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 * For the key types DH and DHX the allowed settable parameters are now different.
27 Previously (in 1.1.1) these conflicting parameters were allowed, but will now
28 result in errors. See EVP_PKEY-DH(7) for further details. This affects the
29 behaviour of openssl-genpkey(1) for DH parameter generation.
33 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
37 * Added support for Kernel TLS (KTLS). In order to use KTLS, support for it
38 must be compiled in using the "enable-ktls" compile time option. It must
39 also be enabled at run time using the SSL_OP_ENABLE_KTLS option.
41 *Boris Pismenny, John Baldwin and Andrew Gallatin*
43 * The error return values from some control calls (ctrl) have changed.
44 One significant change is that controls which used to return -2 for
45 invalid inputs, now return -1 indicating a generic error condition instead.
49 * A public key check is now performed during EVP_PKEY_derive_set_peer().
50 Previously DH was internally doing this during EVP_PKEY_derive().
51 To disable this check use EVP_PKEY_derive_set_peer_ex(dh, peer, 0). This
52 may mean that an error can occur in EVP_PKEY_derive_set_peer() rather than
53 during EVP_PKEY_derive().
57 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
58 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
59 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
60 are deprecated. They are not invoked by the OpenSSL library anymore and
61 are replaced by direct checks of the key operation against the key type
62 when the operation is initialized.
66 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
67 more key types including RSA, DSA, ED25519, X25519, ED448 and X448.
68 Previously (in 1.1.1) they would return -2. For key types that do not have
69 parameters then EVP_PKEY_param_check() will always return 1.
71 * The output from numerous "printing" functions such as X509_signature_print(),
72 X509_print_ex(), X509_CRL_print_ex(), and other similar functions has been
73 amended such that there may be cosmetic differences between the output
74 observed in 1.1.1 and 3.0. This also applies to the "-text" output from the
75 x509 and crl applications.
79 * Windows thread synchronization uses read/write primitives (SRWLock) when
80 supported by the OS, otherwise CriticalSection continues to be used.
84 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
85 work on read only BIO source/sinks that do not support these functions.
86 This allows piping or redirection of a file BIO using stdin to be buffered
87 into memory. This is used internally in OSSL_DECODER_from_bio().
91 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
92 this function would return one of the values OSSL_STORE_INFO_NAME,
93 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
94 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
95 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
96 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
97 using this function should be amended to handle the changed return value.
101 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
102 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
103 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
104 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
105 contains more than one certificate identifier: This means that all
106 certificates referenced there MUST be part of the validation chain.
110 * The implementation of the EVP ciphers CAST5-ECB, CAST5-CBC, CAST5-OFB,
111 CAST5-CFB, BF-ECB, BF-CBC, BF-OFB, BF-CFB, IDEA-ECB, IDEC-CBC, IDEA-OFB,
112 IDEA-CFB, SEED-ECB, SEED-CBC, SEED-OFB, SEED-CFB, RC2-ECB, RC2-CBC,
113 RC2-40-CBC, RC2-64-CBC, RC2-OFB, RC2-CFB, RC4, RC4-40, RC4-HMAC-MD5, RC5-ECB,
114 RC5-CBC, RC5-OFB, RC5-CFB, DESX-CBC, DES-ECB, DES-CBC, DES-OFB, DES-CFB,
115 DES-CFB1 and DES-CFB8 have been moved to the legacy provider. Applications
116 using the EVP APIs to access these ciphers should instead use more modern
117 ciphers. If that is not possible then these applications should ensure that
118 the legacy provider has been loaded. This can be achieved either
119 programmatically or via configuration. See the provider(7) man page for
124 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
125 RIPEMD-160 have been moved to the legacy provider. Applications using the
126 EVP APIs to access these digests should instead use more modern digests. If
127 that is not possible then these applications should ensure that the legacy
128 provider has been loaded. This can be achieved either programmatically or via
129 configuration. See the provider(7) man page for further details.
133 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
138 * The deprecated functions EVP_PKEY_get0_RSA(),
139 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
140 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
141 well as the similarly named "get1" functions behave slightly differently in
142 OpenSSL 3.0. Previously they returned a pointer to the low-level key used
143 internally by libcrypto. From OpenSSL 3.0 this key may now be held in a
144 provider. Calling these functions will only return a handle on the internal
145 key where the EVP_PKEY was constructed using this key in the first place, for
146 example using a function or macro such as EVP_PKEY_assign_RSA(),
147 EVP_PKEY_set1_RSA(), etc. Where the EVP_PKEY holds a provider managed key,
148 then these functions now return a cached copy of the key. Changes to
149 the internal provider key that take place after the first time the cached key
150 is accessed will not be reflected back in the cached copy. Similarly any
151 changes made to the cached copy by application code will not be reflected
152 back in the internal provider key.
154 For the above reasons the keys returned from these functions should typically
155 be treated as read-only. To emphasise this the value returned from
156 EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(),
157 EVP_PKEY_get0_EC_KEY() and EVP_PKEY_get0_DH() has been made const. This may
158 break some existing code. Applications broken by this change should be
159 modified. The preferred solution is to refactor the code to avoid the use of
160 these deprecated functions. Failing this the code should be modified to use a
161 const pointer instead. The EVP_PKEY_get1_RSA(), EVP_PKEY_get1_DSA(),
162 EVP_PKEY_get1_EC_KEY() and EVP_PKEY_get1_DH() functions continue to return a
163 non-const pointer to enable them to be "freed". However they should also be
164 treated as read-only.
168 * A number of functions handling low level keys or engines were deprecated
169 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
170 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
171 EVP_PKEY_get0_siphash(). Applications using engines should instead use
172 providers. Applications getting or setting low level keys in an EVP_PKEY
173 should instead use the OSSL_ENCODER or OSSL_DECODER APIs, or alternatively
174 use EVP_PKEY_fromdata() or EVP_PKEY_get_params().
178 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
179 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions. They are not needed
180 and require returning octet ptr parameters from providers that
181 would like to support them which complicates provider implementations.
185 * The RAND_METHOD APIs have been deprecated. The functions deprecated are:
186 RAND_OpenSSL(), RAND_get_rand_method(), RAND_set_rand_engine() and
187 RAND_set_rand_method(). Provider based random number generators should
188 be used instead via EVP_RAND(3).
192 * The SRP APIs have been deprecated. The old APIs do not work via providers,
193 and there is no EVP interface to them. Unfortunately there is no replacement
194 for these APIs at this time.
198 * Add a compile time option to prevent the caching of provider fetched
199 algorithms. This is enabled by including the no-cached-fetch option
200 at configuration time.
204 * The default algorithms for pkcs12 creation with the PKCS12_create() function
205 were changed to more modern PBKDF2 and AES based algorithms. The default
206 MAC iteration count was changed to PKCS12_DEFAULT_ITER to make it equal
207 with the password-based encryption iteration count. The default digest
208 algorithm for the MAC computation was changed to SHA-256. The pkcs12
209 application now supports -legacy option that restores the previous
210 default algorithms to support interoperability with legacy systems.
212 *Tomáš Mráz and Sahana Prasad*
214 * The openssl speed command does not use low-level API calls anymore. This
215 implies some of the performance numbers might not be fully comparable
216 with the previous releases due to higher overhead. This applies
217 particularly to measuring performance on smaller data chunks.
221 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
224 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
226 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
227 Typically if OpenSSL has no EC or DH algorithms then it cannot support
228 connections with TLSv1.3. However OpenSSL now supports "pluggable" groups
229 through providers. Therefore third party providers may supply group
230 implementations even where there are no built-in ones. Attempting to create
231 TLS connections in such a build without also disabling TLSv1.3 at run time or
232 using third party provider groups may result in handshake failures. TLSv1.3
233 can be disabled at compile time using the "no-tls1_3" Configure option.
237 * The undocumented function X509_certificate_type() has been deprecated;
238 applications can use X509_get0_pubkey() and X509_get0_signature() to
239 get the same information.
243 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range()
244 functions. They are identical to BN_rand() and BN_rand_range()
249 * Removed RSA padding mode for SSLv23 (which was only used for
250 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
251 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
256 * Deprecated the obsolete X9.31 RSA key generation related functions
257 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
258 BN_X931_generate_prime_ex().
262 * The default key generation method for the regular 2-prime RSA keys was
263 changed to the FIPS 186-4 B.3.6 method (Generation of Probable Primes with
264 Conditions Based on Auxiliary Probable Primes). This method is slower
265 than the original method.
269 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
270 They are replaced with the BN_check_prime() function that avoids possible
271 misuse and always uses at least 64 rounds of the Miller-Rabin
272 primality test. At least 64 rounds of the Miller-Rabin test are now also
273 used for all prime generation, including RSA key generation.
274 This increases key generation time, especially for larger keys.
278 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn()
279 as they are not useful with non-deprecated functions.
283 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
284 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
285 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
286 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
287 were used to collect all necessary data to form a HTTP request, and to
288 perform the HTTP transfer with that request. With OpenSSL 3.0, the
289 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
290 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
291 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
292 OSSL_HTTP_REQ_CTX_set1_req(), OSSL_HTTP_REQ_CTX_nbio(),
293 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
294 OSSL_HTTP_REQ_CTX_set_max_response_length().
296 *Rich Salz and Richard Levitte*
298 * Deprecated `OCSP_parse_url()`, which is replaced with `OSSL_HTTP_parse_url`.
302 * Validation of SM2 keys has been separated from the validation of regular EC
303 keys, allowing to improve the SM2 validation process to reject loaded private
304 keys that are not conforming to the SM2 ISO standard.
305 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
310 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
311 switches: a validation failure triggers an early exit, returning a failure
312 exit status to the parent process.
316 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
317 to ignore unknown ciphers.
321 * The `-cipher-commands` and `-digest-commands` options
322 of the command line utility `list` have been deprecated.
323 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
327 * All of the low level EC_KEY functions have been deprecated including:
329 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
330 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
331 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
332 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
333 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
334 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
335 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
336 EC_KEY_METHOD_get_verify,
337 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
338 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
339 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
340 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
341 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
342 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
343 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
344 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
345 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
346 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
347 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
348 Applications that need to implement an EC_KEY_METHOD need to consider
349 implementation of the functionality in a special provider.
350 For replacement of the functions manipulating the EC_KEY objects
351 see the EVP_PKEY-EC(7) manual page.
353 Additionally functions that read and write EC_KEY objects such as
354 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
355 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
356 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
357 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
358 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
359 have also been deprecated. Applications should instead use the
360 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
362 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
363 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
364 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
365 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
366 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
368 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
370 * Deprecated all the libcrypto and libssl error string loading
371 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
372 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
373 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
374 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
375 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
376 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
377 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
378 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
379 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
380 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
382 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
383 now loads error strings automatically.
387 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
388 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
389 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
390 are to be used by servers requiring ephemeral DH keys. Instead applications
391 should consider using the built-in DH parameters that are available by
392 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
393 necessary then applications can use the alternative functions
394 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
395 replacement for the "callback" functions. The callback was originally useful
396 in order to have different parameters for export and non-export ciphersuites.
397 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
398 functions should be replaced by one of the other methods described above.
402 * The `-crypt` option to the `passwd` command line tool has been removed.
406 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
411 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
413 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
414 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
415 The inverse ciphers use AES decryption for wrapping, and
416 AES encryption for unwrapping.
420 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
421 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
422 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
423 OpenSSL 3.0 these are replaced by the more generic functions
424 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
425 The old versions have been converted to deprecated macros that just call the
430 * The security callback, which can be customised by application code, supports
431 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
432 in the "other" parameter. In most places this is what is passed. All these
433 places occur server side. However there was one client side call of this
434 security operation and it passed a DH object instead. This is incorrect
435 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
436 of the other locations. Therefore this client side call has been changed to
437 pass an EVP_PKEY instead.
441 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
442 interface. Their functionality remains unchanged.
446 * Added new option for 'openssl list', '-providers', which will display the
447 list of loaded providers, their names, version and status. It optionally
448 displays their gettable parameters.
452 * Removed EVP_PKEY_set_alias_type(). This function was previously
453 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
454 type is internally recognised so the workaround is no longer needed.
456 This is a breaking change from previous OpenSSL versions.
460 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
461 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
465 * Changed all "STACK" functions to be macros instead of inline functions. Macro
466 parameters are still checked for type safety at compile time via helper
471 * Remove the RAND_DRBG API
473 The RAND_DRBG API did not fit well into the new provider concept as
474 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
475 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
476 and some of its API calls are rather low-level. This holds in particular
477 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
479 Adding a compatibility layer to continue supporting the RAND_DRBG API as
480 a legacy API for a regular deprecation period turned out to come at the
481 price of complicating the new provider API unnecessarily. Since the
482 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
485 *Paul Dale and Matthias St. Pierre*
487 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
488 as well as actual hostnames.
492 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
493 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
494 conversely, silently ignore DTLS protocol version bounds when configuring
495 TLS-based contexts. The commands can be repeated to set bounds of both
496 types. The same applies with the corresponding "min_protocol" and
497 "max_protocol" command-line switches, in case some application uses both TLS
500 SSL_CTX instances that are created for a fixed protocol version (e.g.
501 `TLSv1_server_method()`) also silently ignore version bounds. Previously
502 attempts to apply bounds to these protocol versions would result in an
503 error. Now only the "version-flexible" SSL_CTX instances are subject to
504 limits in configuration files in command-line options.
508 * Deprecated the `ENGINE` API. Engines should be replaced with providers
513 * Reworked the recorded ERR codes to make better space for system errors.
514 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
515 given code is a system error (true) or an OpenSSL error (false).
519 * Reworked the test perl framework to better allow parallel testing.
521 *Nicola Tuveri and David von Oheimb*
523 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
524 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
528 * 'Configure' has been changed to figure out the configuration target if
529 none is given on the command line. Consequently, the 'config' script is
530 now only a mere wrapper. All documentation is changed to only mention
533 *Rich Salz and Richard Levitte*
535 * Added a library context `OSSL_LIB_CTX` that applications as well as
536 other libraries can use to form a separate context within which
537 libcrypto operations are performed.
539 There are two ways this can be used:
541 - Directly, by passing a library context to functions that take
542 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
544 - Indirectly, by creating a new library context and then assigning
545 it as the new default, with `OSSL_LIB_CTX_set0_default`.
547 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
548 apart from the functions directly related to `OSSL_LIB_CTX`, accept
549 NULL to indicate that the default library context should be used.
551 Library code that changes the default library context using
552 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
553 second call before returning to the caller.
555 _(Note: the library context was initially called `OPENSSL_CTX` and
556 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
560 * Handshake now fails if Extended Master Secret extension is dropped
565 * Dropped interactive mode from the `openssl` program. From now on,
566 running it without arguments is equivalent to `openssl help`.
570 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
571 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
572 While the old function names have been retained for backward compatibility
573 they should not be used in new developments
574 because their return values are confusing: Unlike other `_cmp()` functions
575 they do not return 0 in case their arguments are equal.
579 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
580 `EC_GROUP_get_field_type()`.
584 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
585 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
586 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
587 Applications should rely on the library automatically assigning a suitable
588 EC_METHOD internally upon EC_GROUP construction.
592 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
593 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
594 assigned internally without application intervention.
595 Users of EC_GROUP_new() should switch to a different suitable constructor.
599 * Add CAdES-BES signature verification support, mostly derived
600 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
602 *Filipe Raimundo da Silva*
604 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
608 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
609 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS). Its purpose
610 is to support encryption and decryption of a digital envelope that is both
611 authenticated and encrypted using AES GCM mode.
615 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
616 functions are not widely used and now OpenSSL automatically perform this
617 conversion when needed.
621 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
622 EC_KEY_precompute_mult(). These functions are not widely used and
623 applications should instead switch to named curves which OpenSSL has
624 hardcoded lookup tables for.
628 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
629 should instead use the L<EC_POINT_mul(3)> function.
633 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
634 that are not applicable to the new provider model. Applications should
635 instead use EVP_default_properties_is_fips_enabled() and
636 EVP_default_properties_enable_fips().
640 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
641 is set, an unexpected EOF is ignored, it pretends a close notify was received
642 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
646 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
647 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
648 used and applications should instead use the
649 L<EC_POINT_set_affine_coordinates(3)> and
650 L<EC_POINT_get_affine_coordinates(3)> functions.
654 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
655 arrays to be more easily constructed via a series of utility functions.
656 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
657 the various push functions and finally convert to a passable OSSL_PARAM
658 array using OSSL_PARAM_BLD_to_param().
662 * The security strength of SHA1 and MD5 based signatures in TLS has been
663 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
664 working at the default security level of 1 and instead requires security
665 level 0. The security level can be changed either using the cipher string
666 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`. This also means
667 that where the signature algorithms extension is missing from a ClientHello
668 then the handshake will fail in TLS 1.2 at security level 1. This is because,
669 although this extension is optional, failing to provide one means that
670 OpenSSL will fallback to a default set of signature algorithms. This default
671 set requires the availability of SHA1.
675 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
676 contain a provider side internal key.
680 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
681 They are old functions that we don't use, and that you could disable with
682 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
686 * Project text documents not yet having a proper file name extension
687 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
688 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
692 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
693 have been converted to Markdown with the goal to produce documents
694 which not only look pretty when viewed online in the browser, but
695 remain well readable inside a plain text editor.
697 To achieve this goal, a 'minimalistic' Markdown style has been applied
698 which avoids formatting elements that interfere too much with the
699 reading flow in the text file. For example, it
701 * avoids [ATX headings][] and uses [setext headings][] instead
702 (which works for `<h1>` and `<h2>` headings only).
