to output the public key.
16604 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16605 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16606 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16607 that do the same as the `EVP_PKEY_assign_*()` except they up the
16608 reference count of the added key (they don't "swallow" the
16613 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16614 CRLs would fail if the file contained no certificates or no CRLs:
16615 added a new function to read in both types and return the number
16616 read: this means that if none are read it will be an error. The
16617 DER versions of the certificate and CRL reader would always fail
16618 because it isn't possible to mix certificates and CRLs in DER format
16619 without choking one or the other routine. Changed this to just read
16620 a certificate: this is the best we can do. Also modified the code
16621 in `apps/verify.c` to take notice of return codes: it was previously
16622 attempting to read in certificates from NULL pointers and ignoring
16623 any errors: this is one reason why the cert and CRL reader seemed
16624 to work. It doesn't check return codes from the default certificate
16625 routines: these may well fail if the certificates aren't installed.
16629 * Code to support otherName option in GeneralName.
16633 * First update to verify code. Change the verify utility
16634 so it warns if it is passed a self signed certificate:
16635 for consistency with the normal behaviour. X509_verify
16636 has been modified to it will now verify a self signed
16637 certificate if *exactly* the same certificate appears
16638 in the store: it was previously impossible to trust a
16639 single self signed certificate. This means that:
16640 openssl verify ss.pem
16641 now gives a warning about a self signed certificate but
16642 openssl verify -CAfile ss.pem ss.pem
16647 * For servers, store verify_result in SSL_SESSION data structure
16648 (and add it to external session representation).
16649 This is needed when client certificate verifications fails,
16650 but an application-provided verification callback (set by
16651 SSL_CTX_set_cert_verify_callback) allows accepting the session
16652 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16653 but returns 1): When the session is reused, we have to set
16654 ssl->verify_result to the appropriate error code to avoid
16657 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16659 * Fix a bug in the new PKCS#7 code: it didn't consider the
16660 case in PKCS7_dataInit() where the signed PKCS7 structure
16661 didn't contain any existing data because it was being created.
16663 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16665 * Add a salt to the key derivation routines in enc.c. This
16666 forms the first 8 bytes of the encrypted file. Also add a
16667 -S option to allow a salt to be input on the command line.
16671 * New function X509_cmp(). Oddly enough there wasn't a function
16672 to compare two certificates. We do this by working out the SHA1
16673 hash and comparing that. X509_cmp() will be needed by the trust
16678 * SSL_get1_session() is like SSL_get_session(), but increments
16679 the reference count in the SSL_SESSION returned.
16681 *Geoff Thorpe <geoff@eu.c2.net>*
16683 * Fix for 'req': it was adding a null to request attributes.
16684 Also change the X509_LOOKUP and X509_INFO code to handle
16685 certificate auxiliary information.
16689 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16694 * Add the possibility to add extra information to the memory leak
16695 detecting output, to form tracebacks, showing from where each
16696 allocation was originated: CRYPTO_push_info("constant string") adds
16697 the string plus current file name and line number to a per-thread
16698 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16699 is like calling CYRPTO_pop_info() until the stack is empty.
16700 Also updated memory leak detection code to be multi-thread-safe.
16704 * Add options -text and -noout to pkcs7 utility and delete the
16705 encryption options which never did anything. Update docs.
16709 * Add options to some of the utilities to allow the pass phrase
16710 to be included on either the command line (not recommended on
16711 OSes like Unix) or read from the environment. Update the
16712 manpages and fix a few bugs.
16716 * Add a few manpages for some of the openssl commands.
16720 * Fix the -revoke option in ca. It was freeing up memory twice,
16721 leaking and not finding already revoked certificates.
16725 * Extensive changes to support certificate auxiliary information.
16726 This involves the use of X509_CERT_AUX structure and X509_AUX
16727 functions. An X509_AUX function such as PEM_read_X509_AUX()
16728 can still read in a certificate file in the usual way but it
16729 will also read in any additional "auxiliary information". By
16730 doing things this way a fair degree of compatibility can be
16731 retained: existing certificates can have this information added
16732 using the new 'x509' options.
16734 Current auxiliary information includes an "alias" and some trust
16735 settings. The trust settings will ultimately be used in enhanced
16736 certificate chain verification routines: currently a certificate
16737 can only be trusted if it is self signed and then it is trusted
16742 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16743 The problem was that one of the replacement routines had not been working
16744 since SSLeay releases. For now the offending routine has been replaced
16745 with non-optimised assembler. Even so, this now gives around 95%
16746 performance improvement for 1024 bit RSA signs.
16750 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16751 handling. Most clients have the effective key size in bits equal to
16752 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16753 A few however don't do this and instead use the size of the decrypted key
16754 to determine the RC2 key length and the AlgorithmIdentifier to determine
16755 the effective key length. In this case the effective key length can still
16756 be 40 bits but the key length can be 168 bits for example. This is fixed
16757 by manually forcing an RC2 key into the EVP_PKEY structure because the
16758 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16759 the key length and effective key length are equal.
16763 * Add a bunch of functions that should simplify the creation of
16764 X509_NAME structures. Now you should be able to do:
16765 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16766 and have it automatically work out the correct field type and fill in
16767 the structures. The more adventurous can try:
16768 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16769 and it will (hopefully) work out the correct multibyte encoding.
16773 * Change the 'req' utility to use the new field handling and multibyte
16774 copy routines. Before the DN field creation was handled in an ad hoc
16775 way in req, ca, and x509 which was rather broken and didn't support
16776 BMPStrings or UTF8Strings. Since some software doesn't implement
16777 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16778 using the dirstring_type option. See the new comment in the default
16779 openssl.cnf for more info.
16783 * Make crypto/rand/md_rand.c more robust:
16784 - Assure unique random numbers after fork().
16785 - Make sure that concurrent threads access the global counter and
16786 md serializably so that we never lose entropy in them
16787 or use exactly the same state in multiple threads.
16788 Access to the large state is not always serializable because
16789 the additional locking could be a performance killer, and
16790 md should be large enough anyway.
16794 * New file `apps/app_rand.c` with commonly needed functionality
16795 for handling the random seed file.
16797 Use the random seed file in some applications that previously did not:
16799 dsaparam -genkey (which also ignored its '-rand' option),
16802 x509 (when signing).
16803 Except on systems with /dev/urandom, it is crucial to have a random
16804 seed file at least for key creation, DSA signing, and for DH exchanges;
16805 for RSA signatures we could do without one.
16807 gendh and gendsa (unlike genrsa) used to read only the first byte
16808 of each file listed in the '-rand' option. The function as previously
16809 found in genrsa is now in app_rand.c and is used by all programs
16810 that support '-rand'.
16814 * In RAND_write_file, use mode 0600 for creating files;
16815 don't just chmod when it may be too late.
16819 * Report an error from X509_STORE_load_locations
16820 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16824 * New function ASN1_mbstring_copy() this copies a string in either
16825 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16826 into an ASN1_STRING type. A mask of permissible types is passed
16827 and it chooses the "minimal" type to use or an error if not type
16832 * Add function equivalents to the various macros in asn1.h. The old
16833 macros are retained with an `M_` prefix. Code inside the library can
16834 use the `M_` macros. External code (including the openssl utility)
16835 should *NOT* in order to be "shared library friendly".