703 * avoids [inline links][] and uses [reference links][] instead.
704 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
706 [ATX headings]: https://github.github.com/gfm/#atx-headings
707 [setext headings]: https://github.github.com/gfm/#setext-headings
708 [inline links]: https://github.github.com/gfm/#inline-link
709 [reference links]: https://github.github.com/gfm/#reference-link
710 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
711 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
713 *Matthias St. Pierre*
715 * The test suite is changed to preserve results of each test recipe.
716 A new directory test-runs/ with subdirectories named like the
717 test recipes are created in the build tree for this purpose.
721 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
722 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
723 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
725 *David von Oheimb, Martin Peylo*
727 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
728 The legacy OCSP-focused and only partly documented API is retained for
729 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
733 * Added `util/check-format.pl`, a tool for checking adherence to the
734 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
735 The checks performed are incomplete and yield some false positives.
736 Still the tool should be useful for detecting most typical glitches.
740 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
741 If domain name resolution yields multiple IP addresses all of them are tried
742 after `connect()` failures.
746 * All of the low level RSA functions have been deprecated including:
748 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
749 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
750 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
751 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
752 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
753 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
754 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
755 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
756 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
757 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
758 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
759 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
760 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
761 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
762 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
763 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
764 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
765 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
766 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
767 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
768 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
769 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
770 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
771 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
772 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
773 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
774 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
775 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
777 Use of these low level functions has been informally discouraged for a long
778 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
779 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
780 L<EVP_PKEY_decrypt(3)>.
782 All of these low level RSA functions have been deprecated without
785 RSA_blinding_off, RSA_blinding_on, RSA_clear_flags, RSA_get_version,
786 RSAPrivateKey_dup, RSAPublicKey_dup, RSA_set_flags, RSA_setup_blinding and
789 All of these RSA flags have been deprecated without replacement:
791 RSA_FLAG_BLINDING, RSA_FLAG_CACHE_PRIVATE, RSA_FLAG_CACHE_PUBLIC,
792 RSA_FLAG_EXT_PKEY, RSA_FLAG_NO_BLINDING, RSA_FLAG_THREAD_SAFE and
793 RSA_METHOD_FLAG_NO_CHECK.
797 * X509 certificates signed using SHA1 are no longer allowed at security
799 In TLS/SSL the default security level is 1. It can be set either
800 using the cipher string with `@SECLEVEL`, or calling
801 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
802 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
804 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
805 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
806 options of the commands.
810 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
811 modified to use PKEY APIs. These commands are now in maintenance mode
812 and no new features will be added to them.
816 * The command line utility rsautl has been deprecated.
817 Instead use the pkeyutl program.
821 * The command line utilities genrsa and rsa have been modified to use PKEY
822 APIs. They now write PKCS#8 keys by default. These commands are now in
823 maintenance mode and no new features will be added to them.
827 * All of the low level DH functions have been deprecated including:
829 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
830 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
831 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
832 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
833 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
834 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
835 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
836 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
837 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
838 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
839 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
840 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
841 DH_meth_get_generate_params and DH_meth_set_generate_params.
843 Use of these low level functions has been informally discouraged for a long
844 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
845 and L<EVP_PKEY_derive(3)>.
847 These low level DH functions have been deprecated without replacement:
849 DH_clear_flags, DH_get_1024_160, DH_get_2048_224, DH_get_2048_256,
850 DH_set_flags and DH_test_flags.
852 The DH_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
853 The DH_FLAG_TYPE_DH and DH_FLAG_TYPE_DHX have been deprecated. Use
854 EVP_PKEY_is_a() to determine the type of a key. There is no replacement for
857 Additionally functions that read and write DH objects such as d2i_DHparams,
858 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
859 functions have also been deprecated. Applications should instead use the
860 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
862 Finally functions that assign or obtain DH objects from an EVP_PKEY such as
863 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
864 `EVP_PKEY_set1_DH()` are also deprecated.
865 Applications should instead either read or write an
866 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
867 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
869 *Paul Dale and Matt Caswell*
871 * All of the low level DSA functions have been deprecated including:
873 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
874 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
875 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
876 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
877 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
878 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
879 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
880 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
881 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
882 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
883 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
884 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
885 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
886 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
887 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
888 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
889 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
891 Use of these low level functions has been informally discouraged for a long
892 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
893 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
895 These low level DSA functions have been deprecated without replacement:
897 DSA_clear_flags, DSA_dup_DH, DSAparams_dup, DSA_set_flags and
900 The DSA_FLAG_CACHE_MONT_P flag has been deprecated without replacement.
902 Finally functions that assign or obtain DSA objects from an EVP_PKEY such as
903 `EVP_PKEY_assign_DSA()`, `EVP_PKEY_get0_DSA()`, `EVP_PKEY_get1_DSA()`, and
904 `EVP_PKEY_set1_DSA()` are also deprecated.
905 Applications should instead either read or write an
906 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs,
907 or load an EVP_PKEY directly from DSA data using `EVP_PKEY_fromdata()`.
911 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
912 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC. This is a breaking
913 change from previous OpenSSL versions.
915 Unlike in previous OpenSSL versions, this means that applications must not
916 call `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
917 The `EVP_PKEY_set_alias_type` function has now been removed.
919 Parameter and key generation is also reworked to make it possible
920 to generate EVP_PKEY_SM2 parameters and keys. Applications must now generate
921 SM2 keys directly and must not create an EVP_PKEY_EC key first.
925 * Deprecated low level ECDH and ECDSA functions. These include:
927 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
928 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
931 Use of these low level functions has been informally discouraged for a long
932 time. Instead applications should use the EVP_PKEY_derive(3),
933 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
937 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
938 and EVP_PKEY_decrypt() instead.
939 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
940 and EVP_PKEY_encrypt() instead.
944 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
945 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
946 a new formulation to include all the things it can be used for,
947 as well as words of caution.
951 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
952 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
956 * All of the low level HMAC functions have been deprecated including:
958 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
959 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
962 Use of these low level functions has been informally discouraged for a long
963 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
964 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
965 and L<EVP_MAC_final(3)>.
969 * Over two thousand fixes were made to the documentation, including:
970 - Common options (such as -rand/-writerand, TLS version control, etc)
971 were refactored and point to newly-enhanced descriptions in openssl.pod.
972 - Added style conformance for all options (with help from Richard Levitte),
973 documented all reported missing options, added a CI build to check
974 that all options are documented and that no unimplemented options
976 - Documented some internals, such as all use of environment variables.
977 - Addressed all internal broken L<> references.
981 * All of the low level CMAC functions have been deprecated including:
983 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
984 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
986 Use of these low level functions has been informally discouraged for a long
987 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
988 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
989 and L<EVP_MAC_final(3)>.
993 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
994 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
997 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
998 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
999 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
1000 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
1001 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
1002 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
1003 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
1004 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
1005 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
1006 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
1008 Use of these low level functions has been informally discouraged
1009 for a long time. Applications should use the EVP_DigestInit_ex(3),
1010 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
1014 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1015 set of functions. The documentation mentioned negative values for some
1016 errors, but this was never the case, so the mention of negative values
1019 Code that followed the documentation and thereby check with something
1020 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1024 * All of the low level cipher functions have been deprecated including:
1026 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
1027 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
1028 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
1029 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
1030 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
1031 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
1032 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
1033 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
1034 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
1035 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
1036 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
1037 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
1038 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
1039 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
1040 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
1041 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
1042 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
1043 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
1044 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
1045 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
1046 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
1047 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
1048 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
1049 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
1050 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
1051 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
1052 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
1053 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
1054 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
1056 Use of these low level functions has been informally discouraged for
1057 a long time. Applications should use the high level EVP APIs, e.g.
1058 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
1059 equivalently named decrypt functions instead.
1061 *Matt Caswell and Paul Dale*
1063 * Removed include/openssl/opensslconf.h.in and replaced it with
1064 include/openssl/configuration.h.in, which differs in not including
1065 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1066 was added to include both.
1068 This allows internal hacks where one might need to modify the set
1069 of configured macros, for example this if deprecated symbols are
1070 still supposed to be available internally:
1072 #include <openssl/configuration.h>
1074 #undef OPENSSL_NO_DEPRECATED
1075 #define OPENSSL_SUPPRESS_DEPRECATED
1077 #include <openssl/macros.h>
1079 This should not be used by applications that use the exported
1080 symbols, as that will lead to linking errors.
1084 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1085 used in exponentiation with 512-bit moduli. No EC algorithms are
1086 affected. Analysis suggests that attacks against 2-prime RSA1024,
1087 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1088 difficult to perform and are not believed likely. Attacks against DH512
1089 are considered just feasible. However, for an attack the target would
1090 have to re-use the DH512 private key, which is not recommended anyway.
1091 Also applications directly using the low level API BN_mod_exp may be
1092 affected if they use BN_FLG_CONSTTIME.
1097 * Most memory-debug features have been deprecated, and the functionality
1098 replaced with no-ops.
1102 * Added documentation for the STACK API.
1106 * Introduced a new method type and API, OSSL_ENCODER, to
1107 represent generic encoders. An implementation is expected to
1108 be able to encode an object associated with a given name (such
1109 as an algorithm name for an asymmetric key) into forms given by
1110 implementation properties.
1112 Encoders are primarily used from inside libcrypto, through
1113 calls to functions like EVP_PKEY_print_private(),
1114 PEM_write_bio_PrivateKey() and similar.
1116 Encoders are specified in such a way that they can be made to
1117 directly handle the provider side portion of an object, if this
1118 provider side part comes from the same provider as the encoder
1119 itself, but can also be made to handle objects in parametrized
1120 form (as an OSSL_PARAM array of data). This allows a provider to
1121 offer generic encoders as a service for any other provider.
1125 * Added a .pragma directive to the syntax of configuration files, to
1126 allow varying behavior in a supported and predictable manner.
1127 Currently added pragma:
1131 This allows dollar signs to be a keyword character unless it's
1132 followed by a opening brace or parenthesis. This is useful for
1133 platforms where dollar signs are commonly used in names, such as
1134 volume names and system directory names on VMS.
1138 * Added functionality to create an EVP_PKEY from user data. This
1139 is effectively the same as creating a RSA, DH or DSA object and
1140 then assigning them to an EVP_PKEY, but directly using algorithm
1141 agnostic EVP functions. A benefit is that this should be future
1142 proof for public key algorithms to come.
1146 * Change the interpretation of the '--api' configuration option to
1147 mean that this is a desired API compatibility level with no
1148 further meaning. The previous interpretation, that this would
1149 also mean to remove all deprecated symbols up to and including
1150 the given version, no requires that 'no-deprecated' is also used
1151 in the configuration.
1153 When building applications, the desired API compatibility level
1154 can be set with the OPENSSL_API_COMPAT macro like before. For
1155 API compatibility version below 3.0, the old style numerical
1156 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1157 For version 3.0 and on, the value is expected to be the decimal
1158 value calculated from the major and minor version like this:
1160 MAJOR * 10000 + MINOR * 100
1164 -DOPENSSL_API_COMPAT=30000 For 3.0
1165 -DOPENSSL_API_COMPAT=30200 For 3.2
1167 To hide declarations that are deprecated up to and including the
1168 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1169 given when building the application as well.
1173 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1174 access to certificate and CRL stores via URIs and OSSL_STORE
1177 This adds the following functions:
1179 - X509_LOOKUP_store()
1180 - X509_STORE_load_file()
1181 - X509_STORE_load_path()
1182 - X509_STORE_load_store()
1183 - SSL_add_store_cert_subjects_to_stack()
1184 - SSL_CTX_set_default_verify_store()
1185 - SSL_CTX_load_verify_file()
1186 - SSL_CTX_load_verify_dir()
1187 - SSL_CTX_load_verify_store()
1191 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1192 The presence of this system service is determined at run-time.
1196 * Added functionality to create an EVP_PKEY context based on data
1197 for methods from providers. This takes an algorithm name and a
1198 property query string and simply stores them, with the intent
1199 that any operation that uses this context will use those strings
1200 to fetch the needed methods implicitly, thereby making the port
1201 of application written for pre-3.0 OpenSSL easier.
1205 * The undocumented function NCONF_WIN32() has been deprecated; for
1206 conversion details see the HISTORY section of doc/man5/config.pod
1210 * Introduced the new functions EVP_DigestSignInit_ex() and
1211 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1212 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1213 pages for further details.
1217 * Over two thousand fixes were made to the documentation, including:
1218 adding missing command flags, better style conformance, documentation
1221 *Rich Salz, Richard Levitte*
1223 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1224 X25519, X448, Ed25519 and Ed448.
1228 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1233 * Deprecated the public definition of `ERR_STATE` as well as the function
1234 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1239 * Added ERR functionality to give callers access to the stored function
1240 names that have replaced the older function code based functions.
1242 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1243 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1244 ERR_peek_error_all() and ERR_peek_last_error_all().
1246 These functions have become deprecated: ERR_get_error_line(),
1247 ERR_get_error_line_data(), ERR_peek_error_line_data(),
1248 ERR_peek_last_error_line_data() and ERR_func_error_string().
1250 Users are recommended to use ERR_get_error_all(), or to pick information
1251 with ERR_peek functions and finish off with getting the error code by using
1256 * Extended testing to be verbose for failing tests only. The make variables
1257 VERBOSE_FAILURE or VF can be used to enable this:
1259 $ make VF=1 test # Unix
1260 $ mms /macro=(VF=1) test ! OpenVMS
1261 $ nmake VF=1 test # Windows
1265 * Added the `-copy_extensions` option to the `x509` command for use with
1266 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1267 all extensions in the request are copied to the certificate or vice versa.
1269 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1271 * Added the `-copy_extensions` option to the `req` command for use with
1272 `-x509`. When given with the `copy` or `copyall` argument,
1273 all extensions in the certification request are copied to the certificate.
1277 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1278 they generate are by default RFC 5280 compliant in the following sense:
1279 There is a subjectKeyIdentifier extension with a hash value of the public key
1280 and for not self-signed certs there is an authorityKeyIdentifier extension
1281 with a keyIdentifier field or issuer information identifying the signing key.
1282 This is done unless some configuration overrides the new default behavior,
1283 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1287 * Added several checks to `X509_verify_cert()` according to requirements in
1288 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1289 (which may be done by using the CLI option `-x509_strict`):
1290 * The basicConstraints of CA certificates must be marked critical.
1291 * CA certificates must explicitly include the keyUsage extension.
1292 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1293 * The issuer name of any certificate must not be empty.
1294 * The subject name of CA certs, certs with keyUsage crlSign,
1295 and certs without subjectAlternativeName must not be empty.
1296 * If a subjectAlternativeName extension is given it must not be empty.
1297 * The signatureAlgorithm field and the cert signature must be consistent.
1298 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1299 must not be marked critical.
1300 * The authorityKeyIdentifier must be given for X.509v3 certs
1301 unless they are self-signed.
1302 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1306 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1307 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1311 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1312 used even when parsing explicit parameters, when loading a encoded key
1313 or calling `EC_GROUP_new_from_ecpkparameters()`/
1314 `EC_GROUP_new_from_ecparameters()`.
1315 This prevents bypass of security hardening and performance gains,
1316 especially for curves with specialized EC_METHODs.
1317 By default, if a key encoded with explicit parameters is loaded and later
1318 encoded, the output is still encoded with explicit parameters, even if
1319 internally a "named" EC_GROUP is used for computation.
1323 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1324 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1325 NULL. After this change, only the cofactor parameter can be NULL. It also
1326 does some minimal sanity checks on the passed order.
1331 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1332 An attack is simple, if the first CMS_recipientInfo is valid but the
1333 second CMS_recipientInfo is chosen ciphertext. If the second
1334 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1335 encryption key will be replaced by garbage, and the message cannot be
1336 decoded, but if the RSA decryption fails, the correct encryption key is
1337 used and the recipient will not notice the attack.
1338 As a work around for this potential attack the length of the decrypted
1339 key must be equal to the cipher default key length, in case the
1340 certifiate is not given and all recipientInfo are tried out.
1341 The old behaviour can be re-enabled in the CMS code by setting the
1342 CMS_DEBUG_DECRYPT flag.
1346 * Early start up entropy quality from the DEVRANDOM seed source has been
1347 improved for older Linux systems. The RAND subsystem will wait for
1348 /dev/random to be producing output before seeding from /dev/urandom.
1349 The seeded state is stored for future library initialisations using
1350 a system global shared memory segment. The shared memory identifier
1351 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1352 the desired value. The default identifier is 114.
1356 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1357 when primes for RSA keys are computed.
1358 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1359 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1360 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1361 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1362 This avoids possible fingerprinting of newly generated RSA modules.
1366 * Correct the extended master secret constant on EBCDIC systems. Without this
1367 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1368 negotiate EMS will fail. Unfortunately this also means that TLS connections
1369 between EBCDIC systems with this fix, and EBCDIC systems without this
1370 fix will fail if they negotiate EMS.
1374 * Changed the library initialisation so that the config file is now loaded
1375 by default. This was already the case for libssl. It now occurs for both
1376 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1377 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1381 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1382 where the former acts as a replacement for `ERR_put_error()`, and the
1383 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1384 `ERR_raise_data()` adds more flexibility by taking a format string and
1385 an arbitrary number of arguments following it, to be processed with
1390 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1391 to check if a named provider is loaded and available. When called, it
1392 will also activate all fallback providers if such are still present.
1396 * Enforce a minimum DH modulus size of 512 bits.
1400 * Changed DH parameters to generate the order q subgroup instead of 2q.
1401 Previously generated DH parameters are still accepted by DH_check
1402 but DH_generate_key works around that by clearing bit 0 of the
1403 private key for those. This avoids leaking bit 0 of the private key.
1407 * Significantly reduce secure memory usage by the randomness pools.
1411 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1416 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1417 algorithms. An implementation of a key exchange algorithm can be obtained
1418 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1419 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1420 the older EVP_PKEY_derive_init() function. See the man pages for the new
1421 functions for further details.
1425 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1429 * Removed the function names from error messages and deprecated the
1432 * Removed NextStep support and the macro OPENSSL_UNISTD
1436 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1437 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1438 Also removed "export var as function" capability; we do not export
1439 variables, only functions.
1443 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1444 an error and 1 indicating success. In previous versions of OpenSSL this
1445 was a void type. If a key was set longer than the maximum possible this
1450 * Support SM2 signing and verification schemes with X509 certificate.
1454 * Use SHA256 as the default digest for TS query in the `ts` app.
1458 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1459 This checks that the salt length is at least 128 bits, the derived key
1460 length is at least 112 bits, and that the iteration count is at least 1000.
1461 For backwards compatibility these checks are disabled by default in the
1462 default provider, but are enabled by default in the fips provider.
1463 To enable or disable these checks use the control
1464 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1468 * Default cipher lists/suites are now available via a function, the
1469 #defines are deprecated.
1473 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1474 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1475 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1479 * Join the directories crypto/x509 and crypto/x509v3
1483 * Added command 'openssl kdf' that uses the EVP_KDF API.
1487 * Added command 'openssl mac' that uses the EVP_MAC API.
1491 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1492 as default directories. Also added the command 'openssl info'
1493 for scripting purposes.
1497 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1498 deprecated. These undocumented functions were never integrated into the EVP
1499 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1500 Bi-directional IGE mode. These modes were never formally standardised and
1501 usage of these functions is believed to be very small. In particular
1502 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1503 is ever used. The security implications are believed to be minimal, but
1504 this issue was never fixed for backwards compatibility reasons. New code
1505 should not use these modes.
1509 * Add prediction resistance to the DRBG reseeding process.
1513 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1514 mandated by IEEE Std 1619-2018.
1518 * Added newline escaping functionality to a filename when using openssl dgst.
1519 This output format is to replicate the output format found in the `*sum`
1520 checksum programs. This aims to preserve backward compatibility.
1522 *Matt Eaton, Richard Levitte, and Paul Dale*
1524 * Removed the heartbeat message in DTLS feature, as it has very
1525 little usage and doesn't seem to fulfill a valuable purpose.
1526 The configuration option is now deprecated.
1530 * Changed the output of 'openssl {digestname} < file' to display the
1531 digest name in its output.
1535 * Added a new generic trace API which provides support for enabling
1536 instrumentation through trace output. This feature is mainly intended
1537 as an aid for developers and is disabled by default. To utilize it,
1538 OpenSSL needs to be configured with the `enable-trace` option.
1540 If the tracing API is enabled, the application can activate trace output
1541 by registering BIOs as trace channels for a number of tracing and debugging
1544 The `openssl` program has been expanded to enable any of the types
1545 available via environment variables defined by the user, and serves as
1546 one possible example on how to use this functionality.
1548 *Richard Levitte & Matthias St. Pierre*
1550 * Added build tests for C++. These are generated files that only do one
1551 thing, to include one public OpenSSL head file each. This tests that
1552 the public header files can be usefully included in a C++ application.
1554 This test isn't enabled by default. It can be enabled with the option
1555 'enable-buildtest-c++'.
1559 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1563 * Add KMAC to EVP_MAC.
1567 * Added property based algorithm implementation selection framework to
1572 * Added SCA hardening for modular field inversion in EC_GROUP through
1573 a new dedicated field_inv() pointer in EC_METHOD.
1574 This also addresses a leakage affecting conversions from projective
1575 to affine coordinates.
1577 *Billy Bob Brumley, Nicola Tuveri*
1579 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1580 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1581 those algorithms that were already supported through the EVP_PKEY API
1582 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1583 and scrypt are now wrappers that call EVP_KDF.
1587 * Build devcrypto engine as a dynamic engine.
1589 *Eneas U de Queiroz*
1591 * Add keyed BLAKE2 to EVP_MAC.
1595 * Fix a bug in the computation of the endpoint-pair shared secret used
1596 by DTLS over SCTP. This breaks interoperability with older versions
1597 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1598 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1599 interoperability with such broken implementations. However, enabling
1600 this switch breaks interoperability with correct implementations.
1602 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1603 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1607 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1611 * Changed the license to the Apache License v2.0.
1615 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1617 - Major releases (indicated by incrementing the MAJOR release number)
1618 may introduce incompatible API/ABI changes.
1619 - Minor releases (indicated by incrementing the MINOR release number)
1620 may introduce new features but retain API/ABI compatibility.
1621 - Patch releases (indicated by incrementing the PATCH number)
1622 are intended for bug fixes and other improvements of existing
1623 features only (like improving performance or adding documentation)
1624 and retain API/ABI compatibility.
1628 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1632 * Remove the 'dist' target and add a tarball building script. The
1633 'dist' target has fallen out of use, and it shouldn't be
1634 necessary to configure just to create a source distribution.
1638 * Recreate the OS390-Unix config target. It no longer relies on a
1639 special script like it did for OpenSSL pre-1.1.0.
1643 * Instead of having the source directories listed in Configure, add
1644 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1649 * Add GMAC to EVP_MAC.
1653 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1657 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1658 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1659 to facilitate the continued use of MACs through raw private keys in
1660 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1664 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1665 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1669 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1670 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1671 are retained for backwards compatibility.
1675 * AES-XTS mode now enforces that its two keys are different to mitigate
1676 the attacked described in "Efficient Instantiations of Tweakable
1677 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1678 Details of this attack can be obtained from:
1679 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1683 * Rename the object files, i.e. give them other names than in previous
1684 versions. Their names now include the name of the final product, as
1685 well as its type mnemonic (bin, lib, shlib).
1689 * Added new option for 'openssl list', '-objects', which will display the
1690 list of built in objects, i.e. OIDs with names.
1694 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1695 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1700 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1701 improves application performance by removing data copies and providing
1702 applications with zero-copy system calls such as sendfile and splice.
1706 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1707 option is set, openssl cleanses (zeroize) plaintext bytes from
1708 internal buffers after delivering them to the application. Note,
1709 the application is still responsible for cleansing other copies
1710 (e.g.: data received by SSL_read(3)).
1714 * `PKCS12_parse` now maintains the order of the parsed certificates
1715 when outputting them via `*ca` (rather than reversing it).
1719 * Deprecated pthread fork support methods. These were unused so no
1720 replacement is required.
1722 - OPENSSL_fork_prepare()
1723 - OPENSSL_fork_parent()
1724 - OPENSSL_fork_child()
1728 * Added support for FFDHE key exchange in TLS 1.3.
1735 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1737 * Fixed a problem with verifying a certificate chain when using the
1738 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1739 the certificates present in a certificate chain. It is not set by default.
1741 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1742 the chain that have explicitly encoded elliptic curve parameters was added
1743 as an additional strict check.
1745 An error in the implementation of this check meant that the result of a
1746 previous check to confirm that certificates in the chain are valid CA
1747 certificates was overwritten. This effectively bypasses the check
1748 that non-CA certificates must not be able to issue other certificates.
1750 If a "purpose" has been configured then there is a subsequent opportunity
1751 for checks that the certificate is a valid CA. All of the named "purpose"
1752 values implemented in libcrypto perform this check. Therefore, where
1753 a purpose is set the certificate chain will still be rejected even when the
1754 strict flag has been used. A purpose is set by default in libssl client and
1755 server certificate verification routines, but it can be overridden or
1756 removed by an application.
1758 In order to be affected, an application must explicitly set the
1759 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1760 for the certificate verification or, in the case of TLS client or server
1761 applications, override the default purpose.
1766 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1767 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1768 renegotiation ClientHello omits the signature_algorithms extension (where it
1769 was present in the initial ClientHello), but includes a
1770 signature_algorithms_cert extension then a NULL pointer dereference will
1771 result, leading to a crash and a denial of service attack.
1773 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1774 (which is the default configuration). OpenSSL TLS clients are not impacted by
1778 *Peter Kästle and Samuel Sapalski*
1780 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1782 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1783 create a unique hash value based on the issuer and serial number data
1784 contained within an X509 certificate. However it was failing to correctly
1785 handle any errors that may occur while parsing the issuer field (which might
1786 occur if the issuer field is maliciously constructed). This may subsequently
1787 result in a NULL pointer deref and a crash leading to a potential denial of
1793 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1794 padding mode to correctly check for rollback attacks. This is considered a
1795 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1800 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1801 functions. Previously they could overflow the output length argument in some
1802 cases where the input length is close to the maximum permissable length for
1803 an integer on the platform. In such cases the return value from the function
1804 call would be 1 (indicating success), but the output length value would be
1805 negative. This could cause applications to behave incorrectly or crash.
1810 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1811 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1812 could be exploited in a side channel attack to recover the password. Since
1813 the attack is local host only this is outside of the current OpenSSL
1814 threat model and therefore no CVE is assigned.
1816 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1821 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1823 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1824 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1825 If an attacker can control both items being compared then this could lead
1826 to a possible denial of service attack. OpenSSL itself uses the
1827 GENERAL_NAME_cmp function for two purposes:
1828 1) Comparing CRL distribution point names between an available CRL and a
1829 CRL distribution point embedded in an X509 certificate
1830 2) When verifying that a timestamp response token signer matches the
1831 timestamp authority name (exposed via the API functions
1832 TS_RESP_verify_response and TS_RESP_verify_token)
1837 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1839 * Certificates with explicit curve parameters are now disallowed in
1840 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1844 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1845 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1846 conversely, silently ignore DTLS protocol version bounds when configuring
1847 TLS-based contexts. The commands can be repeated to set bounds of both
1848 types. The same applies with the corresponding "min_protocol" and
1849 "max_protocol" command-line switches, in case some application uses both TLS
1852 SSL_CTX instances that are created for a fixed protocol version (e.g.
1853 TLSv1_server_method()) also silently ignore version bounds. Previously
1854 attempts to apply bounds to these protocol versions would result in an
1855 error. Now only the "version-flexible" SSL_CTX instances are subject to
1856 limits in configuration files in command-line options.
1860 * Handshake now fails if Extended Master Secret extension is dropped
1865 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1867 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1869 * Fixed segmentation fault in SSL_check_chain()
1870 Server or client applications that call the SSL_check_chain() function
1871 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1872 dereference as a result of incorrect handling of the
1873 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1874 or unrecognised signature algorithm is received from the peer. This could
1875 be exploited by a malicious peer in a Denial of Service attack.
1880 * Added AES consttime code for no-asm configurations
1881 an optional constant time support for AES was added
1882 when building openssl for no-asm.
1883 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1884 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1885 At this time this feature is by default disabled.
1886 It will be enabled by default in 3.0.
1890 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1892 * Revert the change of EOF detection while reading in libssl to avoid
1893 regressions in applications depending on the current way of reporting
1894 the EOF. As the existing method is not fully accurate the change to
1895 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1896 branch and will be present in the 3.0 release.
1900 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1901 when primes for RSA keys are computed.
1902 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1903 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1904 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1905 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1906 This avoids possible fingerprinting of newly generated RSA modules.
1910 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1912 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1913 while reading in libssl then we would report an error back to the
1914 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1915 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1916 therefore give a hint as to what went wrong.
1920 * Check that ed25519 and ed448 are allowed by the security level. Previously
1921 signature algorithms not using an MD were not being checked that they were
1922 allowed by the security level.
1926 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1927 was not quite right. The behaviour was not consistent between resumption
1928 and normal handshakes, and also not quite consistent with historical
1929 behaviour. The behaviour in various scenarios has been clarified and
1930 it has been updated to make it match historical behaviour as closely as
1935 * *[VMS only]* The header files that the VMS compilers include automatically,
1936 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1937 that the C++ compiler doesn't understand. This is a shortcoming in the
1938 compiler, but can be worked around with `__cplusplus` guards.
1940 C++ applications that use OpenSSL libraries must be compiled using the
1941 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1942 functions. Otherwise, only functions with symbols of less than 31
1943 characters can be used, as the linker will not be able to successfully
1944 resolve symbols with longer names.
1948 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1949 The presence of this system service is determined at run-time.
1953 * Added newline escaping functionality to a filename when using openssl dgst.
1954 This output format is to replicate the output format found in the `*sum`
1955 checksum programs. This aims to preserve backward compatibility.
1957 *Matt Eaton, Richard Levitte, and Paul Dale*
1959 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1964 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1966 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1967 number generator (RNG). This was intended to include protection in the
1968 event of a fork() system call in order to ensure that the parent and child
1969 processes did not share the same RNG state. However this protection was not
1970 being used in the default case.
1972 A partial mitigation for this issue is that the output from a high
1973 precision timer is mixed into the RNG state so the likelihood of a parent
1974 and child process sharing state is significantly reduced.
1976 If an application already calls OPENSSL_init_crypto() explicitly using
1977 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1980 *Matthias St. Pierre*
1982 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1983 used even when parsing explicit parameters, when loading a encoded key
1984 or calling `EC_GROUP_new_from_ecpkparameters()`/
1985 `EC_GROUP_new_from_ecparameters()`.
1986 This prevents bypass of security hardening and performance gains,
1987 especially for curves with specialized EC_METHODs.
1988 By default, if a key encoded with explicit parameters is loaded and later
1989 encoded, the output is still encoded with explicit parameters, even if
1990 internally a "named" EC_GROUP is used for computation.
1994 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1995 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1996 NULL. After this change, only the cofactor parameter can be NULL. It also
1997 does some minimal sanity checks on the passed order.
2002 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2003 An attack is simple, if the first CMS_recipientInfo is valid but the
2004 second CMS_recipientInfo is chosen ciphertext. If the second
2005 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2006 encryption key will be replaced by garbage, and the message cannot be
2007 decoded, but if the RSA decryption fails, the correct encryption key is
2008 used and the recipient will not notice the attack.
2009 As a work around for this potential attack the length of the decrypted
2010 key must be equal to the cipher default key length, in case the
2011 certifiate is not given and all recipientInfo are tried out.
2012 The old behaviour can be re-enabled in the CMS code by setting the
2013 CMS_DEBUG_DECRYPT flag.
2018 * Early start up entropy quality from the DEVRANDOM seed source has been
2019 improved for older Linux systems. The RAND subsystem will wait for
2020 /dev/random to be producing output before seeding from /dev/urandom.
2021 The seeded state is stored for future library initialisations using
2022 a system global shared memory segment. The shared memory identifier
2023 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2024 the desired value. The default identifier is 114.
2028 * Correct the extended master secret constant on EBCDIC systems. Without this
2029 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2030 negotiate EMS will fail. Unfortunately this also means that TLS connections
2031 between EBCDIC systems with this fix, and EBCDIC systems without this
2032 fix will fail if they negotiate EMS.
2036 * Use Windows installation paths in the mingw builds
2038 Mingw isn't a POSIX environment per se, which means that Windows
2039 paths should be used for installation.
2044 * Changed DH_check to accept parameters with order q and 2q subgroups.
2045 With order 2q subgroups the bit 0 of the private key is not secret
2046 but DH_generate_key works around that by clearing bit 0 of the
2047 private key for those. This avoids leaking bit 0 of the private key.
2051 * Significantly reduce secure memory usage by the randomness pools.
2055 * Revert the DEVRANDOM_WAIT feature for Linux systems
2057 The DEVRANDOM_WAIT feature added a select() call to wait for the
2058 /dev/random device to become readable before reading from the
2059 /dev/urandom device.
2061 It turned out that this change had negative side effects on
2062 performance which were not acceptable. After some discussion it
2063 was decided to revert this feature and leave it up to the OS
2064 resp. the platform maintainer to ensure a proper initialization
2065 during early boot time.
2067 *Matthias St. Pierre*
2069 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2071 * Add build tests for C++. These are generated files that only do one
2072 thing, to include one public OpenSSL head file each. This tests that
2073 the public header files can be usefully included in a C++ application.
2075 This test isn't enabled by default. It can be enabled with the option
2076 'enable-buildtest-c++'.
2080 * Enable SHA3 pre-hashing for ECDSA and DSA.
2084 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2085 This changes the size when using the `genpkey` command when no size is given.
2086 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2087 generation commands to use 2048 bits by default.
2091 * Reorganize the manual pages to consistently have RETURN VALUES,
2092 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2093 util/fix-doc-nits accordingly.
2095 *Paul Yang, Joshua Lock*
2097 * Add the missing accessor EVP_PKEY_get0_engine()
2101 * Have commands like `s_client` and `s_server` output the signature scheme
2102 along with other cipher suite parameters when debugging.
2106 * Make OPENSSL_config() error agnostic again.
2110 * Do the error handling in RSA decryption constant time.
2114 * Prevent over long nonces in ChaCha20-Poly1305.
2116 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2117 for every encryption operation. RFC 7539 specifies that the nonce value
2118 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2119 and front pads the nonce with 0 bytes if it is less than 12
2120 bytes. However it also incorrectly allows a nonce to be set of up to 16
2121 bytes. In this case only the last 12 bytes are significant and any
2122 additional leading bytes are ignored.
2124 It is a requirement of using this cipher that nonce values are
2125 unique. Messages encrypted using a reused nonce value are susceptible to
2126 serious confidentiality and integrity attacks. If an application changes
2127 the default nonce length to be longer than 12 bytes and then makes a
2128 change to the leading bytes of the nonce expecting the new value to be a
2129 new unique nonce then such an application could inadvertently encrypt
2130 messages with a reused nonce.