16839 * Add various functions that can check a certificate's extensions
16840 to see if it usable for various purposes such as SSL client,
16841 server or S/MIME and CAs of these types. This is currently
16842 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16843 verification. Also added a -purpose flag to x509 utility to
16844 print out all the purposes.
16848 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16853 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16854 for, obtain and decode and extension and obtain its critical flag.
16855 This allows all the necessary extension code to be handled in a
16856 single function call.
16860 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16861 platforms. See crypto/rc4/rc4_enc.c for further details.
16865 * New -noout option to asn1parse. This causes no output to be produced
16866 its main use is when combined with -strparse and -out to extract data
16867 from a file (which may not be in ASN.1 format).
16871 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16872 when producing the local key id.
16874 *Richard Levitte <levitte@stacken.kth.se>*
16876 * New option -dhparam in s_server. This allows a DH parameter file to be
16877 stated explicitly. If it is not stated then it tries the first server
16878 certificate file. The previous behaviour hard coded the filename
16883 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16884 a public key to be input or output. For example:
16885 openssl rsa -in key.pem -pubout -out pubkey.pem
16886 Also added necessary DSA public key functions to handle this.
16890 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16891 in the message. This was handled by allowing
16892 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16894 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16896 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16897 to the end of the strings whereas this didn't. This would cause problems
16898 if strings read with d2i_ASN1_bytes() were later modified.
16900 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16902 * Fix for base64 decode bug. When a base64 bio reads only one line of
16903 data and it contains EOF it will end up returning an error. This is
16904 caused by input 46 bytes long. The cause is due to the way base64
16905 BIOs find the start of base64 encoded data. They do this by trying a
16906 trial decode on each line until they find one that works. When they
16907 do a flag is set and it starts again knowing it can pass all the
16908 data directly through the decoder. Unfortunately it doesn't reset
16909 the context it uses. This means that if EOF is reached an attempt
16910 is made to pass two EOFs through the context and this causes the
16911 resulting error. This can also cause other problems as well. As is
16912 usual with these problems it takes *ages* to find and the fix is
16913 trivial: move one line.
16915 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16917 * Ugly workaround to get s_client and s_server working under Windows. The
16918 old code wouldn't work because it needed to select() on sockets and the
16919 tty (for keypresses and to see if data could be written). Win32 only
16920 supports select() on sockets so we select() with a 1s timeout on the
16921 sockets and then see if any characters are waiting to be read, if none
16922 are present then we retry, we also assume we can always write data to
16923 the tty. This isn't nice because the code then blocks until we've
16924 received a complete line of data and it is effectively polling the
16925 keyboard at 1s intervals: however it's quite a bit better than not
16926 working at all :-) A dedicated Windows application might handle this
16927 with an event loop for example.
16931 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16932 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16933 will be called when RSA_sign() and RSA_verify() are used. This is useful
16934 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16935 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16936 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16937 This necessitated the support of an extra signature type NID_md5_sha1
16938 for SSL signatures and modifications to the SSL library to use it instead
16939 of calling RSA_public_decrypt() and RSA_private_encrypt().
16943 * Add new -verify -CAfile and -CApath options to the crl program, these
16944 will lookup a CRL issuers certificate and verify the signature in a
16945 similar way to the verify program. Tidy up the crl program so it
16946 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16947 less strict. It will now permit CRL extensions even if it is not
16948 a V2 CRL: this will allow it to tolerate some broken CRLs.
16952 * Initialize all non-automatic variables each time one of the openssl
16953 sub-programs is started (this is necessary as they may be started
16954 multiple times from the "OpenSSL>" prompt).
16956 *Lennart Bang, Bodo Moeller*
16958 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16959 removing all other RSA functionality (this is what NO_RSA does). This
16960 is so (for example) those in the US can disable those operations covered
16961 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16966 * Non-copying interface to BIO pairs.
16967 (still largely untested)
16971 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16972 ASCII string. This was handled independently in various places before.
16976 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16977 UTF8 strings a character at a time.
16981 * Use client_version from client hello to select the protocol
16982 (s23_srvr.c) and for RSA client key exchange verification
16983 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16987 * Add various utility functions to handle SPKACs, these were previously
16988 handled by poking round in the structure internals. Added new function
16989 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16990 print, verify and generate SPKACs. Based on an original idea from
16991 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16995 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16999 * Allow the config file extension section to be overwritten on the
17000 command line. Based on an original idea from Massimiliano Pala
17001 <madwolf@comune.modena.it>. The new option is called -extensions
17002 and can be applied to ca, req and x509. Also -reqexts to override
17003 the request extensions in req and -crlexts to override the crl extensions
17008 * Add new feature to the SPKAC handling in ca. Now you can include
17009 the same field multiple times by preceding it by "XXXX." for example:
17012 this is the same syntax as used in the req config file.
17016 * Allow certificate extensions to be added to certificate requests. These
17017 are specified in a 'req_extensions' option of the req section of the
17018 config file. They can be printed out with the -text option to req but
17019 are otherwise ignored at present.
17023 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17024 data read consists of only the final block it would not decrypted because
17025 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17026 A misplaced 'break' also meant the decrypted final block might not be
17027 copied until the next read.
17031 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17032 a few extra parameters to the DH structure: these will be useful if
17033 for example we want the value of 'q' or implement X9.42 DH.
17037 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17038 provides hooks that allow the default DSA functions or functions on a
17039 "per key" basis to be replaced. This allows hardware acceleration and
17040 hardware key storage to be handled without major modification to the
17041 library. Also added low level modexp hooks and CRYPTO_EX structure and
17042 associated functions.
17046 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17047 as "read only": it can't be written to and the buffer it points to will
17048 not be freed. Reading from a read only BIO is much more efficient than
17049 a normal memory BIO. This was added because there are several times when
17050 an area of memory needs to be read from a BIO. The previous method was
17051 to create a memory BIO and write the data to it, this results in two
17052 copies of the data and an O(n^2) reading algorithm. There is a new
17053 function BIO_new_mem_buf() which creates a read only memory BIO from
17054 an area of memory. Also modified the PKCS#7 routines to use read only
17059 * Bugfix: ssl23_get_client_hello did not work properly when called in
17060 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17061 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17062 but a retry condition occurred while trying to read the rest.
17066 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17067 NID_pkcs7_encrypted by default: this was wrong since this should almost
17068 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17069 the encrypted data type: this is a more sensible place to put it and it
17070 allows the PKCS#12 code to be tidied up that duplicated this
17075 * Changed obj_dat.pl script so it takes its input and output files on
17076 the command line. This should avoid shell escape redirection problems
17081 * Initial support for certificate extension requests, these are included
17082 in things like Xenroll certificate requests. Included functions to allow
17083 extensions to be obtained and added.
17087 * -crlf option to s_client and s_server for sending newlines as
17088 CRLF (as required by many protocols).
17092 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17094 * Install libRSAglue.a when OpenSSL is built with RSAref.