2132 Additionally the ignored bytes in a long nonce are not covered by the
2133 integrity guarantee of this cipher. Any application that relies on the
2134 integrity of these ignored leading bytes of a long nonce may be further
2135 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2136 is safe because no such use sets such a long nonce value. However user
2137 applications that use this cipher directly and set a non-default nonce
2138 length to be longer than 12 bytes may be vulnerable.
2140 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2146 * Add DEVRANDOM_WAIT feature for Linux systems
2148 On older Linux systems where the getrandom() system call is not available,
2149 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2150 Contrary to getrandom(), the /dev/urandom device will not block during
2151 early boot when the kernel CSPRNG has not been seeded yet.
2153 To mitigate this known weakness, use select() to wait for /dev/random to
2154 become readable before reading from /dev/urandom.
2156 * Ensure that SM2 only uses SM3 as digest algorithm
2160 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2162 * Change the info callback signals for the start and end of a post-handshake
2163 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2164 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2165 confused by this and assume that a TLSv1.2 renegotiation has started. This
2166 can break KeyUpdate handling. Instead we no longer signal the start and end
2167 of a post handshake message exchange (although the messages themselves are
2168 still signalled). This could break some applications that were expecting
2169 the old signals. However without this KeyUpdate is not usable for many
2174 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2176 * Timing vulnerability in DSA signature generation
2178 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2179 timing side channel attack. An attacker could use variations in the signing
2180 algorithm to recover the private key.
2182 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2187 * Timing vulnerability in ECDSA signature generation
2189 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2190 timing side channel attack. An attacker could use variations in the signing
2191 algorithm to recover the private key.
2193 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2198 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2199 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2200 of two gigabytes and the error handling improved.
2202 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2203 categorized as a normal bug, not a security issue, because the DRBG reseeds
2204 automatically and is fully functional even without additional randomness
2205 provided by the application.
2207 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2209 * Add a new ClientHello callback. Provides a callback interface that gives
2210 the application the ability to adjust the nascent SSL object at the
2211 earliest stage of ClientHello processing, immediately after extensions have
2212 been collected but before they have been processed. In particular, this
2213 callback can adjust the supported TLS versions in response to the contents
2218 * Add SM2 base algorithm support.
2222 * s390x assembly pack: add (improved) hardware-support for the following
2223 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2224 aes-cfb/cfb8, aes-ecb.
2228 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2229 parameter is no longer accepted, as it leads to a corrupt table. NULL
2230 pem_str is reserved for alias entries only.
2234 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2235 step for prime curves. The new implementation is based on formulae from
2236 differential addition-and-doubling in homogeneous projective coordinates
2237 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2238 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2239 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2240 to work in projective coordinates.
2242 *Billy Bob Brumley, Nicola Tuveri*
2244 * Change generating and checking of primes so that the error rate of not
2245 being prime depends on the intended use based on the size of the input.
2246 For larger primes this will result in more rounds of Miller-Rabin.
2247 The maximal error rate for primes with more than 1080 bits is lowered
2250 *Kurt Roeckx, Annie Yousar*
2252 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2256 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2257 moving between systems, and to avoid confusion when a Windows build is
2258 done with mingw vs with MSVC. For POSIX installs, there's still a
2259 symlink or copy named 'tsget' to avoid that confusion as well.
2263 * Revert blinding in ECDSA sign and instead make problematic addition
2264 length-invariant. Switch even to fixed-length Montgomery multiplication.
2268 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2269 step for binary curves. The new implementation is based on formulae from
2270 differential addition-and-doubling in mixed Lopez-Dahab projective
2271 coordinates, modified to independently blind the operands.
2273 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2275 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2276 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2277 EC_METHODs to implement their own specialized "ladder step", to take
2278 advantage of more favorable coordinate systems or more efficient
2279 differential addition-and-doubling algorithms.
2281 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2283 * Modified the random device based seed sources to keep the relevant
2284 file descriptors open rather than reopening them on each access.
2285 This allows such sources to operate in a chroot() jail without
2286 the associated device nodes being available. This behaviour can be
2287 controlled using RAND_keep_random_devices_open().
2291 * Numerous side-channel attack mitigations have been applied. This may have
2292 performance impacts for some algorithms for the benefit of improved
2293 security. Specific changes are noted in this change log by their respective
2298 * AIX shared library support overhaul. Switch to AIX "natural" way of
2299 handling shared libraries, which means collecting shared objects of
2300 different versions and bitnesses in one common archive. This allows to
2301 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2302 doesn't affect the way 3rd party applications are linked, only how
2303 multi-version installation is managed.
2307 * Make ec_group_do_inverse_ord() more robust and available to other
2308 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2309 mitigations are applied to the fallback BN_mod_inverse().
2310 When using this function rather than BN_mod_inverse() directly, new
2311 EC cryptosystem implementations are then safer-by-default.
2315 * Add coordinate blinding for EC_POINT and implement projective
2316 coordinate blinding for generic prime curves as a countermeasure to
2317 chosen point SCA attacks.
2319 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2321 * Add blinding to ECDSA and DSA signatures to protect against side channel
2322 attacks discovered by Keegan Ryan (NCC Group).
2326 * Enforce checking in the `pkeyutl` command to ensure that the input
2327 length does not exceed the maximum supported digest length when performing
2328 a sign, verify or verifyrecover operation.
2332 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2333 I/O in combination with something like select() or poll() will hang. This
2334 can be turned off again using SSL_CTX_clear_mode().
2335 Many applications do not properly handle non-application data records, and
2336 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2337 around the problems in those applications, but can also break some.
2338 It's recommended to read the manpages about SSL_read(), SSL_write(),
2339 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2340 SSL_CTX_set_read_ahead() again.
2344 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2345 now allow empty (zero character) pass phrases.
2349 * Apply blinding to binary field modular inversion and remove patent
2350 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2354 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2355 binary and prime elliptic curves.
2359 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2360 constant time fixed point multiplication.
2364 * Revise elliptic curve scalar multiplication with timing attack
2365 defenses: ec_wNAF_mul redirects to a constant time implementation
2366 when computing fixed point and variable point multiplication (which
2367 in OpenSSL are mostly used with secret scalars in keygen, sign,
2368 ECDH derive operations).
2369 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2372 * Updated CONTRIBUTING
2376 * Updated DRBG / RAND to request nonce and additional low entropy
2377 randomness from the system.
2379 *Matthias St. Pierre*
2381 * Updated 'openssl rehash' to use OpenSSL consistent default.
2385 * Moved the load of the ssl_conf module to libcrypto, which helps
2386 loading engines that libssl uses before libssl is initialised.
2390 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2394 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2396 *Ingo Schwarze, Rich Salz*
2398 * Added output of accepting IP address and port for 'openssl s_server'
2402 * Added a new API for TLSv1.3 ciphersuites:
2403 SSL_CTX_set_ciphersuites()
2404 SSL_set_ciphersuites()
2408 * Memory allocation failures consistently add an error to the error
2413 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2414 in libcrypto when run as setuid/setgid.
2418 * Load any config file by default when libssl is used.
2422 * Added new public header file <openssl/rand_drbg.h> and documentation
2423 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2425 *Matthias St. Pierre*
2427 * QNX support removed (cannot find contributors to get their approval
2428 for the license change).
2432 * TLSv1.3 replay protection for early data has been implemented. See the
2433 SSL_read_early_data() man page for further details.
2437 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2438 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2439 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2440 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2441 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2442 configuration has been separated out. See the ciphers man page or the
2443 SSL_CTX_set_ciphersuites() man page for more information.
2447 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2448 in responder mode now supports the new "-multi" option, which
2449 spawns the specified number of child processes to handle OCSP
2450 requests. The "-timeout" option now also limits the OCSP
2451 responder's patience to wait to receive the full client request
2452 on a newly accepted connection. Child processes are respawned
2453 as needed, and the CA index file is automatically reloaded
2454 when changed. This makes it possible to run the "ocsp" responder
2455 as a long-running service, making the OpenSSL CA somewhat more
2456 feature-complete. In this mode, most diagnostic messages logged
2457 after entering the event loop are logged via syslog(3) rather than
2462 * Added support for X448 and Ed448. Heavily based on original work by
2467 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2468 objects loaded. This adds the functions OSSL_STORE_expect() and
2469 OSSL_STORE_find() as well as needed tools to construct searches and
2470 get the search data out of them.
2474 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2475 version of OpenSSL should review their configuration settings to ensure
2476 that they are still appropriate for TLSv1.3. For further information see:
2477 <https://wiki.openssl.org/index.php/TLS1.3>
2481 * Grand redesign of the OpenSSL random generator
2483 The default RAND method now utilizes an AES-CTR DRBG according to
2484 NIST standard SP 800-90Ar1. The new random generator is essentially
2485 a port of the default random generator from the OpenSSL FIPS 2.0
2486 object module. It is a hybrid deterministic random bit generator
2487 using an AES-CTR bit stream and which seeds and reseeds itself
2488 automatically using trusted system entropy sources.
2490 Some of its new features are:
2491 - Support for multiple DRBG instances with seed chaining.
2492 - The default RAND method makes use of a DRBG.
2493 - There is a public and private DRBG instance.
2494 - The DRBG instances are fork-safe.
2495 - Keep all global DRBG instances on the secure heap if it is enabled.
2496 - The public and private DRBG instance are per thread for lock free
2499 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2501 * Changed Configure so it only says what it does and doesn't dump
2502 so much data. Instead, ./configdata.pm should be used as a script
2503 to display all sorts of configuration data.
2507 * Added processing of "make variables" to Configure.
2511 * Added SHA512/224 and SHA512/256 algorithm support.
2515 * The last traces of Netware support, first removed in 1.1.0, have
2520 * Get rid of Makefile.shared, and in the process, make the processing
2521 of certain files (rc.obj, or the .def/.map/.opt files produced from
2522 the ordinal files) more visible and hopefully easier to trace and
2523 debug (or make silent).
2527 * Make it possible to have environment variable assignments as
2528 arguments to config / Configure.
2532 * Add multi-prime RSA (RFC 8017) support.
2536 * Add SM3 implemented according to GB/T 32905-2016
2537 *Jack Lloyd <jack.lloyd@ribose.com>,*
2538 *Ronald Tse <ronald.tse@ribose.com>,*
2539 *Erick Borsboom <erick.borsboom@ribose.com>*
2541 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2542 as documented in RFC6066.
2543 Based on a patch from Tomasz Moń
2545 *Filipe Raimundo da Silva*
2547 * Add SM4 implemented according to GB/T 32907-2016.
2548 *Jack Lloyd <jack.lloyd@ribose.com>,*
2549 *Ronald Tse <ronald.tse@ribose.com>,*
2550 *Erick Borsboom <erick.borsboom@ribose.com>*
2552 * Reimplement -newreq-nodes and ERR_error_string_n; the
2553 original author does not agree with the license change.
2557 * Add ARIA AEAD TLS support.
2561 * Some macro definitions to support VS6 have been removed. Visual
2562 Studio 6 has not worked since 1.1.0
2566 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2567 without clearing the errors.
2571 * Add "atfork" functions. If building on a system that without
2572 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2573 requirements. The RAND facility now uses/requires this.
2581 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2582 not possible to disable entirely. However, it's still possible to
2583 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2586 To disable, configure with 'no-ui-console'. 'no-ui' is still
2587 possible to use as an alias. Check at compile time with the
2588 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2589 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2593 * Add a STORE module, which implements a uniform and URI based reader of
2594 stores that can contain keys, certificates, CRLs and numerous other
2595 objects. The main API is loosely based on a few stdio functions,
2596 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2597 OSSL_STORE_error and OSSL_STORE_close.
2598 The implementation uses backends called "loaders" to implement arbitrary
2599 URI schemes. There is one built in "loader" for the 'file' scheme.
2603 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2604 then adjusted to work on FreeBSD 8.4 as well.
2605 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2606 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2610 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2611 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2612 error code calls like this:
2614 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2616 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2617 that can be encoded in C. For the foreseeable future, this will only
2620 *Richard Levitte and Tim Hudson*
2622 * Removed BSD cryptodev engine.
2626 * Add a build target 'build_all_generated', to build all generated files
2627 and only that. This can be used to prepare everything that requires
2628 things like perl for a system that lacks perl and then move everything
2629 to that system and do the rest of the build there.
2633 * In the UI interface, make it possible to duplicate the user data. This
2634 can be used by engines that need to retain the data for a longer time
2635 than just the call where this user data is passed.
2639 * Ignore the '-named_curve auto' value for compatibility of applications
2642 *Tomáš Mráz <tmraz@fedoraproject.org>*
2644 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2645 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2646 alerts across multiple records (some of which could be empty). In practice
2647 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2648 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2649 support this at all. Supporting it adds significant complexity to the
2650 record layer, and its removal is unlikely to cause interoperability
2655 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2656 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2657 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2662 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2663 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2665 *Richard Levitte, Andy Polyakov*
2667 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2672 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2673 platform rather than 'mingw'.
2677 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2678 success if they are asked to add an object which already exists
2679 in the store. This change cascades to other functions which load
2680 certificates and CRLs.
2684 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2685 facilitate stack unwinding even from assembly subroutines.
2689 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2690 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2694 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2695 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2696 which is the minimum version we support.
2700 * Certificate time validation (X509_cmp_time) enforces stricter
2701 compliance with RFC 5280. Fractional seconds and timezone offsets
2702 are no longer allowed.
2706 * Add support for ARIA
2710 * s_client will now send the Server Name Indication (SNI) extension by
2711 default unless the new "-noservername" option is used. The server name is
2712 based on the host provided to the "-connect" option unless overridden by
2713 using "-servername".
2717 * Add support for SipHash
2721 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2722 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2723 prevent issues where no progress is being made and the peer continually
2724 sends unrecognised record types, using up resources processing them.
2728 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2729 using the algorithm defined in
2730 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2734 * Heartbeat support has been removed; the ABI is changed for now.
2736 *Richard Levitte, Rich Salz*
2738 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2742 * The RSA "null" method, which was partially supported to avoid patent
2743 issues, has been replaced to always returns NULL.
2750 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2752 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2753 used even when parsing explicit parameters, when loading a encoded key
2754 or calling `EC_GROUP_new_from_ecpkparameters()`/
2755 `EC_GROUP_new_from_ecparameters()`.
2756 This prevents bypass of security hardening and performance gains,
2757 especially for curves with specialized EC_METHODs.
2758 By default, if a key encoded with explicit parameters is loaded and later
2759 encoded, the output is still encoded with explicit parameters, even if
2760 internally a "named" EC_GROUP is used for computation.
2764 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2765 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2766 NULL. After this change, only the cofactor parameter can be NULL. It also
2767 does some minimal sanity checks on the passed order.
2772 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2773 An attack is simple, if the first CMS_recipientInfo is valid but the
2774 second CMS_recipientInfo is chosen ciphertext. If the second
2775 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2776 encryption key will be replaced by garbage, and the message cannot be
2777 decoded, but if the RSA decryption fails, the correct encryption key is
2778 used and the recipient will not notice the attack.
2779 As a work around for this potential attack the length of the decrypted
2780 key must be equal to the cipher default key length, in case the
2781 certifiate is not given and all recipientInfo are tried out.
2782 The old behaviour can be re-enabled in the CMS code by setting the
2783 CMS_DEBUG_DECRYPT flag.
2788 * Use Windows installation paths in the mingw builds
2790 Mingw isn't a POSIX environment per se, which means that Windows
2791 paths should be used for installation.
2796 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2798 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2799 This changes the size when using the `genpkey` command when no size is given.
2800 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2801 generation commands to use 2048 bits by default.
2805 * Prevent over long nonces in ChaCha20-Poly1305.
2807 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2808 for every encryption operation. RFC 7539 specifies that the nonce value
2809 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2810 and front pads the nonce with 0 bytes if it is less than 12
2811 bytes. However it also incorrectly allows a nonce to be set of up to 16
2812 bytes. In this case only the last 12 bytes are significant and any
2813 additional leading bytes are ignored.
2815 It is a requirement of using this cipher that nonce values are
2816 unique. Messages encrypted using a reused nonce value are susceptible to
2817 serious confidentiality and integrity attacks. If an application changes
2818 the default nonce length to be longer than 12 bytes and then makes a
2819 change to the leading bytes of the nonce expecting the new value to be a
2820 new unique nonce then such an application could inadvertently encrypt
2821 messages with a reused nonce.
2823 Additionally the ignored bytes in a long nonce are not covered by the
2824 integrity guarantee of this cipher. Any application that relies on the
2825 integrity of these ignored leading bytes of a long nonce may be further
2826 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2827 is safe because no such use sets such a long nonce value. However user
2828 applications that use this cipher directly and set a non-default nonce
2829 length to be longer than 12 bytes may be vulnerable.
2831 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2837 * Added SCA hardening for modular field inversion in EC_GROUP through
2838 a new dedicated field_inv() pointer in EC_METHOD.
2839 This also addresses a leakage affecting conversions from projective
2840 to affine coordinates.
2842 *Billy Bob Brumley, Nicola Tuveri*
2844 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2845 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2849 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2853 * Remove the 'dist' target and add a tarball building script. The
2854 'dist' target has fallen out of use, and it shouldn't be
2855 necessary to configure just to create a source distribution.
2859 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2861 * Timing vulnerability in DSA signature generation
2863 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2864 timing side channel attack. An attacker could use variations in the signing
2865 algorithm to recover the private key.
2867 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2872 * Timing vulnerability in ECDSA signature generation
2874 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2875 timing side channel attack. An attacker could use variations in the signing
2876 algorithm to recover the private key.
2878 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2883 * Add coordinate blinding for EC_POINT and implement projective
2884 coordinate blinding for generic prime curves as a countermeasure to
2885 chosen point SCA attacks.
2887 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2889 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2891 * Client DoS due to large DH parameter
2893 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2894 malicious server can send a very large prime value to the client. This will
2895 cause the client to spend an unreasonably long period of time generating a
2896 key for this prime resulting in a hang until the client has finished. This
2897 could be exploited in a Denial Of Service attack.
2899 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2904 * Cache timing vulnerability in RSA Key Generation
2906 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2907 a cache timing side channel attack. An attacker with sufficient access to
2908 mount cache timing attacks during the RSA key generation process could
2909 recover the private key.
2911 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2912 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2917 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2918 parameter is no longer accepted, as it leads to a corrupt table. NULL
2919 pem_str is reserved for alias entries only.
2923 * Revert blinding in ECDSA sign and instead make problematic addition
2924 length-invariant. Switch even to fixed-length Montgomery multiplication.
2928 * Change generating and checking of primes so that the error rate of not
2929 being prime depends on the intended use based on the size of the input.
2930 For larger primes this will result in more rounds of Miller-Rabin.
2931 The maximal error rate for primes with more than 1080 bits is lowered
2934 *Kurt Roeckx, Annie Yousar*
2936 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2940 * Add blinding to ECDSA and DSA signatures to protect against side channel
2941 attacks discovered by Keegan Ryan (NCC Group).
2945 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2946 now allow empty (zero character) pass phrases.
2950 * Certificate time validation (X509_cmp_time) enforces stricter
2951 compliance with RFC 5280. Fractional seconds and timezone offsets
2952 are no longer allowed.
2956 * Fixed a text canonicalisation bug in CMS
2958 Where a CMS detached signature is used with text content the text goes
2959 through a canonicalisation process first prior to signing or verifying a
2960 signature. This process strips trailing space at the end of lines, converts
2961 line terminators to CRLF and removes additional trailing line terminators
2962 at the end of a file. A bug in the canonicalisation process meant that
2963 some characters, such as form-feed, were incorrectly treated as whitespace
2964 and removed. This is contrary to the specification (RFC5485). This fix
2965 could mean that detached text data signed with an earlier version of
2966 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2967 signed with a fixed OpenSSL may fail to verify with an earlier version of
2968 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2969 and use the "-binary" flag (for the "cms" command line application) or set
2970 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2974 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2976 * Constructed ASN.1 types with a recursive definition could exceed the stack
2978 Constructed ASN.1 types with a recursive definition (such as can be found
2979 in PKCS7) could eventually exceed the stack given malicious input with
2980 excessive recursion. This could result in a Denial Of Service attack. There
2981 are no such structures used within SSL/TLS that come from untrusted sources
2982 so this is considered safe.
2984 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2990 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2992 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2993 effectively reduced to only comparing the least significant bit of each
2994 byte. This allows an attacker to forge messages that would be considered as
2995 authenticated in an amount of tries lower than that guaranteed by the
2996 security claims of the scheme. The module can only be compiled by the
2997 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2999 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3005 * Add a build target 'build_all_generated', to build all generated files
3006 and only that. This can be used to prepare everything that requires
3007 things like perl for a system that lacks perl and then move everything
3008 to that system and do the rest of the build there.
3012 * Backport SSL_OP_NO_RENGOTIATION
3014 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3015 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3016 changes this is no longer possible in 1.1.0. Therefore the new
3017 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3018 1.1.0 to provide equivalent functionality.
3020 Note that if an application built against 1.1.0h headers (or above) is run
3021 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3022 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3026 * Removed the OS390-Unix config target. It relied on a script that doesn't
3031 * rsaz_1024_mul_avx2 overflow bug on x86_64
3033 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3034 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3035 Analysis suggests that attacks against RSA and DSA as a result of this
3036 defect would be very difficult to perform and are not believed likely.
3037 Attacks against DH1024 are considered just feasible, because most of the
3038 work necessary to deduce information about a private key may be performed
3039 offline. The amount of resources required for such an attack would be
3040 significant. However, for an attack on TLS to be meaningful, the server
3041 would have to share the DH1024 private key among multiple clients, which is
3042 no longer an option since CVE-2016-0701.
3044 This only affects processors that support the AVX2 but not ADX extensions
3045 like Intel Haswell (4th generation).
3047 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3048 was originally found via the OSS-Fuzz project.
3053 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3055 * bn_sqrx8x_internal carry bug on x86_64
3057 There is a carry propagating bug in the x86_64 Montgomery squaring
3058 procedure. No EC algorithms are affected. Analysis suggests that attacks
3059 against RSA and DSA as a result of this defect would be very difficult to
3060 perform and are not believed likely. Attacks against DH are considered just
3061 feasible (although very difficult) because most of the work necessary to
3062 deduce information about a private key may be performed offline. The amount
3063 of resources required for such an attack would be very significant and
3064 likely only accessible to a limited number of attackers. An attacker would
3065 additionally need online access to an unpatched system using the target
3066 private key in a scenario with persistent DH parameters and a private
3067 key that is shared between multiple clients.
3069 This only affects processors that support the BMI1, BMI2 and ADX extensions
3070 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3072 This issue was reported to OpenSSL by the OSS-Fuzz project.
3077 * Malformed X.509 IPAddressFamily could cause OOB read
3079 If an X.509 certificate has a malformed IPAddressFamily extension,
3080 OpenSSL could do a one-byte buffer overread. The most likely result
3081 would be an erroneous display of the certificate in text format.
3083 This issue was reported to OpenSSL by the OSS-Fuzz project.
3088 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3090 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3091 platform rather than 'mingw'.
3095 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3096 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3097 which is the minimum version we support.
3101 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3103 * Encrypt-Then-Mac renegotiation crash
3105 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3106 negotiated where it was not in the original handshake (or vice-versa) then
3107 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3108 and servers are affected.
3110 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3115 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3117 * Truncated packet could crash via OOB read
3119 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3120 cipher is being used, then a truncated packet can cause that host to
3121 perform an out-of-bounds read, usually resulting in a crash.
3123 This issue was reported to OpenSSL by Robert Święcki of Google.
3128 * Bad (EC)DHE parameters cause a client crash
3130 If a malicious server supplies bad parameters for a DHE or ECDHE key
3131 exchange then this can result in the client attempting to dereference a
3132 NULL pointer leading to a client crash. This could be exploited in a Denial
3135 This issue was reported to OpenSSL by Guido Vranken.
3140 * BN_mod_exp may produce incorrect results on x86_64
3142 There is a carry propagating bug in the x86_64 Montgomery squaring
3143 procedure. No EC algorithms are affected. Analysis suggests that attacks
3144 against RSA and DSA as a result of this defect would be very difficult to
3145 perform and are not believed likely. Attacks against DH are considered just
3146 feasible (although very difficult) because most of the work necessary to
3147 deduce information about a private key may be performed offline. The amount
3148 of resources required for such an attack would be very significant and
3149 likely only accessible to a limited number of attackers. An attacker would
3150 additionally need online access to an unpatched system using the target
3151 private key in a scenario with persistent DH parameters and a private
3152 key that is shared between multiple clients. For example this can occur by
3153 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3154 similar to CVE-2015-3193 but must be treated as a separate problem.
3156 This issue was reported to OpenSSL by the OSS-Fuzz project.
3161 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3163 * ChaCha20/Poly1305 heap-buffer-overflow
3165 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3166 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3167 crash. This issue is not considered to be exploitable beyond a DoS.
3169 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3174 * CMS Null dereference
3176 Applications parsing invalid CMS structures can crash with a NULL pointer
3177 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3178 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3179 structure callback if an attempt is made to free certain invalid encodings.
3180 Only CHOICE structures using a callback which do not handle NULL value are
3183 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3188 * Montgomery multiplication may produce incorrect results
3190 There is a carry propagating bug in the Broadwell-specific Montgomery
3191 multiplication procedure that handles input lengths divisible by, but
3192 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3193 and DH private keys are impossible. This is because the subroutine in
3194 question is not used in operations with the private key itself and an input
3195 of the attacker's direct choice. Otherwise the bug can manifest itself as
3196 transient authentication and key negotiation failures or reproducible
3197 erroneous outcome of public-key operations with specially crafted input.
3198 Among EC algorithms only Brainpool P-512 curves are affected and one
3199 presumably can attack ECDH key negotiation. Impact was not analyzed in
3200 detail, because pre-requisites for attack are considered unlikely. Namely
3201 multiple clients have to choose the curve in question and the server has to
3202 share the private key among them, neither of which is default behaviour.
3203 Even then only clients that chose the curve will be affected.
3205 This issue was publicly reported as transient failures and was not
3206 initially recognized as a security issue. Thanks to Richard Morgan for
3207 providing reproducible case.
3212 * Removed automatic addition of RPATH in shared libraries and executables,
3213 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3217 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3219 * Fix Use After Free for large message sizes
3221 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3222 message larger than approx 16k is received then the underlying buffer to
3223 store the incoming message is reallocated and moved. Unfortunately a
3224 dangling pointer to the old location is left which results in an attempt to
3225 write to the previously freed location. This is likely to result in a
3226 crash, however it could potentially lead to execution of arbitrary code.
3228 This issue only affects OpenSSL 1.1.0a.
3230 This issue was reported to OpenSSL by Robert Święcki.
3235 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3237 * OCSP Status Request extension unbounded memory growth
3239 A malicious client can send an excessively large OCSP Status Request
3240 extension. If that client continually requests renegotiation, sending a
3241 large OCSP Status Request extension each time, then there will be unbounded
3242 memory growth on the server. This will eventually lead to a Denial Of
3243 Service attack through memory exhaustion. Servers with a default
3244 configuration are vulnerable even if they do not support OCSP. Builds using
3245 the "no-ocsp" build time option are not affected.
3247 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3252 * SSL_peek() hang on empty record
3254 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3255 sends an empty record. This could be exploited by a malicious peer in a
3256 Denial Of Service attack.
3258 This issue was reported to OpenSSL by Alex Gaynor.
3263 * Excessive allocation of memory in tls_get_message_header() and
3264 dtls1_preprocess_fragment()
3266 A (D)TLS message includes 3 bytes for its length in the header for the
3267 message. This would allow for messages up to 16Mb in length. Messages of
3268 this length are excessive and OpenSSL includes a check to ensure that a
3269 peer is sending reasonably sized messages in order to avoid too much memory
3270 being consumed to service a connection. A flaw in the logic of version
3271 1.1.0 means that memory for the message is allocated too early, prior to
3272 the excessive message length check. Due to way memory is allocated in
3273 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3274 to service a connection. This could lead to a Denial of Service through
3275 memory exhaustion. However, the excessive message length check still takes
3276 place, and this would cause the connection to immediately fail. Assuming
3277 that the application calls SSL_free() on the failed connection in a timely
3278 manner then the 21Mb of allocated memory will then be immediately freed
3279 again. Therefore the excessive memory allocation will be transitory in
3280 nature. This then means that there is only a security impact if:
3282 1) The application does not call SSL_free() in a timely manner in the event
3283 that the connection fails
3285 2) The application is working in a constrained environment where there is
3286 very little free memory
3288 3) The attacker initiates multiple connection attempts such that there are
3289 multiple connections in a state where memory has been allocated for the
3290 connection; SSL_free() has not yet been called; and there is insufficient
3291 memory to service the multiple requests.
3293 Except in the instance of (1) above any Denial Of Service is likely to be
3294 transitory because as soon as the connection fails the memory is
3295 subsequently freed again in the SSL_free() call. However there is an
3296 increased risk during this period of application crashes due to the lack of
3297 memory - which would then mean a more serious Denial of Service.
3299 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3300 (CVE-2016-6307 and CVE-2016-6308)
3304 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3305 had to be removed. Primary reason is that vendor assembler can't
3306 assemble our modules with -KPIC flag. As result it, assembly
3307 support, was not even available as option. But its lack means
3308 lack of side-channel resistant code, which is incompatible with
3309 security by todays standards. Fortunately gcc is readily available
3310 prepackaged option, which we firmly point at...
3314 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3316 * Windows command-line tool supports UTF-8 opt-in option for arguments
3317 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3318 (to any value) allows Windows user to access PKCS#12 file generated
3319 with Windows CryptoAPI and protected with non-ASCII password, as well
3320 as files generated under UTF-8 locale on Linux also protected with
3325 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3326 have been disabled by default and removed from DEFAULT, just like RC4.
3327 See the RC4 item below to re-enable both.
3331 * The method for finding the storage location for the Windows RAND seed file
3332 has changed. First we check %RANDFILE%. If that is not set then we check
3333 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3334 all else fails we fall back to C:\.
3338 * The EVP_EncryptUpdate() function has had its return type changed from void
3339 to int. A return of 0 indicates and error while a return of 1 indicates
3344 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3345 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3346 off the constant time implementation for RSA, DSA and DH have been made
3347 no-ops and deprecated.
3351 * Windows RAND implementation was simplified to only get entropy by
3352 calling CryptGenRandom(). Various other RAND-related tickets
3355 *Joseph Wylie Yandle, Rich Salz*
3357 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3358 and `OPENSSL_LH_`, respectively. The old names are available
3359 with API compatibility. They new names are now completely documented.
3363 * Unify TYPE_up_ref(obj) methods signature.
3364 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3365 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3366 int (instead of void) like all others TYPE_up_ref() methods.
3367 So now these methods also check the return value of CRYPTO_atomic_add(),
3368 and the validity of object reference counter.
3370 *fdasilvayy@gmail.com*
3372 * With Windows Visual Studio builds, the .pdb files are installed
3373 alongside the installed libraries and executables. For a static
3374 library installation, ossl_static.pdb is the associate compiler
3375 generated .pdb file to be used when linking programs.
3379 * Remove openssl.spec. Packaging files belong with the packagers.
3383 * Automatic Darwin/OSX configuration has had a refresh, it will now
3384 recognise x86_64 architectures automatically. You can still decide
3385 to build for a different bitness with the environment variable
3386 KERNEL_BITS (can be 32 or 64), for example:
3388 KERNEL_BITS=32 ./config
3392 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3393 256 bit AES and HMAC with SHA256.
3397 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3401 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3405 * To enable users to have their own config files and build file templates,
3406 Configure looks in the directory indicated by the environment variable
3407 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3408 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3409 name and is used as is.
3413 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3414 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3415 X509_CERT_FILE_CTX was removed.
3419 * "shared" builds are now the default. To create only static libraries use
3420 the "no-shared" Configure option.
3424 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3425 All of these option have not worked for some while and are fundamental
3430 * Make various cleanup routines no-ops and mark them as deprecated. Most
3431 global cleanup functions are no longer required because they are handled
3432 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3433 Explicitly de-initing can cause problems (e.g. where a library that uses
3434 OpenSSL de-inits, but an application is still using it). The affected
3435 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3436 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3437 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3438 COMP_zlib_cleanup().
3442 * --strict-warnings no longer enables runtime debugging options
3443 such as REF_DEBUG. Instead, debug options are automatically
3444 enabled with '--debug' builds.
3446 *Andy Polyakov, Emilia Käsper*
3448 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3449 have been moved out of the public header files. New functions for managing
3450 these have been added.
3454 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3455 objects have been moved out of the public header files. New
3456 functions for managing these have been added.
3460 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3461 have been moved out of the public header files. New functions for managing
3462 these have been added.
3466 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3467 moved out of the public header files. New functions for managing these
3472 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3476 * Removed the mk1mf build scripts.
3480 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3481 it is always safe to #include a header now.
3485 * Removed the aged BC-32 config and all its supporting scripts
3489 * Removed support for Ultrix, Netware, and OS/2.
3493 * Add support for HKDF.
3495 *Alessandro Ghedini*
3497 * Add support for blake2b and blake2s
3501 * Added support for "pipelining". Ciphers that have the
3502 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3503 encryptions/decryptions simultaneously. There are currently no built-in
3504 ciphers with this property but the expectation is that engines will be able
3505 to offer it to significantly improve throughput. Support has been extended
3506 into libssl so that multiple records for a single connection can be
3507 processed in one go (for >=TLS 1.1).
3511 * Added the AFALG engine. This is an async capable engine which is able to
3512 offload work to the Linux kernel. In this initial version it only supports
3513 AES128-CBC. The kernel must be version 4.1.0 or greater.
3517 * OpenSSL now uses a new threading API. It is no longer necessary to
3518 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3519 are two supported threading models: pthreads and windows threads. It is
3520 also possible to configure OpenSSL at compile time for "no-threads". The
3521 old threading API should no longer be used. The functions have been
3522 replaced with "no-op" compatibility macros.
3524 *Alessandro Ghedini, Matt Caswell*
3526 * Modify behavior of ALPN to invoke callback after SNI/servername
3527 callback, such that updates to the SSL_CTX affect ALPN.
3531 * Add SSL_CIPHER queries for authentication and key-exchange.
3535 * Changes to the DEFAULT cipherlist:
3536 - Prefer (EC)DHE handshakes over plain RSA.
3537 - Prefer AEAD ciphers over legacy ciphers.
3538 - Prefer ECDSA over RSA when both certificates are available.
3539 - Prefer TLSv1.2 ciphers/PRF.
3540 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3545 * Change the ECC default curve list to be this, in order: x25519,
3546 secp256r1, secp521r1, secp384r1.
3550 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3551 disabled by default. They can be re-enabled using the
3552 enable-weak-ssl-ciphers option to Configure.