17096 *Ralf S. Engelschall*
17098 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17100 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17102 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17107 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17108 DH parameters/keys (q is lost during that conversion, but the resulting
17109 DH parameters contain its length).
17111 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17112 much faster than DH_generate_parameters (which creates parameters
17113 where `p = 2*q + 1`), and also the smaller q makes DH computations
17114 much more efficient (160-bit exponentiation instead of 1024-bit
17115 exponentiation); so this provides a convenient way to support DHE
17116 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17117 utter importance to use
17118 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17120 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17121 when such DH parameters are used, because otherwise small subgroup
17122 attacks may become possible!
17126 * Avoid memory leak in i2d_DHparams.
17130 * Allow the -k option to be used more than once in the enc program:
17131 this allows the same encrypted message to be read by multiple recipients.
17135 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17136 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17137 it will always use the numerical form of the OID, even if it has a short
17142 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17143 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17144 otherwise bn_mod_exp was called. In the case of hardware keys for example
17145 no private key components need be present and it might store extra data
17146 in the RSA structure, which cannot be accessed from bn_mod_exp.
17147 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17148 private key operations.
17152 * Added support for SPARC Linux.
17156 * pem_password_cb function type incompatibly changed from
17157 typedef int pem_password_cb(char *buf, int size, int rwflag);
17159 ....(char *buf, int size, int rwflag, void *userdata);
17160 so that applications can pass data to their callbacks:
17161 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17162 additional void * argument, which is just handed through whenever
17163 the password callback is called.
17165 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17167 New function SSL_CTX_set_default_passwd_cb_userdata.
17169 Compatibility note: As many C implementations push function arguments
17170 onto the stack in reverse order, the new library version is likely to
17171 interoperate with programs that have been compiled with the old
17172 pem_password_cb definition (PEM_whatever takes some data that
17173 happens to be on the stack as its last argument, and the callback
17174 just ignores this garbage); but there is no guarantee whatsoever that
17177 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17178 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17179 problems not only on Windows, but also on some Unix platforms.
17180 To avoid problematic command lines, these definitions are now in an
17181 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17182 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17186 * MIPS III/IV assembler module is reimplemented.
17190 * More DES library cleanups: remove references to srand/rand and
17191 delete an unused file.
17195 * Add support for the free Netwide assembler (NASM) under Win32,
17196 since not many people have MASM (ml) and it can be hard to obtain.
17197 This is currently experimental but it seems to work OK and pass all
17198 the tests. Check out INSTALL.W32 for info.
17202 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17203 without temporary keys kept an extra copy of the server key,
17204 and connections with temporary keys did not free everything in case
17209 * New function RSA_check_key and new openssl rsa option -check
17210 for verifying the consistency of RSA keys.
17212 *Ulf Moeller, Bodo Moeller*
17214 * Various changes to make Win32 compile work:
17215 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17216 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17217 comparison" warnings.
17218 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17222 * Add a debugging option to PKCS#5 v2 key generation function: when
17223 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17224 derived keys are printed to stderr.
17228 * Copy the flags in ASN1_STRING_dup().
17230 *Roman E. Pavlov <pre@mo.msk.ru>*
17232 * The x509 application mishandled signing requests containing DSA
17233 keys when the signing key was also DSA and the parameters didn't match.
17235 It was supposed to omit the parameters when they matched the signing key:
17236 the verifying software was then supposed to automatically use the CA's
17237 parameters if they were absent from the end user certificate.
17239 Omitting parameters is no longer recommended. The test was also
17240 the wrong way round! This was probably due to unusual behaviour in
17241 EVP_cmp_parameters() which returns 1 if the parameters match.
17242 This meant that parameters were omitted when they *didn't* match and
17243 the certificate was useless. Certificates signed with 'ca' didn't have
17246 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17248 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17249 The interface is as follows:
17250 Applications can use
17251 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17252 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17253 "off" is now the default.
17254 The library internally uses
17255 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17256 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17257 to disable memory-checking temporarily.
17259 Some inconsistent states that previously were possible (and were
17260 even the default) are now avoided.
17262 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17263 with each memory chunk allocated; this is occasionally more helpful
17264 than just having a counter.
17266 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17268 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17273 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17274 which largely parallels "options", but is for changing API behaviour,
17275 whereas "options" are about protocol behaviour.
17276 Initial "mode" flags are:
17278 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17279 a single record has been written.
17280 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17281 retries use the same buffer location.
17282 (But all of the contents must be
17287 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17290 * Fix problems with no-hmac etc.
17292 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17294 * New functions RSA_get_default_method(), RSA_set_method() and
17295 RSA_get_method(). These allows replacement of RSA_METHODs without having
17296 to mess around with the internals of an RSA structure.
17300 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17301 Also really enable memory leak checks in openssl.c and in some
17304 *Chad C. Mulligan, Bodo Moeller*
17306 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17307 up the length of negative integers. This has now been simplified to just
17308 store the length when it is first determined and use it later, rather
17309 than trying to keep track of where data is copied and updating it to
17311 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17313 * Add a new function PKCS7_signatureVerify. This allows the verification
17314 of a PKCS#7 signature but with the signing certificate passed to the
17315 function itself. This contrasts with PKCS7_dataVerify which assumes the
17316 certificate is present in the PKCS#7 structure. This isn't always the
17317 case: certificates can be omitted from a PKCS#7 structure and be
17318 distributed by "out of band" means (such as a certificate database).
17322 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17323 function prototypes in pem.h, also change util/mkdef.pl to add the
17324 necessary function names.
17328 * mk1mf.pl (used by Windows builds) did not properly read the
17329 options set by Configure in the top level Makefile, and Configure
17330 was not even able to write more than one option correctly.
17331 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17335 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17336 file to be loaded from a BIO or FILE pointer. The BIO version will
17337 for example allow memory BIOs to contain config info.
17341 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17342 Whoever hopes to achieve shared-library compatibility across versions
17343 must use this, not the compile-time macro.
17344 (Exercise 0.9.4: Which is the minimum library version required by
17346 Note: All this applies only to multi-threaded programs, others don't
17351 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17352 through a BIO pair triggered the default case, i.e.
17353 SSLerr(...,SSL_R_UNKNOWN_STATE).
17357 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17358 can use the SSL library even if none of the specific BIOs is
17363 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17364 for the encoded length.
17366 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17368 * Add initial documentation of the X509V3 functions.
17372 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17373 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17374 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17375 secure PKCS#8 private key format with a high iteration count.
17379 * Fix determination of Perl interpreter: A perl or perl5
17380 *directory* in $PATH was also accepted as the interpreter.
17382 *Ralf S. Engelschall*
17384 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17385 wrong with it but it was very old and did things like calling
17386 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17387 unusual formatting.
17391 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17392 to use the new extension code.
17396 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17397 with macros. This should make it easier to change their form, add extra
17398 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17403 * Add to configuration table a new entry that can specify an alternative
17404 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17405 according to Mark Crispin <MRC@Panda.COM>.
17409 * DES CBC did not update the IV. Weird.
17413 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17414 Changing the behaviour of the former might break existing programs --
17415 where IV updating is needed, des_ncbc_encrypt can be used.
17418 * When bntest is run from "make test" it drives bc to check its
17419 calculations, as well as internally checking them. If an internal check
17420 fails, it needs to cause bc to give a non-zero result or make test carries
17421 on without noticing the failure. Fixed.
17425 * DES library cleanups.
17429 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17430 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17431 ciphers. NOTE: although the key derivation function has been verified
17432 against some published test vectors it has not been extensively tested
17433 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17438 * Instead of "mkdir -p", which is not fully portable, use new
17439 Perl script "util/mkdir-p.pl".
17443 * Rewrite the way password based encryption (PBE) is handled. It used to
17444 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17445 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17446 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17447 the 'parameter' field of the AlgorithmIdentifier is passed to the
17448 underlying key generation function so it must do its own ASN1 parsing.
17449 This has also changed the EVP_PBE_CipherInit() function which now has a
17450 'parameter' argument instead of literal salt and iteration count values
17451 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17455 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17456 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17457 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17458 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17459 value was just used as a "magic string" and not used directly its
17460 value doesn't matter.
17464 * Introduce some semblance of const correctness to BN. Shame C doesn't
17469 * "linux-sparc64" configuration (ultrapenguin).
17471 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17472 "linux-sparc" configuration.
17474 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17476 * config now generates no-xxx options for missing ciphers.
17480 * Support the EBCDIC character set (work in progress).
17481 File ebcdic.c not yet included because it has a different license.
17483 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17485 * Support BS2000/OSD-POSIX.
17487 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17489 * Make callbacks for key generation use `void *` instead of `char *`.
17493 * Make S/MIME samples compile (not yet tested).
17497 * Additional typesafe stacks.
17501 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17505 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17507 * New configuration variant "sco5-gcc".
17509 * Updated some demos.
17511 *Sean O Riordain, Wade Scholine*
17513 * Add missing BIO_free at exit of pkcs12 application.
17517 * Fix memory leak in conf.c.
17521 * Updates for Win32 to assembler version of MD5.
17525 * Set #! path to perl in `apps/der_chop` to where we found it
17526 instead of using a fixed path.
17530 * SHA library changes for irix64-mips4-cc.
17534 * Improvements for VMS support.
17538 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17540 * Bignum library bug fix. IRIX 6 passes "make test" now!
17541 This also avoids the problems with SC4.2 and unpatched SC5.
17543 *Andy Polyakov <appro@fy.chalmers.se>*
17545 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17546 These are required because of the typesafe stack would otherwise break
17547 existing code. If old code used a structure member which used to be STACK
17548 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17549 sk_num or sk_value it would produce an error because the num, data members
17550 are not present in STACK_OF. Now it just produces a warning. sk_set
17551 replaces the old method of assigning a value to sk_value
17552 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17553 that does this will no longer work (and should use sk_set instead) but
17554 this could be regarded as a "questionable" behaviour anyway.
17558 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17559 correctly handle encrypted S/MIME data.
17563 * Change type of various DES function arguments from des_cblock
17564 (which means, in function argument declarations, pointer to char)
17565 to des_cblock * (meaning pointer to array with 8 char elements),
17566 which allows the compiler to do more typechecking; it was like
17567 that back in SSLeay, but with lots of ugly casts.
17569 Introduce new type const_des_cblock.
17573 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17574 problems: find RecipientInfo structure that matches recipient certificate
17575 and initialise the ASN1 structures properly based on passed cipher.
17579 * Belatedly make the BN tests actually check the results.
17583 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17584 to and from BNs: it was completely broken. New compilation option
17585 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17586 key elements as negative integers.
17590 * Reorganize and speed up MD5.
17592 *Andy Polyakov <appro@fy.chalmers.se>*
17596 *Richard Levitte <richard@levitte.org>*
17598 * New option -out to asn1parse to allow the parsed structure to be
17599 output to a file. This is most useful when combined with the -strparse
17600 option to examine the output of things like OCTET STRINGS.
17604 * Make SSL library a little more fool-proof by not requiring any longer
17605 that `SSL_set_{accept,connect}_state` be called before
17606 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17607 in many applications because usually everything *appeared* to work as
17608 intended anyway -- now it really works as intended).
17612 * Move openssl.cnf out of lib/.
17616 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17617 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17618 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17620 *Ralf S. Engelschall*
17622 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17623 handle PKCS#7 enveloped data properly.
17625 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17627 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17628 copying pointers. The cert_st handling is changed by this in
17629 various ways (and thus what used to be known as ctx->default_cert
17630 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17631 any longer when s->cert does not give us what we need).
17632 ssl_cert_instantiate becomes obsolete by this change.
17633 As soon as we've got the new code right (possibly it already is?),
17634 we have solved a couple of bugs of the earlier code where s->cert
17635 was used as if it could not have been shared with other SSL structures.
17637 Note that using the SSL API in certain dirty ways now will result
17638 in different behaviour than observed with earlier library versions:
17639 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17640 does not influence s as it used to.
17642 In order to clean up things more thoroughly, inside SSL_SESSION
17643 we don't use CERT any longer, but a new structure SESS_CERT
17644 that holds per-session data (if available); currently, this is
17645 the peer's certificate chain and, for clients, the server's certificate
17646 and temporary key. CERT holds only those values that can have
17647 meaningful defaults in an SSL_CTX.
17651 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17652 from the internal representation. Various PKCS#7 fixes: remove some
17653 evil casts and set the enc_dig_alg field properly based on the signing
17658 * Allow PKCS#12 password to be set from the command line or the
17659 environment. Let 'ca' get its config file name from the environment
17660 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17665 * Allow certificate policies extension to use an IA5STRING for the
17666 organization field. This is contrary to the PKIX definition but
17667 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17672 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17673 without disallowing inline assembler and the like for non-pedantic builds.
17677 * Support Borland C++ builder.
17679 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17685 * SHA-1 cleanups and performance enhancements.
17687 *Andy Polyakov <appro@fy.chalmers.se>*
17689 * Sparc v8plus assembler for the bignum library.
17691 *Andy Polyakov <appro@fy.chalmers.se>*
17693 * Accept any -xxx and +xxx compiler options in Configure.
17697 * Update HPUX configuration.
17701 * Add missing `sk_<type>_unshift()` function to safestack.h
17703 *Ralf S. Engelschall*
17705 * New function SSL_CTX_use_certificate_chain_file that sets the
17706 "extra_cert"s in addition to the certificate. (This makes sense
17707 only for "PEM" format files, as chains as a whole are not
17712 * Support verify_depth from the SSL API.
17713 x509_vfy.c had what can be considered an off-by-one-error:
17714 Its depth (which was not part of the external interface)
17715 was actually counting the number of certificates in a chain;
17716 now it really counts the depth.
17720 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17721 instead of X509err, which often resulted in confusing error
17722 messages since the error codes are not globally unique
17723 (e.g. an alleged error in ssl3_accept when a certificate
17724 didn't match the private key).
17726 * New function SSL_CTX_set_session_id_context that allows to set a default
17727 value (so that you don't need SSL_set_session_id_context for each
17728 connection using the SSL_CTX).
17732 * OAEP decoding bug fix.
17736 * Support INSTALL_PREFIX for package builders, as proposed by
17741 * New Configure options "threads" and "no-threads". For systems
17742 where the proper compiler options are known (currently Solaris
17743 and Linux), "threads" is the default.
17747 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17751 * Install various scripts to $(OPENSSLDIR)/misc, not to
17752 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17753 such as /usr/local/bin.
17757 * "make linux-shared" to build shared libraries.
17759 *Niels Poppe <niels@netbox.org>*
17761 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17765 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17766 extension adding in x509 utility.