3556 * If the server has ALPN configured, but supports no protocols that the
3557 client advertises, send a fatal "no_application_protocol" alert.
3558 This behaviour is SHALL in RFC 7301, though it isn't universally
3559 implemented by other servers.
3563 * Add X25519 support.
3564 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3565 for public and private key encoding using the format documented in
3566 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3567 key generation and key derivation.
3569 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3574 * Deprecate SRP_VBASE_get_by_user.
3575 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3576 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3577 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3578 seed, even if the seed is configured.
3580 Users should use SRP_VBASE_get1_by_user instead. Note that in
3581 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3582 also that even though configuring the SRP seed attempts to hide
3583 invalid usernames by continuing the handshake with fake
3584 credentials, this behaviour is not constant time and no strong
3585 guarantees are made that the handshake is indistinguishable from
3586 that of a valid user.
3590 * Configuration change; it's now possible to build dynamic engines
3591 without having to build shared libraries and vice versa. This
3592 only applies to the engines in `engines/`, those in `crypto/engine/`
3593 will always be built into libcrypto (i.e. "static").
3595 Building dynamic engines is enabled by default; to disable, use
3596 the configuration option "disable-dynamic-engine".
3598 The only requirements for building dynamic engines are the
3599 presence of the DSO module and building with position independent
3600 code, so they will also automatically be disabled if configuring
3601 with "disable-dso" or "disable-pic".
3603 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3604 are also taken away from openssl/opensslconf.h, as they are
3609 * Configuration change; if there is a known flag to compile
3610 position independent code, it will always be applied on the
3611 libcrypto and libssl object files, and never on the application
3612 object files. This means other libraries that use routines from
3613 libcrypto / libssl can be made into shared libraries regardless
3614 of how OpenSSL was configured.
3616 If this isn't desirable, the configuration options "disable-pic"
3617 or "no-pic" can be used to disable the use of PIC. This will
3618 also disable building shared libraries and dynamic engines.
3622 * Removed JPAKE code. It was experimental and has no wide use.
3626 * The INSTALL_PREFIX Makefile variable has been renamed to
3627 DESTDIR. That makes for less confusion on what this variable
3628 is for. Also, the configuration option --install_prefix is
3633 * Heartbeat for TLS has been removed and is disabled by default
3634 for DTLS; configure with enable-heartbeats. Code that uses the
3635 old #define's might need to be updated.
3637 *Emilia Käsper, Rich Salz*
3639 * Rename REF_CHECK to REF_DEBUG.
3643 * New "unified" build system
3645 The "unified" build system is aimed to be a common system for all
3646 platforms we support. With it comes new support for VMS.
3648 This system builds supports building in a different directory tree
3649 than the source tree. It produces one Makefile (for unix family
3650 or lookalikes), or one descrip.mms (for VMS).
3652 The source of information to make the Makefile / descrip.mms is
3653 small files called 'build.info', holding the necessary
3654 information for each directory with source to compile, and a
3655 template in Configurations, like unix-Makefile.tmpl or
3658 With this change, the library names were also renamed on Windows
3659 and on VMS. They now have names that are closer to the standard
3660 on Unix, and include the major version number, and in certain
3661 cases, the architecture they are built for. See "Notes on shared
3662 libraries" in INSTALL.
3664 We rely heavily on the perl module Text::Template.
3668 * Added support for auto-initialisation and de-initialisation of the library.
3669 OpenSSL no longer requires explicit init or deinit routines to be called,
3670 except in certain circumstances. See the OPENSSL_init_crypto() and
3671 OPENSSL_init_ssl() man pages for further information.
3675 * The arguments to the DTLSv1_listen function have changed. Specifically the
3676 "peer" argument is now expected to be a BIO_ADDR object.
3678 * Rewrite of BIO networking library. The BIO library lacked consistent
3679 support of IPv6, and adding it required some more extensive
3680 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3681 which hold all types of addresses and chains of address information.
3682 It also introduces a new API, with functions like BIO_socket,
3683 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3684 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3685 have been adapted accordingly.
3689 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3694 * CRIME protection: disable compression by default, even if OpenSSL is
3695 compiled with zlib enabled. Applications can still enable compression
3696 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3697 using the SSL_CONF library to configure compression.
3701 * The signature of the session callback configured with
3702 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3703 was explicitly marked as `const unsigned char*` instead of
3708 * Always DPURIFY. Remove the use of uninitialized memory in the
3709 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3713 * Removed many obsolete configuration items, including
3714 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3715 MD2_CHAR, MD2_INT, MD2_LONG
3717 IDEA_SHORT, IDEA_LONG
3718 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3720 *Rich Salz, with advice from Andy Polyakov*
3722 * Many BN internals have been moved to an internal header file.
3724 *Rich Salz with help from Andy Polyakov*
3726 * Configuration and writing out the results from it has changed.
3727 Files such as Makefile include/openssl/opensslconf.h and are now
3728 produced through general templates, such as Makefile.in and
3729 crypto/opensslconf.h.in and some help from the perl module
3732 Also, the center of configuration information is no longer
3733 Makefile. Instead, Configure produces a perl module in
3734 configdata.pm which holds most of the config data (in the hash
3735 table %config), the target data that comes from the target
3736 configuration in one of the `Configurations/*.conf` files (in
3741 * To clarify their intended purposes, the Configure options
3742 --prefix and --openssldir change their semantics, and become more
3743 straightforward and less interdependent.
3745 --prefix shall be used exclusively to give the location INSTALLTOP
3746 where programs, scripts, libraries, include files and manuals are
3747 going to be installed. The default is now /usr/local.
3749 --openssldir shall be used exclusively to give the default
3750 location OPENSSLDIR where certificates, private keys, CRLs are
3751 managed. This is also where the default openssl.cnf gets
3753 If the directory given with this option is a relative path, the
3754 values of both the --prefix value and the --openssldir value will
3755 be combined to become OPENSSLDIR.
3756 The default for --openssldir is INSTALLTOP/ssl.
3758 Anyone who uses --openssldir to specify where OpenSSL is to be
3759 installed MUST change to use --prefix instead.
3763 * The GOST engine was out of date and therefore it has been removed. An up
3764 to date GOST engine is now being maintained in an external repository.
3765 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3766 support for GOST ciphersuites (these are only activated if a GOST engine
3771 * EGD is no longer supported by default; use enable-egd when
3774 *Ben Kaduk and Rich Salz*
3776 * The distribution now has Makefile.in files, which are used to
3777 create Makefile's when Configure is run. *Configure must be run
3778 before trying to build now.*
3782 * The return value for SSL_CIPHER_description() for error conditions
3787 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3789 Obtaining and performing DNSSEC validation of TLSA records is
3790 the application's responsibility. The application provides
3791 the TLSA records of its choice to OpenSSL, and these are then
3792 used to authenticate the peer.
3794 The TLSA records need not even come from DNS. They can, for
3795 example, be used to implement local end-entity certificate or
3796 trust-anchor "pinning", where the "pin" data takes the form
3797 of TLSA records, which can augment or replace verification
3798 based on the usual WebPKI public certification authorities.
3802 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3803 continues to support deprecated interfaces in default builds.
3804 However, applications are strongly advised to compile their
3805 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3806 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3807 or the 1.1.0 releases.
3809 In environments in which all applications have been ported to
3810 not use any deprecated interfaces OpenSSL's Configure script
3811 should be used with the --api=1.1.0 option to entirely remove
3812 support for the deprecated features from the library and
3813 unconditionally disable them in the installed headers.
3814 Essentially the same effect can be achieved with the "no-deprecated"
3815 argument to Configure, except that this will always restrict
3816 the build to just the latest API, rather than a fixed API
3819 As applications are ported to future revisions of the API,
3820 they should update their compile-time OPENSSL_API_COMPAT define
3821 accordingly, but in most cases should be able to continue to
3822 compile with later releases.
3824 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3825 0x10000000L and 0x00908000L, respectively. However those
3826 versions did not support the OPENSSL_API_COMPAT feature, and
3827 so applications are not typically tested for explicit support
3828 of just the undeprecated features of either release.
3832 * Add support for setting the minimum and maximum supported protocol.
3833 It can bet set via the SSL_set_min_proto_version() and
3834 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3835 MaxProtocol. It's recommended to use the new APIs to disable
3836 protocols instead of disabling individual protocols using
3837 SSL_set_options() or SSL_CONF's Protocol. This change also
3838 removes support for disabling TLS 1.2 in the OpenSSL TLS
3839 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3843 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3847 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3848 and integrates ECDSA and ECDH functionality into EC. Implementations can
3849 now redirect key generation and no longer need to convert to or from
3852 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3853 include the ec.h header file instead.
3857 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3858 ciphers who are no longer supported and drops support the ephemeral RSA key
3859 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3863 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3864 opaque. For HMAC_CTX, the following constructors and destructors
3867 HMAC_CTX *HMAC_CTX_new(void);
3868 void HMAC_CTX_free(HMAC_CTX *ctx);
3870 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3871 destroy such methods has been added. See EVP_MD_meth_new(3) and
3872 EVP_CIPHER_meth_new(3) for documentation.
3875 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3876 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3877 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3878 an already created structure.
3879 2) For consistency with the majority of our object creators and
3880 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3881 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3882 for deprecated builds.
3886 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3887 cryptographic operations to be performed asynchronously as long as an
3888 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3889 further details. Libssl has also had this capability integrated with the
3890 introduction of the new mode SSL_MODE_ASYNC and associated error
3891 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3892 pages. This work was developed in partnership with Intel Corp.
3896 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3897 always enabled now. If you want to disable the support you should
3898 exclude it using the list of supported ciphers. This also means that the
3899 "-no_ecdhe" option has been removed from s_server.
3903 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3904 SSL_{CTX_}set1_curves() which can set a list.
3908 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3909 curve you want to support using SSL_{CTX_}set1_curves().
3913 * State machine rewrite. The state machine code has been significantly
3914 refactored in order to remove much duplication of code and solve issues
3915 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3916 further details). This change does have some associated API changes.
3917 Notably the SSL_state() function has been removed and replaced by
3918 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3919 SSL_set_state() has been removed altogether. The previous handshake states
3920 defined in ssl.h and ssl3.h have also been removed.
3924 * All instances of the string "ssleay" in the public API were replaced
3925 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3926 Some error codes related to internal RSA_eay API's were renamed.
3930 * The demo files in crypto/threads were moved to demo/threads.
3934 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3937 *Matt Caswell, Rich Salz*
3939 * New ASN.1 embed macro.
3941 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3942 structure is not allocated: it is part of the parent. That is instead of
3950 This reduces memory fragmentation and make it impossible to accidentally
3951 set a mandatory field to NULL.
3953 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3954 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3955 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3960 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3964 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3965 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3966 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3967 DES and RC4 ciphersuites.
3971 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3972 This changes the decoding behaviour for some invalid messages,
3973 though the change is mostly in the more lenient direction, and
3974 legacy behaviour is preserved as much as possible.
3978 * Fix no-stdio build.
3979 *David Woodhouse <David.Woodhouse@intel.com> and also*
3980 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3982 * New testing framework
3983 The testing framework has been largely rewritten and is now using
3984 perl and the perl modules Test::Harness and an extended variant of
3985 Test::More called OpenSSL::Test to do its work. All test scripts in
3986 test/ have been rewritten into test recipes, and all direct calls to
3987 executables in test/Makefile have become individual recipes using the
3988 simplified testing OpenSSL::Test::Simple.
3990 For documentation on our testing modules, do:
3992 perldoc test/testlib/OpenSSL/Test/Simple.pm
3993 perldoc test/testlib/OpenSSL/Test.pm
3997 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3998 are used; the latter aborts on memory leaks (usually checked on exit).
3999 Some undocumented "set malloc, etc., hooks" functions were removed
4000 and others were changed. All are now documented.
4004 * In DSA_generate_parameters_ex, if the provided seed is too short,
4007 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4009 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4010 from RFC4279, RFC4785, RFC5487, RFC5489.
4012 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4013 original RSA_PSK patch.
4017 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4018 era flag was never set throughout the codebase (only read). Also removed
4019 SSL3_FLAGS_POP_BUFFER which was only used if
4020 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4024 * Changed the default name options in the "ca", "crl", "req" and "x509"
4025 to be "oneline" instead of "compat".
4029 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4030 not aware of clients that still exhibit this bug, and the workaround
4031 hasn't been working properly for a while.
4035 * The return type of BIO_number_read() and BIO_number_written() as well as
4036 the corresponding num_read and num_write members in the BIO structure has
4037 changed from unsigned long to uint64_t. On platforms where an unsigned
4038 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4043 * Given the pervasive nature of TLS extensions it is inadvisable to run
4044 OpenSSL without support for them. It also means that maintaining
4045 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4046 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4050 * Removed support for the two export grade static DH ciphersuites
4051 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4052 were newly added (along with a number of other static DH ciphersuites) to
4053 1.0.2. However the two export ones have *never* worked since they were
4054 introduced. It seems strange in any case to be adding new export
4055 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4059 * Version negotiation has been rewritten. In particular SSLv23_method(),
4060 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4061 and turned into macros which simply call the new preferred function names
4062 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4063 should use the new names instead. Also as part of this change the ssl23.h
4064 header file has been removed.
4068 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4069 code and the associated standard is no longer considered fit-for-purpose.
4073 * RT2547 was closed. When generating a private key, try to make the
4074 output file readable only by the owner. This behavior change might
4075 be noticeable when interacting with other software.
4077 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4082 * Added HTTP GET support to the ocsp command.
4086 * Changed default digest for the dgst and enc commands from MD5 to
4091 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4095 * Added support for TLS extended master secret from
4096 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4097 initial patch which was a great help during development.
4101 * All libssl internal structures have been removed from the public header
4102 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4103 now redundant). Users should not attempt to access internal structures
4104 directly. Instead they should use the provided API functions.
4108 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4109 Access to deprecated functions can be re-enabled by running config with
4110 "enable-deprecated". In addition applications wishing to use deprecated
4111 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4112 will, by default, disable some transitive includes that previously existed
4113 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4117 * Added support for OCB mode. OpenSSL has been granted a patent license
4118 compatible with the OpenSSL license for use of OCB. Details are available
4119 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4120 for OCB can be removed by calling config with no-ocb.
4124 * SSLv2 support has been removed. It still supports receiving a SSLv2
4125 compatible client hello.
4129 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4130 done while fixing the error code for the key-too-small case.
4132 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4134 * CA.sh has been removed; use CA.pl instead.
4138 * Removed old DES API.
4142 * Remove various unsupported platforms:
4148 Sinix/ReliantUNIX RM400
4153 16-bit platforms such as WIN16
4157 * Clean up OPENSSL_NO_xxx #define's
4158 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4159 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4160 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4161 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4162 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4163 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4164 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4165 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4166 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4167 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4171 * Cleaned up dead code
4172 Remove all but one '#ifdef undef' which is to be looked at.
4176 * Clean up calling of xxx_free routines.
4177 Just like free(), fix most of the xxx_free routines to accept
4178 NULL. Remove the non-null checks from callers. Save much code.
4182 * Add secure heap for storage of private keys (when possible).
4183 Add BIO_s_secmem(), CBIGNUM, etc.
4184 Contributed by Akamai Technologies under our Corporate CLA.
4188 * Experimental support for a new, fast, unbiased prime candidate generator,
4189 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4191 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4193 * New output format NSS in the sess_id command line tool. This allows
4194 exporting the session id and the master key in NSS keylog format.
4196 *Martin Kaiser <martin@kaiser.cx>*
4198 * Harmonize version and its documentation. -f flag is used to display
4201 *mancha <mancha1@zoho.com>*
4203 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4204 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4206 *mancha <mancha1@zoho.com>*
4208 * Fix some double frees. These are not thought to be exploitable.
4210 *mancha <mancha1@zoho.com>*
4212 * A missing bounds check in the handling of the TLS heartbeat extension
4213 can be used to reveal up to 64k of memory to a connected client or
4216 Thanks for Neel Mehta of Google Security for discovering this bug and to
4217 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4218 preparing the fix ([CVE-2014-0160])
4220 *Adam Langley, Bodo Moeller*
4222 * Fix for the attack described in the paper "Recovering OpenSSL
4223 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4224 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4225 <http://eprint.iacr.org/2014/140>
4227 Thanks to Yuval Yarom and Naomi Benger for discovering this
4228 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4230 *Yuval Yarom and Naomi Benger*
4232 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4233 this fixes a limitation in previous versions of OpenSSL.
4237 * Experimental encrypt-then-mac support.
4239 Experimental support for encrypt then mac from
4240 draft-gutmann-tls-encrypt-then-mac-02.txt
4242 To enable it set the appropriate extension number (0x42 for the test
4243 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4245 For non-compliant peers (i.e. just about everything) this should have no
4248 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4252 * Add EVP support for key wrapping algorithms, to avoid problems with
4253 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4254 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4255 algorithms and include tests cases.
4259 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4264 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4265 MGF1 digest and OAEP label.
4269 * Make openssl verify return errors.
4271 *Chris Palmer <palmer@google.com> and Ben Laurie*
4273 * New function ASN1_TIME_diff to calculate the difference between two
4274 ASN1_TIME structures or one structure and the current time.
4278 * Update fips_test_suite to support multiple command line options. New
4279 test to induce all self test errors in sequence and check expected
4284 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4285 sign or verify all in one operation.
4289 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4290 test programs and fips_test_suite. Includes functionality to parse
4291 the minimal script output of fipsalgest.pl directly.
4295 * Add authorisation parameter to FIPS_module_mode_set().