17770 * Remove NOPROTO sections and error code comments.
17774 * Partial rewrite of the DEF file generator to now parse the ANSI
17779 * New Configure options --prefix=DIR and --openssldir=DIR.
17783 * Complete rewrite of the error code script(s). It is all now handled
17784 by one script at the top level which handles error code gathering,
17785 header rewriting and C source file generation. It should be much better
17786 than the old method: it now uses a modified version of Ulf's parser to
17787 read the ANSI prototypes in all header files (thus the old K&R definitions
17788 aren't needed for error creation any more) and do a better job of
17789 translating function codes into names. The old 'ASN1 error code embedded
17790 in a comment' is no longer necessary and it doesn't use .err files which
17791 have now been deleted. Also the error code call doesn't have to appear all
17792 on one line (which resulted in some large lines...).
17796 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17800 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17801 0 (which usually indicates a closed connection), but continue reading.
17805 * Fix some race conditions.
17809 * Add support for CRL distribution points extension. Add Certificate
17810 Policies and CRL distribution points documentation.
17814 * Move the autogenerated header file parts to crypto/opensslconf.h.
17818 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17819 8 of keying material. Merlin has also confirmed interop with this fix
17820 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17822 *Merlin Hughes <merlin@baltimore.ie>*
17824 * Fix lots of warnings.
17826 *Richard Levitte <levitte@stacken.kth.se>*
17828 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17829 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17831 *Richard Levitte <levitte@stacken.kth.se>*
17833 * Fix problems with sizeof(long) == 8.
17835 *Andy Polyakov <appro@fy.chalmers.se>*
17837 * Change functions to ANSI C.
17841 * Fix typos in error codes.
17843 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17845 * Remove defunct assembler files from Configure.
17849 * SPARC v8 assembler BIGNUM implementation.
17851 *Andy Polyakov <appro@fy.chalmers.se>*
17853 * Support for Certificate Policies extension: both print and set.
17854 Various additions to support the r2i method this uses.
17858 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17859 return a const string when you are expecting an allocated buffer.
17863 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17864 types DirectoryString and DisplayText.
17868 * Add code to allow r2i extensions to access the configuration database,
17869 add an LHASH database driver and add several ctx helper functions.
17873 * Fix an evil bug in bn_expand2() which caused various BN functions to
17874 fail when they extended the size of a BIGNUM.
17878 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17879 support typesafe stack.
17883 * Fix typo in SSL_[gs]et_options().
17885 *Nils Frostberg <nils@medcom.se>*
17887 * Delete various functions and files that belonged to the (now obsolete)
17888 old X509V3 handling code.
17892 * New Configure option "rsaref".
17896 * Don't auto-generate pem.h.
17900 * Introduce type-safe ASN.1 SETs.
17904 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17906 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17908 * Introduce type-safe STACKs. This will almost certainly break lots of code
17909 that links with OpenSSL (well at least cause lots of warnings), but fear
17910 not: the conversion is trivial, and it eliminates loads of evil casts. A
17911 few STACKed things have been converted already. Feel free to convert more.
17912 In the fullness of time, I'll do away with the STACK type altogether.
17916 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17917 specified in `<certfile>` by updating the entry in the index.txt file.
17918 This way one no longer has to edit the index.txt file manually for
17919 revoking a certificate. The -revoke option does the gory details now.
17921 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17923 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17924 `-text` option at all and this way the `-noout -text` combination was
17925 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17927 *Ralf S. Engelschall*
17929 * Make sure a corresponding plain text error message exists for the
17930 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17931 verify callback function determined that a certificate was revoked.
17933 *Ralf S. Engelschall*
17935 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17936 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17937 all available ciphers including rc5, which was forgotten until now.
17938 In order to let the testing shell script know which algorithms
17939 are available, a new (up to now undocumented) command
17940 `openssl list-cipher-commands` is used.
17944 * Bugfix: s_client occasionally would sleep in select() when
17945 it should have checked SSL_pending() first.
17949 * New functions DSA_do_sign and DSA_do_verify to provide access to
17950 the raw DSA values prior to ASN.1 encoding.
17954 * Tweaks to Configure
17956 *Niels Poppe <niels@netbox.org>*
17958 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17963 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17967 * New config option to avoid instructions that are illegal on the 80386.
17968 The default code is faster, but requires at least a 486.
17972 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17973 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17974 same as SSL2_VERSION anyway.
17978 * New "-showcerts" option for s_client.
17982 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17983 application. Various cleanups and fixes.
17987 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17988 modify error routines to work internally. Add error codes and PBE init
17989 to library startup routines.
17993 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17994 packing functions to asn1 and evp. Changed function names and error
17995 codes along the way.
17999 * PKCS12 integration: and so it begins... First of several patches to
18000 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18001 objects to objects.h
18005 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18006 and display support for Thawte strong extranet extension.
18010 * Add LinuxPPC support.
18012 *Jeff Dubrule <igor@pobox.org>*
18014 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18015 bn_div_words in alpha.s.
18017 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18019 * Make sure the RSA OAEP test is skipped under -DRSAref because
18020 OAEP isn't supported when OpenSSL is built with RSAref.
18022 *Ulf Moeller <ulf@fitug.de>*
18024 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18025 so they no longer are missing under -DNOPROTO.
18027 *Soren S. Jorvang <soren@t.dk>*
18029 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18031 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18032 doesn't work when the session is reused. Coming soon!
18036 * Fix a security hole, that allows sessions to be reused in the wrong
18037 context thus bypassing client cert protection! All software that uses
18038 client certs and session caches in multiple contexts NEEDS PATCHING to
18039 allow session reuse! A fuller solution is in the works.
18041 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18043 * Some more source tree cleanups (removed obsolete files
18044 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18045 permission on "config" script to be executable) and a fix for the INSTALL
18048 *Ulf Moeller <ulf@fitug.de>*
18050 * Remove some legacy and erroneous uses of malloc, free instead of
18053 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18055 * Make rsa_oaep_test return non-zero on error.
18057 *Ulf Moeller <ulf@fitug.de>*
18059 * Add support for native Solaris shared libraries. Configure
18060 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18061 if someone would make that last step automatic.
18063 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18065 * ctx_size was not built with the right compiler during "make links". Fixed.
18069 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18070 except NULL ciphers". This means the default cipher list will no longer
18071 enable NULL ciphers. They need to be specifically enabled e.g. with
18072 the string "DEFAULT:eNULL".
18076 * Fix to RSA private encryption routines: if p < q then it would
18077 occasionally produce an invalid result. This will only happen with
18078 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18082 * Be less restrictive and allow also `perl util/perlpath.pl
18083 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18084 because this way one can also use an interpreter named `perl5` (which is
18085 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18086 installed as `perl`).
18088 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18090 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18092 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18094 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18095 advapi32.lib to Win32 build and change the pem test comparison
18096 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18097 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18098 and crypto/des/ede_cbcm_enc.c.
18102 * DES quad checksum was broken on big-endian architectures. Fixed.
18106 * Comment out two functions in bio.h that aren't implemented. Fix up the
18107 Win32 test batch file so it (might) work again. The Win32 test batch file
18108 is horrible: I feel ill....
18112 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18113 in e_os.h. Audit of header files to check ANSI and non ANSI
18114 sections: 10 functions were absent from non ANSI section and not exported
18115 from Windows DLLs. Fixed up libeay.num for new functions.
18119 * Make `openssl version` output lines consistent.
18121 *Ralf S. Engelschall*
18123 * Fix Win32 symbol export lists for BIO functions: Added
18124 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18125 to ms/libeay{16,32}.def.
18127 *Ralf S. Engelschall*
18129 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18130 fine under Unix and passes some trivial tests I've now added. But the
18131 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18132 added to make sure no one expects that this stuff really works in the
18133 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18134 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18137 *Ralf S. Engelschall*
18139 * Fix the generation of two part addresses in perl.
18141 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18143 * Add config entry for Linux on MIPS.
18145 *John Tobey <jtobey@channel1.com>*
18147 * Make links whenever Configure is run, unless we are on Windoze.
18151 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18152 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18157 * Add a useful kludge to allow package maintainers to specify compiler and
18158 other platforms details on the command line without having to patch the
18159 Configure script every time: One now can use
18160 `perl Configure <id>:<details>`,
18161 i.e. platform ids are allowed to have details appended
18162 to them (separated by colons). This is treated as there would be a static
18163 pre-configured entry in Configure's %table under key `<id>` with value
18164 `<details>` and `perl Configure <id>` is called. So, when you want to
18165 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18166 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18167 now, which overrides the FreeBSD-elf entry on-the-fly.
18169 *Ralf S. Engelschall*
18171 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18175 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18176 on the `perl Configure ...` command line. This way one can compile
18177 OpenSSL libraries with Position Independent Code (PIC) which is needed
18178 for linking it into DSOs.
18180 *Ralf S. Engelschall*
18182 * Remarkably, export ciphers were totally broken and no-one had noticed!
18187 * Cleaned up the LICENSE document: The official contact for any license
18188 questions now is the OpenSSL core team under openssl-core@openssl.org.
18189 And add a paragraph about the dual-license situation to make sure people
18190 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18191 to the OpenSSL toolkit.
18193 *Ralf S. Engelschall*
18195 * General source tree makefile cleanups: Made `making xxx in yyy...`
18196 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18197 Additionally cleaned up the `make links` target: Remove unnecessary
18198 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18199 to speed processing and no longer clutter the display with confusing
18200 stuff. Instead only the actually done links are displayed.
18202 *Ralf S. Engelschall*
18204 * Permit null encryption ciphersuites, used for authentication only. It used
18205 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18206 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18211 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18212 signed attributes when verifying signatures (this would break them),
18213 the detached data encoding was wrong and public keys obtained using
18214 X509_get_pubkey() weren't freed.
18218 * Add text documentation for the BUFFER functions. Also added a work around
18219 to a Win95 console bug. This was triggered by the password read stuff: the
18220 last character typed gets carried over to the next fread(). If you were
18221 generating a new cert request using 'req' for example then the last
18222 character of the passphrase would be CR which would then enter the first
18227 * Added the new 'Includes OpenSSL Cryptography Software' button as
18228 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18229 button and can be used by applications based on OpenSSL to show the
18230 relationship to the OpenSSL project.
18232 *Ralf S. Engelschall*
18234 * Remove confusing variables in function signatures in files
18235 ssl/ssl_lib.c and ssl/ssl.h.
18237 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18239 * Don't install bss_file.c under PREFIX/include/
18241 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18243 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18244 functions that return function pointers and has support for NT specific
18245 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18246 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18247 unsigned to signed types: this was killing the Win32 compile.
18251 * Add new certificate file to stack functions,
18252 SSL_add_dir_cert_subjects_to_stack() and
18253 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18254 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18255 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18256 This means that Apache-SSL and similar packages don't have to mess around
18257 to add as many CAs as they want to the preferred list.
18261 * Experiment with doxygen documentation. Currently only partially applied to
18263 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18264 openssl.doxy as the configuration file.
18268 * Get rid of remaining C++-style comments which strict C compilers hate.
18270 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18272 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18273 compiled in by default: it has problems with large keys.
18277 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18278 DH private keys and/or callback functions which directly correspond to
18279 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18280 is needed for applications which have to configure certificates on a
18281 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18283 For the RSA certificate situation is makes no difference, but
18284 for the DSA certificate situation this fixes the "no shared cipher"
18285 problem where the OpenSSL cipher selection procedure failed because the
18286 temporary keys were not overtaken from the context and the API provided
18287 no way to reconfigure them.
18288 The new functions now let applications reconfigure the stuff and they
18289 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18290 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18291 non-public-API function ssl_cert_instantiate() is used as a helper
18292 function and also to reduce code redundancy inside ssl_rsa.c.
18294 *Ralf S. Engelschall*
18296 * Move s_server -dcert and -dkey options out of the undocumented feature
18297 area because they are useful for the DSA situation and should be
18298 recognized by the users.
18300 *Ralf S. Engelschall*
18302 * Fix the cipher decision scheme for export ciphers: the export bits are
18303 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18304 SSL_EXP_MASK. So, the original variable has to be used instead of the
18305 already masked variable.
18307 *Richard Levitte <levitte@stacken.kth.se>*
18309 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18311 *Richard Levitte <levitte@stacken.kth.se>*
18313 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18314 from `int` to `unsigned int` because it is a length and initialized by
18315 EVP_DigestFinal() which expects an `unsigned int *`.
18317 *Richard Levitte <levitte@stacken.kth.se>*
18319 * Don't hard-code path to Perl interpreter on shebang line of Configure
18320 script. Instead use the usual Shell->Perl transition trick.
18322 *Ralf S. Engelschall*
18324 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18325 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18326 -noout -modulus` as it's already the case for `openssl rsa -noout
18327 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18328 currently the public key is printed (a decision which was already done by
18329 `openssl dsa -modulus` in the past) which serves a similar purpose.
18330 Additionally the NO_RSA no longer completely removes the whole -modulus
18331 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18334 *Ralf S. Engelschall*
18336 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18337 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18339 *Arne Ansper <arne@ats.cyber.ee>*
18341 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18342 to be added. Now both 'req' and 'ca' can use new objects defined in the
18347 * Add cool BIO that does syslog (or event log on NT).
18349 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18351 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18352 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18353 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18354 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18358 * Add preliminary config info for new extension code.
18362 * Make RSA_NO_PADDING really use no padding.
18364 *Ulf Moeller <ulf@fitug.de>*
18366 * Generate errors when private/public key check is done.
18370 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18371 for some CRL extensions and new objects added.
18375 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18376 key usage extension and fuller support for authority key id.
18380 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18381 padding method for RSA, which is recommended for new applications in PKCS
18382 #1 v2.0 (RFC 2437, October 1998).
18383 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18384 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18385 against Bleichbacher's attack on RSA.
18386 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18389 * Updates to the new SSL compression code
18391 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18393 * Fix so that the version number in the master secret, when passed
18394 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18395 (because the server will not accept higher), that the version number
18396 is 0x03,0x01, not 0x03,0x00
18398 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18400 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18401 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18402 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18406 * Support for RAW extensions where an arbitrary extension can be
18407 created by including its DER encoding. See `apps/openssl.cnf` for
18412 * Make sure latest Perl versions don't interpret some generated C array
18413 code as Perl array code in the crypto/err/err_genc.pl script.