4299 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4303 * Use separate DRBG fields for internal and external flags. New function
4304 FIPS_drbg_health_check() to perform on demand health checking. Add
4305 generation tests to fips_test_suite with reduced health check interval to
4306 demonstrate periodic health checking. Add "nodh" option to
4307 fips_test_suite to skip very slow DH test.
4311 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4316 * More extensive health check for DRBG checking many more failure modes.
4317 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4318 combination: call this in fips_test_suite.
4322 * Add support for canonical generation of DSA parameter 'g'. See
4325 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4326 POST to handle HMAC cases.
4330 * Add functions FIPS_module_version() and FIPS_module_version_text()
4331 to return numerical and string versions of the FIPS module number.
4335 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4336 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4337 outside the validated module in the FIPS capable OpenSSL.
4341 * Minor change to DRBG entropy callback semantics. In some cases
4342 there is no multiple of the block length between min_len and
4343 max_len. Allow the callback to return more than max_len bytes
4344 of entropy but discard any extra: it is the callback's responsibility
4345 to ensure that the extra data discarded does not impact the
4346 requested amount of entropy.
4350 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4351 information in FIPS186-3, SP800-57 and SP800-131A.
4355 * CCM support via EVP. Interface is very similar to GCM case except we
4356 must supply all data in one chunk (i.e. no update, final) and the
4357 message length must be supplied if AAD is used. Add algorithm test
4362 * Initial version of POST overhaul. Add POST callback to allow the status
4363 of POST to be monitored and/or failures induced. Modify fips_test_suite
4364 to use callback. Always run all selftests even if one fails.
4368 * XTS support including algorithm test driver in the fips_gcmtest program.
4369 Note: this does increase the maximum key length from 32 to 64 bytes but
4370 there should be no binary compatibility issues as existing applications
4371 will never use XTS mode.
4375 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4376 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4377 performs algorithm blocking for unapproved PRNG types. Also do not
4378 set PRNG type in FIPS_mode_set(): leave this to the application.
4379 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4380 the standard OpenSSL PRNG: set additional data to a date time vector.
4384 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4385 This shouldn't present any incompatibility problems because applications
4386 shouldn't be using these directly and any that are will need to rethink
4387 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4391 * Extensive self tests and health checking required by SP800-90 DRBG.
4392 Remove strength parameter from FIPS_drbg_instantiate and always
4393 instantiate at maximum supported strength.
4397 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4401 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4405 * New function DH_compute_key_padded() to compute a DH key and pad with
4406 leading zeroes if needed: this complies with SP800-56A et al.
4410 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4411 anything, incomplete, subject to change and largely untested at present.
4415 * Modify fipscanisteronly build option to only build the necessary object
4416 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4420 * Add experimental option FIPSSYMS to give all symbols in
4421 fipscanister.o and FIPS or fips prefix. This will avoid
4422 conflicts with future versions of OpenSSL. Add perl script
4423 util/fipsas.pl to preprocess assembly language source files
4424 and rename any affected symbols.
4428 * Add selftest checks and algorithm block of non-fips algorithms in
4429 FIPS mode. Remove DES2 from selftests.
4433 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4434 return internal method without any ENGINE dependencies. Add new
4435 tiny fips sign and verify functions.
4439 * New build option no-ec2m to disable characteristic 2 code.
4443 * New build option "fipscanisteronly". This only builds fipscanister.o
4444 and (currently) associated fips utilities. Uses the file Makefile.fips
4445 instead of Makefile.org as the prototype.
4449 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4450 Update fips_gcmtest to use IV generator.
4454 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4455 setting output buffer to NULL. The `*Final` function must be
4456 called although it will not retrieve any additional data. The tag
4457 can be set or retrieved with a ctrl. The IV length is by default 12
4458 bytes (96 bits) but can be set to an alternative value. If the IV
4459 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4464 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4465 underlying do_cipher function handles all cipher semantics itself
4466 including padding and finalisation. This is useful if (for example)
4467 an ENGINE cipher handles block padding itself. The behaviour of
4468 do_cipher is subtly changed if this flag is set: the return value
4469 is the number of characters written to the output buffer (zero is
4470 no longer an error code) or a negative error code. Also if the
4471 input buffer is NULL and length 0 finalisation should be performed.
4475 * If a candidate issuer certificate is already part of the constructed
4476 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4480 * Improve forward-security support: add functions
4482 void SSL_CTX_set_not_resumable_session_callback(
4483 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4484 void SSL_set_not_resumable_session_callback(
4485 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4487 for use by SSL/TLS servers; the callback function will be called whenever a
4488 new session is created, and gets to decide whether the session may be
4489 cached to make it resumable (return 0) or not (return 1). (As by the
4490 SSL/TLS protocol specifications, the session_id sent by the server will be
4491 empty to indicate that the session is not resumable; also, the server will
4492 not generate RFC 4507 (RFC 5077) session tickets.)
4494 A simple reasonable callback implementation is to return is_forward_secure.
4495 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4496 by the SSL/TLS server library, indicating whether it can provide forward
4499 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4501 * New -verify_name option in command line utilities to set verification
4506 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4507 Add CMAC pkey methods.
4511 * Experimental renegotiation in s_server -www mode. If the client
4512 browses /reneg connection is renegotiated. If /renegcert it is
4513 renegotiated requesting a certificate.
4517 * Add an "external" session cache for debugging purposes to s_server. This
4518 should help trace issues which normally are only apparent in deployed
4519 multi-process servers.
4523 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4524 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4525 BIO_set_cipher() and some obscure PEM functions were changed so they
4526 can now return an error. The RAND changes required a change to the
4527 RAND_METHOD structure.
4531 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4532 a gcc attribute to warn if the result of a function is ignored. This
4533 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4534 whose return value is often ignored.
4538 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4539 These allow SCTs (signed certificate timestamps) to be requested and
4540 validated when establishing a connection.
4542 *Rob Percival <robpercival@google.com>*
4547 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4549 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4550 used even when parsing explicit parameters, when loading a encoded key
4551 or calling `EC_GROUP_new_from_ecpkparameters()`/
4552 `EC_GROUP_new_from_ecparameters()`.
4553 This prevents bypass of security hardening and performance gains,
4554 especially for curves with specialized EC_METHODs.
4555 By default, if a key encoded with explicit parameters is loaded and later
4556 encoded, the output is still encoded with explicit parameters, even if
4557 internally a "named" EC_GROUP is used for computation.
4561 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4562 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4563 NULL. After this change, only the cofactor parameter can be NULL. It also
4564 does some minimal sanity checks on the passed order.
4569 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4570 An attack is simple, if the first CMS_recipientInfo is valid but the
4571 second CMS_recipientInfo is chosen ciphertext. If the second
4572 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4573 encryption key will be replaced by garbage, and the message cannot be
4574 decoded, but if the RSA decryption fails, the correct encryption key is
4575 used and the recipient will not notice the attack.
4576 As a work around for this potential attack the length of the decrypted
4577 key must be equal to the cipher default key length, in case the
4578 certifiate is not given and all recipientInfo are tried out.
4579 The old behaviour can be re-enabled in the CMS code by setting the
4580 CMS_DEBUG_DECRYPT flag.
4585 * Document issue with installation paths in diverse Windows builds
4587 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4588 binaries and run-time config file.
4593 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4595 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4596 This changes the size when using the `genpkey` command when no size is given.
4597 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4598 generation commands to use 2048 bits by default.
4602 * Add FIPS support for Android Arm 64-bit
4604 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4605 Module in Version 2.0.10. For some reason, the corresponding target
4606 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4607 built with FIPS support on Android Arm 64-bit. This omission has been
4610 *Matthias St. Pierre*
4612 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4614 * 0-byte record padding oracle
4616 If an application encounters a fatal protocol error and then calls
4617 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4618 then OpenSSL can respond differently to the calling application if a 0 byte
4619 record is received with invalid padding compared to if a 0 byte record is
4620 received with an invalid MAC. If the application then behaves differently
4621 based on that in a way that is detectable to the remote peer, then this
4622 amounts to a padding oracle that could be used to decrypt data.
4624 In order for this to be exploitable "non-stitched" ciphersuites must be in
4625 use. Stitched ciphersuites are optimised implementations of certain
4626 commonly used ciphersuites. Also the application must call SSL_shutdown()
4627 twice even if a protocol error has occurred (applications should not do
4628 this but some do anyway).
4630 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4631 Aviram, with additional investigation by Steven Collison and Andrew
4632 Hourselt. It was reported to OpenSSL on 10th December 2018.
4637 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4641 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4643 * Microarchitecture timing vulnerability in ECC scalar multiplication
4645 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4646 shown to be vulnerable to a microarchitecture timing side channel attack.
4647 An attacker with sufficient access to mount local timing attacks during
4648 ECDSA signature generation could recover the private key.
4650 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4651 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4657 * Timing vulnerability in DSA signature generation
4659 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4660 timing side channel attack. An attacker could use variations in the signing
4661 algorithm to recover the private key.
4663 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4668 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4669 Module, accidentally introduced while backporting security fixes from the
4670 development branch and hindering the use of ECC in FIPS mode.
4674 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4676 * Client DoS due to large DH parameter
4678 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4679 malicious server can send a very large prime value to the client. This will
4680 cause the client to spend an unreasonably long period of time generating a
4681 key for this prime resulting in a hang until the client has finished. This
4682 could be exploited in a Denial Of Service attack.
4684 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4689 * Cache timing vulnerability in RSA Key Generation
4691 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4692 a cache timing side channel attack. An attacker with sufficient access to
4693 mount cache timing attacks during the RSA key generation process could
4694 recover the private key.
4696 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4697 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4702 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4703 parameter is no longer accepted, as it leads to a corrupt table. NULL
4704 pem_str is reserved for alias entries only.
4708 * Revert blinding in ECDSA sign and instead make problematic addition
4709 length-invariant. Switch even to fixed-length Montgomery multiplication.
4713 * Change generating and checking of primes so that the error rate of not
4714 being prime depends on the intended use based on the size of the input.
4715 For larger primes this will result in more rounds of Miller-Rabin.
4716 The maximal error rate for primes with more than 1080 bits is lowered
4719 *Kurt Roeckx, Annie Yousar*
4721 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4725 * Add blinding to ECDSA and DSA signatures to protect against side channel
4726 attacks discovered by Keegan Ryan (NCC Group).
4730 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4731 now allow empty (zero character) pass phrases.
4735 * Certificate time validation (X509_cmp_time) enforces stricter
4736 compliance with RFC 5280. Fractional seconds and timezone offsets
4737 are no longer allowed.
4741 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4743 * Constructed ASN.1 types with a recursive definition could exceed the stack
4745 Constructed ASN.1 types with a recursive definition (such as can be found
4746 in PKCS7) could eventually exceed the stack given malicious input with
4747 excessive recursion. This could result in a Denial Of Service attack. There
4748 are no such structures used within SSL/TLS that come from untrusted sources
4749 so this is considered safe.
4751 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4757 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4759 * Read/write after SSL object in error state
4761 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4762 mechanism. The intent was that if a fatal error occurred during a handshake
4763 then OpenSSL would move into the error state and would immediately fail if
4764 you attempted to continue the handshake. This works as designed for the
4765 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4766 SSL_connect()), however due to a bug it does not work correctly if
4767 SSL_read() or SSL_write() is called directly. In that scenario, if the
4768 handshake fails then a fatal error will be returned in the initial function
4769 call. If SSL_read()/SSL_write() is subsequently called by the application
4770 for the same SSL object then it will succeed and the data is passed without
4771 being decrypted/encrypted directly from the SSL/TLS record layer.
4773 In order to exploit this issue an application bug would have to be present
4774 that resulted in a call to SSL_read()/SSL_write() being issued after having
4775 already received a fatal error.
4777 This issue was reported to OpenSSL by David Benjamin (Google).
4782 * rsaz_1024_mul_avx2 overflow bug on x86_64
4784 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4785 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4786 Analysis suggests that attacks against RSA and DSA as a result of this
4787 defect would be very difficult to perform and are not believed likely.
4788 Attacks against DH1024 are considered just feasible, because most of the
4789 work necessary to deduce information about a private key may be performed
4790 offline. The amount of resources required for such an attack would be
4791 significant. However, for an attack on TLS to be meaningful, the server
4792 would have to share the DH1024 private key among multiple clients, which is
4793 no longer an option since CVE-2016-0701.
4795 This only affects processors that support the AVX2 but not ADX extensions
4796 like Intel Haswell (4th generation).
4798 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4799 was originally found via the OSS-Fuzz project.
4804 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4806 * bn_sqrx8x_internal carry bug on x86_64
4808 There is a carry propagating bug in the x86_64 Montgomery squaring
4809 procedure. No EC algorithms are affected. Analysis suggests that attacks
4810 against RSA and DSA as a result of this defect would be very difficult to
4811 perform and are not believed likely. Attacks against DH are considered just
4812 feasible (although very difficult) because most of the work necessary to
4813 deduce information about a private key may be performed offline. The amount
4814 of resources required for such an attack would be very significant and
4815 likely only accessible to a limited number of attackers. An attacker would
4816 additionally need online access to an unpatched system using the target
4817 private key in a scenario with persistent DH parameters and a private
4818 key that is shared between multiple clients.
4820 This only affects processors that support the BMI1, BMI2 and ADX extensions
4821 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4823 This issue was reported to OpenSSL by the OSS-Fuzz project.
4828 * Malformed X.509 IPAddressFamily could cause OOB read
4830 If an X.509 certificate has a malformed IPAddressFamily extension,
4831 OpenSSL could do a one-byte buffer overread. The most likely result
4832 would be an erroneous display of the certificate in text format.
4834 This issue was reported to OpenSSL by the OSS-Fuzz project.
4839 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4841 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4842 platform rather than 'mingw'.
4846 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4848 * Truncated packet could crash via OOB read
4850 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4851 cipher is being used, then a truncated packet can cause that host to
4852 perform an out-of-bounds read, usually resulting in a crash.
4854 This issue was reported to OpenSSL by Robert Święcki of Google.
4859 * BN_mod_exp may produce incorrect results on x86_64
4861 There is a carry propagating bug in the x86_64 Montgomery squaring
4862 procedure. No EC algorithms are affected. Analysis suggests that attacks
4863 against RSA and DSA as a result of this defect would be very difficult to
4864 perform and are not believed likely. Attacks against DH are considered just
4865 feasible (although very difficult) because most of the work necessary to
4866 deduce information about a private key may be performed offline. The amount
4867 of resources required for such an attack would be very significant and
4868 likely only accessible to a limited number of attackers. An attacker would
4869 additionally need online access to an unpatched system using the target
4870 private key in a scenario with persistent DH parameters and a private
4871 key that is shared between multiple clients. For example this can occur by
4872 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4873 similar to CVE-2015-3193 but must be treated as a separate problem.
4875 This issue was reported to OpenSSL by the OSS-Fuzz project.
4880 * Montgomery multiplication may produce incorrect results
4882 There is a carry propagating bug in the Broadwell-specific Montgomery
4883 multiplication procedure that handles input lengths divisible by, but
4884 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4885 and DH private keys are impossible. This is because the subroutine in
4886 question is not used in operations with the private key itself and an input
4887 of the attacker's direct choice. Otherwise the bug can manifest itself as
4888 transient authentication and key negotiation failures or reproducible
4889 erroneous outcome of public-key operations with specially crafted input.
4890 Among EC algorithms only Brainpool P-512 curves are affected and one
4891 presumably can attack ECDH key negotiation. Impact was not analyzed in
4892 detail, because pre-requisites for attack are considered unlikely. Namely
4893 multiple clients have to choose the curve in question and the server has to
4894 share the private key among them, neither of which is default behaviour.
4895 Even then only clients that chose the curve will be affected.
4897 This issue was publicly reported as transient failures and was not
4898 initially recognized as a security issue. Thanks to Richard Morgan for
4899 providing reproducible case.
4904 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4905 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4906 prevent issues where no progress is being made and the peer continually
4907 sends unrecognised record types, using up resources processing them.
4911 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4913 * Missing CRL sanity check
4915 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4916 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4917 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4919 This issue only affects the OpenSSL 1.0.2i
4924 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4926 * OCSP Status Request extension unbounded memory growth
4928 A malicious client can send an excessively large OCSP Status Request
4929 extension. If that client continually requests renegotiation, sending a
4930 large OCSP Status Request extension each time, then there will be unbounded
4931 memory growth on the server. This will eventually lead to a Denial Of
4932 Service attack through memory exhaustion. Servers with a default
4933 configuration are vulnerable even if they do not support OCSP. Builds using
4934 the "no-ocsp" build time option are not affected.
4936 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4941 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4944 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4950 * OOB write in MDC2_Update()
4952 An overflow can occur in MDC2_Update() either if called directly or
4953 through the EVP_DigestUpdate() function using MDC2. If an attacker
4954 is able to supply very large amounts of input data after a previous
4955 call to EVP_EncryptUpdate() with a partial block then a length check
4956 can overflow resulting in a heap corruption.
4958 The amount of data needed is comparable to SIZE_MAX which is impractical
4961 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4966 * Malformed SHA512 ticket DoS
4968 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4969 DoS attack where a malformed ticket will result in an OOB read which will
4972 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4973 a custom server callback and ticket lookup mechanism.
4975 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4980 * OOB write in BN_bn2dec()
4982 The function BN_bn2dec() does not check the return value of BN_div_word().
4983 This can cause an OOB write if an application uses this function with an
4984 overly large BIGNUM. This could be a problem if an overly large certificate
4985 or CRL is printed out from an untrusted source. TLS is not affected because
4986 record limits will reject an oversized certificate before it is parsed.