18415 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18417 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18418 not many people have the assembler. Various Win32 compilation fixes and
18419 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18420 build instructions.
18424 * Modify configure script 'Configure' to automatically create crypto/date.h
18425 file under Win32 and also build pem.h from pem.org. New script
18426 util/mkfiles.pl to create the MINFO file on environments that can't do a
18427 'make files': perl util/mkfiles.pl >MINFO should work.
18431 * Major rework of DES function declarations, in the pursuit of correctness
18432 and purity. As a result, many evil casts evaporated, and some weirdness,
18433 too. You may find this causes warnings in your code. Zapping your evil
18434 casts will probably fix them. Mostly.
18438 * Fix for a typo in asn1.h. Bug fix to object creation script
18439 obj_dat.pl. It considered a zero in an object definition to mean
18440 "end of object": none of the objects in objects.h have any zeros
18441 so it wasn't spotted.
18443 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18445 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18446 Masking (CBCM). In the absence of test vectors, the best I have been able
18447 to do is check that the decrypt undoes the encrypt, so far. Send me test
18448 vectors if you have them.
18452 * Correct calculation of key length for export ciphers (too much space was
18453 allocated for null ciphers). This has not been tested!
18457 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18458 message is now correct (it understands "crypto" and "ssl" on its
18459 command line). There is also now an "update" option. This will update
18460 the util/ssleay.num and util/libeay.num files with any new functions.
18462 perl util/mkdef.pl crypto ssl update
18463 it will update them.
18467 * Overhauled the Perl interface:
18468 - ported BN stuff to OpenSSL's different BN library
18469 - made the perl/ source tree CVS-aware
18470 - renamed the package from SSLeay to OpenSSL (the files still contain
18471 their history because I've copied them in the repository)
18472 - removed obsolete files (the test scripts will be replaced
18473 by better Test::Harness variants in the future)
18475 *Ralf S. Engelschall*
18477 * First cut for a very conservative source tree cleanup:
18478 1. merge various obsolete readme texts into doc/ssleay.txt
18479 where we collect the old documents and readme texts.
18480 2. remove the first part of files where I'm already sure that we no
18481 longer need them because of three reasons: either they are just temporary
18482 files which were left by Eric or they are preserved original files where
18483 I've verified that the diff is also available in the CVS via "cvs diff
18484 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18485 the crypto/md/ stuff).
18487 *Ralf S. Engelschall*
18489 * More extension code. Incomplete support for subject and issuer alt
18490 name, issuer and authority key id. Change the i2v function parameters
18491 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18492 what that's for :-) Fix to ASN1 macro which messed up
18493 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18497 * Preliminary support for ENUMERATED type. This is largely copied from the
18502 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18504 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18506 * Make sure `make rehash` target really finds the `openssl` program.
18508 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18510 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18511 like to hear about it if this slows down other processors.
18515 * Add CygWin32 platform information to Configure script.
18517 *Alan Batie <batie@aahz.jf.intel.com>*
18519 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18521 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18523 * New program nseq to manipulate netscape certificate sequences
18527 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18532 * Fixes to BN code. Previously the default was to define BN_RECURSION
18533 but the BN code had some problems that would cause failures when
18534 doing certificate verification and some other functions.
18536 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18538 * Add ASN1 and PEM code to support netscape certificate sequences.
18542 * Add ASN1 and PEM code to support netscape certificate sequences.
18546 * Add several PKIX and private extended key usage OIDs.
18550 * Modify the 'ca' program to handle the new extension code. Modify
18551 openssl.cnf for new extension format, add comments.
18555 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18556 and add a sample to openssl.cnf so req -x509 now adds appropriate
18561 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18562 error code, add initial support to X509_print() and x509 application.
18566 * Takes a deep breath and start adding X509 V3 extension support code. Add
18567 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18568 stuff is currently isolated and isn't even compiled yet.
18572 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18573 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18574 Removed the versions check from X509 routines when loading extensions:
18575 this allows certain broken certificates that don't set the version
18576 properly to be processed.
18580 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18581 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18582 can still be regenerated with "make depend".
18586 * Spelling mistake in C version of CAST-128.
18588 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18590 * Changes to the error generation code. The perl script err-code.pl
18591 now reads in the old error codes and retains the old numbers, only
18592 adding new ones if necessary. It also only changes the .err files if new
18593 codes are added. The makefiles have been modified to only insert errors
18594 when needed (to avoid needlessly modifying header files). This is done
18595 by only inserting errors if the .err file is newer than the auto generated
18596 C file. To rebuild all the error codes from scratch (the old behaviour)
18597 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18598 or delete all the .err files.
18602 * CAST-128 was incorrectly implemented for short keys. The C version has
18603 been fixed, but is untested. The assembler versions are also fixed, but
18604 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18605 to regenerate it if needed.
18606 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18607 Hagino <itojun@kame.net>*
18609 * File was opened incorrectly in randfile.c.
18611 *Ulf Möller <ulf@fitug.de>*
18613 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18614 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18615 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18616 al: it's just almost always a UTCTime. Note this patch adds new error
18617 codes so do a "make errors" if there are problems.
18621 * Correct Linux 1 recognition in config.
18623 *Ulf Möller <ulf@fitug.de>*
18625 * Remove pointless MD5 hash when using DSA keys in ca.
18627 *Anonymous <nobody@replay.com>*
18629 * Generate an error if given an empty string as a cert directory. Also
18630 generate an error if handed NULL (previously returned 0 to indicate an
18631 error, but didn't set one).
18633 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18635 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18639 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18640 parameters. This was causing a warning which killed off the Win32 compile.
18644 * Remove C++ style comments from crypto/bn/bn_local.h.
18646 *Neil Costigan <neil.costigan@celocom.com>*
18648 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18649 based on a text string, looking up short and long names and finally
18650 "dot" format. The "dot" format stuff didn't work. Added new function
18651 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18652 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18653 OID is not part of the table.
18657 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18658 X509_LOOKUP_by_alias().
18662 * Sort openssl functions by name.
18666 * Get the `gendsa` command working and add it to the `list` command. Remove
18667 encryption from sample DSA keys (in case anyone is interested the password
18672 * Make *all* `*_free` functions accept a NULL pointer.
18674 *Frans Heymans <fheymans@isaserver.be>*
18676 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18679 *Anonymous <nobody@replay.com>*
18681 * s_server should send the CAfile as acceptable CAs, not its own cert.
18683 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18685 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18687 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18689 * Temp key "for export" tests were wrong in s3_srvr.c.
18691 *Anonymous <nobody@replay.com>*
18693 * Add prototype for temp key callback functions
18694 SSL_CTX_set_tmp_{rsa,dh}_callback().
18698 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18699 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18703 * X509_name_add_entry() freed the wrong thing after an error.
18705 *Arne Ansper <arne@ats.cyber.ee>*
18707 * rsa_eay.c would attempt to free a NULL context.
18709 *Arne Ansper <arne@ats.cyber.ee>*
18711 * BIO_s_socket() had a broken should_retry() on Windoze.
18713 *Arne Ansper <arne@ats.cyber.ee>*
18715 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18717 *Arne Ansper <arne@ats.cyber.ee>*
18719 * Make sure the already existing X509_STORE->depth variable is initialized
18720 in X509_STORE_new(), but document the fact that this variable is still
18721 unused in the certificate verification process.
18723 *Ralf S. Engelschall*
18725 * Fix the various library and `apps/` files to free up pkeys obtained from
18726 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18730 * Fix reference counting in X509_PUBKEY_get(). This makes
18731 demos/maurice/example2.c work, amongst others, probably.
18733 *Steve Henson and Ben Laurie*
18735 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18736 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18737 are no longer created. This way we have a single and consistent command
18738 line interface `openssl <command>`, similar to `cvs <command>`.
18740 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18742 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18743 BIT STRING wrapper always have zero unused bits.
18747 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18751 * Make the top-level INSTALL documentation easier to understand.
18755 * Makefiles updated to exit if an error occurs in a sub-directory
18756 make (including if user presses ^C) [Paul Sutton]
18758 * Make Montgomery context stuff explicit in RSA data structure.
18762 * Fix build order of pem and err to allow for generated pem.h.
18766 * Fix renumbering bug in X509_NAME_delete_entry().
18770 * Enhanced the err-ins.pl script so it makes the error library number
18771 global and can add a library name. This is needed for external ASN1 and
18772 other error libraries.
18776 * Fixed sk_insert which never worked properly.
18780 * Fix ASN1 macros so they can handle indefinite length constructed
18781 EXPLICIT tags. Some non standard certificates use these: they can now
18786 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18787 into a single doc/ssleay.txt bundle. This way the information is still
18788 preserved but no longer messes up this directory. Now it's new room for
18789 the new set of documentation files.
18791 *Ralf S. Engelschall*
18793 * SETs were incorrectly DER encoded. This was a major pain, because they
18794 shared code with SEQUENCEs, which aren't coded the same. This means that
18795 almost everything to do with SETs or SEQUENCEs has either changed name or
18796 number of arguments.
18798 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18800 * Fix test data to work with the above.
18804 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18805 was already fixed by Eric for 0.9.1 it seems.
18807 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18809 * Autodetect FreeBSD3.
18813 * Fix various bugs in Configure. This affects the following platforms:
18817 unixware-2.0-pentium
18822 * Eliminate generated files from CVS. Reorder tests to regenerate files
18823 before they are needed.
18827 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18831 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18833 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18834 changed SSLeay to OpenSSL in version strings.
18836 *Ralf S. Engelschall*
18838 * Some fixups to the top-level documents.
18842 * Fixed the nasty bug where rsaref.h was not found under compile-time
18843 because the symlink to include/ was missing.
18845 *Ralf S. Engelschall*
18847 * Incorporated the popular no-RSA/DSA-only patches
18848 which allow to compile a RSA-free SSLeay.
18850 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18852 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18853 when "ssleay" is still not found.
18855 *Ralf S. Engelschall*
18857 * Added more platforms to Configure: Cray T3E, HPUX 11,
18859 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18861 * Updated the README file.
18863 *Ralf S. Engelschall*
18865 * Added various .cvsignore files in the CVS repository subdirs
18866 to make a "cvs update" really silent.
18868 *Ralf S. Engelschall*
18870 * Recompiled the error-definition header files and added
18871 missing symbols to the Win32 linker tables.
18873 *Ralf S. Engelschall*
18875 * Cleaned up the top-level documents;
18876 o new files: CHANGES and LICENSE
18877 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18878 o merged COPYRIGHT into LICENSE
18879 o removed obsolete TODO file
18880 o renamed MICROSOFT to INSTALL.W32
18882 *Ralf S. Engelschall*
18884 * Removed dummy files from the 0.9.1b source tree:
18886 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18887 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18888 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18889 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18890 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18893 *Ralf S. Engelschall*
18895 * Added various platform portability fixes.
18899 * The Genesis of the OpenSSL rpject:
18900 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18901 Young and Tim J. Hudson created while they were working for C2Net until
18904 *The OpenSSL Project*
18906 ### Changes between 0.9.0b and 0.9.1b [not released]
18908 * Updated a few CA certificates under certs/
18912 * Changed some BIGNUM api stuff.
18916 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18917 DGUX x86, Linux Alpha, etc.
18921 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18922 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18927 * Add -strparse option to asn1pars program which parses nested
18930 *Dr Stephen Henson <shenson@bigfoot.com>*
18932 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18936 * DSA fix for "ca" program.
18940 * Added "-genkey" option to "dsaparam" program.
18944 * Added RIPE MD160 (rmd160) message digest.
18948 * Added -a (all) option to "ssleay version" command.
18952 * Added PLATFORM define which is the id given to Configure.
18956 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18960 * Extended the ASN.1 parser routines.
18964 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18968 * Added a BN_CTX to the BN library.
18972 * Fixed the weak key values in DES library
18976 * Changed API in EVP library for cipher aliases.
18980 * Added support for RC2/64bit cipher.
18984 * Converted the lhash library to the crypto/mem.c functions.
18988 * Added more recognized ASN.1 object ids.
18992 * Added more RSA padding checks for SSL/TLS.
18996 * Added BIO proxy/filter functionality.
19000 * Added extra_certs to SSL_CTX which can be used
19001 send extra CA certificates to the client in the CA cert chain sending
19002 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19006 * Now Fortezza is denied in the authentication phase because
19007 this is key exchange mechanism is not supported by SSLeay at all.
19011 * Additional PKCS1 checks.
19015 * Support the string "TLSv1" for all TLS v1 ciphers.
19019 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19020 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19024 * Fixed a few memory leaks.
19028 * Fixed various code and comment typos.
19032 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19033 bytes sent in the client random.
19035 *Edward Bishop <ebishop@spyglass.com>*
19039 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19040 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19041 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19042 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19043 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19044 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19045 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19046 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19047 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19048 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19049 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19050 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19051 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19052 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19053 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19054 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19055 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19056 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19057 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19058 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19059 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19060 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19061 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19062 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19063 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19064 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19065 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19066 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19067 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19068 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19069 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19070 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19071 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19072 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19073 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19074 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19075 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19076 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19077 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19078 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19079 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19080 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19081 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19082 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19083 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19084 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19085 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19086 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19087 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19088 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19089 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19090 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19091 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19092 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19093 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19094 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19095 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19096 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19097 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19098 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19099 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19100 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19101 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19102 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19103 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19104 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19105 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19106 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19107 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19108 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19109 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19110 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19111 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19112 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19113 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19114 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19115 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19116 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19117 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19118 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19119 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19120 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19121 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19122 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19123 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19124 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19125 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19126 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19127 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19128 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19129 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19130 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19131 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19132 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19133 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19134 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19135 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19136 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19137 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19138 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19139 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19140 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19141 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19142 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19143 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19144 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19145 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19146 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19147 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19148 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19149 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19150 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19151 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19152 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19153 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19154 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19155 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19156 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19157 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19158 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19159 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19160 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19161 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19162 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19163 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19164 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19165 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19166 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19167 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19168 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19169 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19170 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19171 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19172 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19173 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19174 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19175 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19176 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19177 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19178 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19179 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19180 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19181 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19182 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19183 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19184 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19185 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19186 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19187 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19188 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19189 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19190 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19191 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19192 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19193 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19194 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19195 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19196 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19197 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19198 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19199 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19200 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655