4988 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4993 * OOB read in TS_OBJ_print_bio()
4995 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4996 the total length the OID text representation would use and not the amount
4997 of data written. This will result in OOB reads when large OIDs are
5000 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5005 * Pointer arithmetic undefined behaviour
5007 Avoid some undefined pointer arithmetic
5009 A common idiom in the codebase is to check limits in the following manner:
5012 Where "p" points to some malloc'd data of SIZE bytes and
5015 "len" here could be from some externally supplied data (e.g. from a TLS
5018 The rules of C pointer arithmetic are such that "p + len" is only well
5019 defined where len <= SIZE. Therefore the above idiom is actually
5020 undefined behaviour.
5022 For example this could cause problems if some malloc implementation
5023 provides an address for "p" such that "p + len" actually overflows for
5024 values of len that are too big and therefore p + len < limit.
5026 This issue was reported to OpenSSL by Guido Vranken
5031 * Constant time flag not preserved in DSA signing
5033 Operations in the DSA signing algorithm should run in constant time in
5034 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5035 implementation means that a non-constant time codepath is followed for
5036 certain operations. This has been demonstrated through a cache-timing
5037 attack to be sufficient for an attacker to recover the private DSA key.
5039 This issue was reported by César Pereida (Aalto University), Billy Brumley
5040 (Tampere University of Technology), and Yuval Yarom (The University of
5041 Adelaide and NICTA).
5046 * DTLS buffered message DoS
5048 In a DTLS connection where handshake messages are delivered out-of-order
5049 those messages that OpenSSL is not yet ready to process will be buffered
5050 for later use. Under certain circumstances, a flaw in the logic means that
5051 those messages do not get removed from the buffer even though the handshake
5052 has been completed. An attacker could force up to approx. 15 messages to
5053 remain in the buffer when they are no longer required. These messages will
5054 be cleared when the DTLS connection is closed. The default maximum size for
5055 a message is 100k. Therefore the attacker could force an additional 1500k
5056 to be consumed per connection. By opening many simulataneous connections an
5057 attacker could cause a DoS attack through memory exhaustion.
5059 This issue was reported to OpenSSL by Quan Luo.
5064 * DTLS replay protection DoS
5066 A flaw in the DTLS replay attack protection mechanism means that records
5067 that arrive for future epochs update the replay protection "window" before
5068 the MAC for the record has been validated. This could be exploited by an
5069 attacker by sending a record for the next epoch (which does not have to
5070 decrypt or have a valid MAC), with a very large sequence number. This means
5071 that all subsequent legitimate packets are dropped causing a denial of
5072 service for a specific DTLS connection.
5074 This issue was reported to OpenSSL by the OCAP audit team.
5079 * Certificate message OOB reads
5081 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5082 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5083 theoretical DoS risk but this has not been observed in practice on common
5086 The messages affected are client certificate, client certificate request
5087 and server certificate. As a result the attack can only be performed
5088 against a client or a server which enables client authentication.
5090 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5095 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5097 * Prevent padding oracle in AES-NI CBC MAC check
5099 A MITM attacker can use a padding oracle attack to decrypt traffic
5100 when the connection uses an AES CBC cipher and the server support
5103 This issue was introduced as part of the fix for Lucky 13 padding
5104 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5105 constant time by making sure that always the same bytes are read and
5106 compared against either the MAC or padding bytes. But it no longer
5107 checked that there was enough data to have both the MAC and padding
5110 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5115 * Fix EVP_EncodeUpdate overflow
5117 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5118 Base64 encoding of binary data. If an attacker is able to supply very large
5119 amounts of input data then a length check can overflow resulting in a heap
5122 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5123 the `PEM_write_bio*` family of functions. These are mainly used within the
5124 OpenSSL command line applications, so any application which processes data
5125 from an untrusted source and outputs it as a PEM file should be considered
5126 vulnerable to this issue. User applications that call these APIs directly
5127 with large amounts of untrusted data may also be vulnerable.
5129 This issue was reported by Guido Vranken.
5134 * Fix EVP_EncryptUpdate overflow
5136 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5137 is able to supply very large amounts of input data after a previous call to
5138 EVP_EncryptUpdate() with a partial block then a length check can overflow
5139 resulting in a heap corruption. Following an analysis of all OpenSSL
5140 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5141 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5142 the first called function after an EVP_EncryptInit(), and therefore that
5143 specific call must be safe. The second form is where the length passed to
5144 EVP_EncryptUpdate() can be seen from the code to be some small value and
5145 therefore there is no possibility of an overflow. Since all instances are
5146 one of these two forms, it is believed that there can be no overflows in
5147 internal code due to this problem. It should be noted that
5148 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5149 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5150 of these calls have also been analysed too and it is believed there are no
5151 instances in internal usage where an overflow could occur.
5153 This issue was reported by Guido Vranken.
5158 * Prevent ASN.1 BIO excessive memory allocation
5160 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5161 a short invalid encoding can cause allocation of large amounts of memory
5162 potentially consuming excessive resources or exhausting memory.
5164 Any application parsing untrusted data through d2i BIO functions is
5165 affected. The memory based functions such as d2i_X509() are *not* affected.
5166 Since the memory based functions are used by the TLS library, TLS
5167 applications are not affected.
5169 This issue was reported by Brian Carpenter.
5176 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5177 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5178 in arbitrary stack data being returned in the buffer.
5180 This issue was reported by Guido Vranken.
5185 * Modify behavior of ALPN to invoke callback after SNI/servername
5186 callback, such that updates to the SSL_CTX affect ALPN.
5190 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5195 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5196 methods are enabled and ssl2 is disabled the methods return NULL.
5200 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5202 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5203 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5204 provide any "EXPORT" or "LOW" strength ciphers.
5208 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5209 is by default disabled at build-time. Builds that are not configured with
5210 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5211 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5212 will need to explicitly call either of:
5214 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5216 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5218 as appropriate. Even if either of those is used, or the application
5219 explicitly uses the version-specific SSLv2_method() or its client and
5220 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5221 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5222 ciphers, and SSLv2 56-bit DES are no longer available.
5227 * Fix a double-free in DSA code
5229 A double free bug was discovered when OpenSSL parses malformed DSA private
5230 keys and could lead to a DoS attack or memory corruption for applications
5231 that receive DSA private keys from untrusted sources. This scenario is
5234 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5240 * Disable SRP fake user seed to address a server memory leak.
5242 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5244 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5245 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5246 was changed to ignore the "fake user" SRP seed, even if the seed
5249 Users should use SRP_VBASE_get1_by_user instead. Note that in
5250 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5251 also that even though configuring the SRP seed attempts to hide
5252 invalid usernames by continuing the handshake with fake
5253 credentials, this behaviour is not constant time and no strong
5254 guarantees are made that the handshake is indistinguishable from
5255 that of a valid user.
5260 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5262 In the BN_hex2bn function the number of hex digits is calculated using an
5263 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5264 large values of `i` this can result in `bn_expand` not allocating any
5265 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5266 field as NULL leading to a subsequent NULL ptr deref. For very large values
5267 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5268 In this case memory is allocated to the internal BIGNUM data field, but it
5269 is insufficiently sized leading to heap corruption. A similar issue exists
5270 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5271 is ever called by user applications with very large untrusted hex/dec data.
5272 This is anticipated to be a rare occurrence.
5274 All OpenSSL internal usage of these functions use data that is not expected
5275 to be untrusted, e.g. config file data or application command line
5276 arguments. If user developed applications generate config file data based
5277 on untrusted data then it is possible that this could also lead to security
5278 consequences. This is also anticipated to be rare.
5280 This issue was reported to OpenSSL by Guido Vranken.
5285 * Fix memory issues in `BIO_*printf` functions
5287 The internal `fmtstr` function used in processing a "%s" format string in
5288 the `BIO_*printf` functions could overflow while calculating the length of a
5289 string and cause an OOB read when printing very long strings.
5291 Additionally the internal `doapr_outch` function can attempt to write to an
5292 OOB memory location (at an offset from the NULL pointer) in the event of a
5293 memory allocation failure. In 1.0.2 and below this could be caused where
5294 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5295 could be in processing a very long "%s" format string. Memory leaks can
5298 The first issue may mask the second issue dependent on compiler behaviour.
5299 These problems could enable attacks where large amounts of untrusted data
5300 is passed to the `BIO_*printf` functions. If applications use these functions
5301 in this way then they could be vulnerable. OpenSSL itself uses these
5302 functions when printing out human-readable dumps of ASN.1 data. Therefore
5303 applications that print this data could be vulnerable if the data is from
5304 untrusted sources. OpenSSL command line applications could also be
5305 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5306 as command line arguments.
5308 Libssl is not considered directly vulnerable. Additionally certificates etc
5309 received via remote connections via libssl are also unlikely to be able to
5310 trigger these issues because of message size limits enforced within libssl.
5312 This issue was reported to OpenSSL Guido Vranken.
5317 * Side channel attack on modular exponentiation
5319 A side-channel attack was found which makes use of cache-bank conflicts on
5320 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5321 of RSA keys. The ability to exploit this issue is limited as it relies on
5322 an attacker who has control of code in a thread running on the same
5323 hyper-threaded core as the victim thread which is performing decryptions.
5325 This issue was reported to OpenSSL by Yuval Yarom, The University of
5326 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5327 Nadia Heninger, University of Pennsylvania with more information at
5328 <http://cachebleed.info>.
5333 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5334 if no keysize is specified with default_bits. This fixes an
5335 omission in an earlier change that changed all RSA/DSA key generation
5336 commands to use 2048 bits by default.
5340 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5342 * DH small subgroups
5344 Historically OpenSSL only ever generated DH parameters based on "safe"
5345 primes. More recently (in version 1.0.2) support was provided for
5346 generating X9.42 style parameter files such as those required for RFC 5114
5347 support. The primes used in such files may not be "safe". Where an
5348 application is using DH configured with parameters based on primes that are
5349 not "safe" then an attacker could use this fact to find a peer's private
5350 DH exponent. This attack requires that the attacker complete multiple
5351 handshakes in which the peer uses the same private DH exponent. For example
5352 this could be used to discover a TLS server's private DH exponent if it's
5353 reusing the private DH exponent or it's using a static DH ciphersuite.
5355 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5356 TLS. It is not on by default. If the option is not set then the server
5357 reuses the same private DH exponent for the life of the server process and
5358 would be vulnerable to this attack. It is believed that many popular
5359 applications do set this option and would therefore not be at risk.
5361 The fix for this issue adds an additional check where a "q" parameter is
5362 available (as is the case in X9.42 based parameters). This detects the
5363 only known attack, and is the only possible defense for static DH
5364 ciphersuites. This could have some performance impact.
5366 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5367 default and cannot be disabled. This could have some performance impact.
5369 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5374 * SSLv2 doesn't block disabled ciphers
5376 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5377 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5378 been disabled, provided that the SSLv2 protocol was not also disabled via
5381 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5382 and Sebastian Schinzel.
5387 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5389 * BN_mod_exp may produce incorrect results on x86_64
5391 There is a carry propagating bug in the x86_64 Montgomery squaring
5392 procedure. No EC algorithms are affected. Analysis suggests that attacks
5393 against RSA and DSA as a result of this defect would be very difficult to
5394 perform and are not believed likely. Attacks against DH are considered just
5395 feasible (although very difficult) because most of the work necessary to
5396 deduce information about a private key may be performed offline. The amount
5397 of resources required for such an attack would be very significant and
5398 likely only accessible to a limited number of attackers. An attacker would
5399 additionally need online access to an unpatched system using the target
5400 private key in a scenario with persistent DH parameters and a private
5401 key that is shared between multiple clients. For example this can occur by
5402 default in OpenSSL DHE based SSL/TLS ciphersuites.
5404 This issue was reported to OpenSSL by Hanno Böck.
5409 * Certificate verify crash with missing PSS parameter
5411 The signature verification routines will crash with a NULL pointer
5412 dereference if presented with an ASN.1 signature using the RSA PSS
5413 algorithm and absent mask generation function parameter. Since these
5414 routines are used to verify certificate signature algorithms this can be
5415 used to crash any certificate verification operation and exploited in a
5416 DoS attack. Any application which performs certificate verification is
5417 vulnerable including OpenSSL clients and servers which enable client
5420 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5425 * X509_ATTRIBUTE memory leak
5427 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5428 memory. This structure is used by the PKCS#7 and CMS routines so any
5429 application which reads PKCS#7 or CMS data from untrusted sources is
5430 affected. SSL/TLS is not affected.
5432 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5438 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5439 This changes the decoding behaviour for some invalid messages,
5440 though the change is mostly in the more lenient direction, and
5441 legacy behaviour is preserved as much as possible.
5445 * In DSA_generate_parameters_ex, if the provided seed is too short,
5448 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5450 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5452 * Alternate chains certificate forgery
5454 During certificate verification, OpenSSL will attempt to find an
5455 alternative certificate chain if the first attempt to build such a chain
5456 fails. An error in the implementation of this logic can mean that an
5457 attacker could cause certain checks on untrusted certificates to be
5458 bypassed, such as the CA flag, enabling them to use a valid leaf
5459 certificate to act as a CA and "issue" an invalid certificate.
5461 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5466 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5468 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5469 incompatibility in the handling of HMAC. The previous ABI has now been
5474 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5476 * Malformed ECParameters causes infinite loop
5478 When processing an ECParameters structure OpenSSL enters an infinite loop
5479 if the curve specified is over a specially malformed binary polynomial
5482 This can be used to perform denial of service against any
5483 system which processes public keys, certificate requests or
5484 certificates. This includes TLS clients and TLS servers with
5485 client authentication enabled.
5487 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5492 * Exploitable out-of-bounds read in X509_cmp_time
5494 X509_cmp_time does not properly check the length of the ASN1_TIME
5495 string and can read a few bytes out of bounds. In addition,
5496 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5499 An attacker can use this to craft malformed certificates and CRLs of
5500 various sizes and potentially cause a segmentation fault, resulting in
5501 a DoS on applications that verify certificates or CRLs. TLS clients
5502 that verify CRLs are affected. TLS clients and servers with client
5503 authentication enabled may be affected if they use custom verification
5506 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5507 independently by Hanno Böck.
5512 * PKCS7 crash with missing EnvelopedContent
5514 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5515 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5516 with missing content and trigger a NULL pointer dereference on parsing.
5518 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5519 structures from untrusted sources are affected. OpenSSL clients and
5520 servers are not affected.
5522 This issue was reported to OpenSSL by Michal Zalewski (Google).
5527 * CMS verify infinite loop with unknown hash function
5529 When verifying a signedData message the CMS code can enter an infinite loop
5530 if presented with an unknown hash function OID. This can be used to perform
5531 denial of service against any system which verifies signedData messages using
5533 This issue was reported to OpenSSL by Johannes Bauer.
5538 * Race condition handling NewSessionTicket
5540 If a NewSessionTicket is received by a multi-threaded client when attempting to
5541 reuse a previous ticket then a race condition can occur potentially leading to
5542 a double free of the ticket data.
5547 * Only support 256-bit or stronger elliptic curves with the
5548 'ecdh_auto' setting (server) or by default (client). Of supported
5549 curves, prefer P-256 (both).
5553 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5555 * ClientHello sigalgs DoS fix
5557 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5558 invalid signature algorithms extension a NULL pointer dereference will
5559 occur. This can be exploited in a DoS attack against the server.
5561 This issue was was reported to OpenSSL by David Ramos of Stanford
5565 *Stephen Henson and Matt Caswell*
5567 * Multiblock corrupted pointer fix
5569 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5570 feature only applies on 64 bit x86 architecture platforms that support AES
5571 NI instructions. A defect in the implementation of "multiblock" can cause
5572 OpenSSL's internal write buffer to become incorrectly set to NULL when
5573 using non-blocking IO. Typically, when the user application is using a
5574 socket BIO for writing, this will only result in a failed connection.
5575 However if some other BIO is used then it is likely that a segmentation
5576 fault will be triggered, thus enabling a potential DoS attack.
5578 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5583 * Segmentation fault in DTLSv1_listen fix
5585 The DTLSv1_listen function is intended to be stateless and processes the
5586 initial ClientHello from many peers. It is common for user code to loop
5587 over the call to DTLSv1_listen until a valid ClientHello is received with
5588 an associated cookie. A defect in the implementation of DTLSv1_listen means
5589 that state is preserved in the SSL object from one invocation to the next
5590 that can lead to a segmentation fault. Errors processing the initial
5591 ClientHello can trigger this scenario. An example of such an error could be
5592 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5595 This issue was reported to OpenSSL by Per Allansson.
5600 * Segmentation fault in ASN1_TYPE_cmp fix
5602 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5603 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5604 certificate signature algorithm consistency this can be used to crash any
5605 certificate verification operation and exploited in a DoS attack. Any
5606 application which performs certificate verification is vulnerable including
5607 OpenSSL clients and servers which enable client authentication.
5612 * Segmentation fault for invalid PSS parameters fix
5614 The signature verification routines will crash with a NULL pointer
5615 dereference if presented with an ASN.1 signature using the RSA PSS
5616 algorithm and invalid parameters. Since these routines are used to verify
5617 certificate signature algorithms this can be used to crash any
5618 certificate verification operation and exploited in a DoS attack. Any
5619 application which performs certificate verification is vulnerable including
5620 OpenSSL clients and servers which enable client authentication.
5622 This issue was was reported to OpenSSL by Brian Carpenter.
5627 * ASN.1 structure reuse memory corruption fix
5629 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5630 memory corruption via an invalid write. Such reuse is and has been
5631 strongly discouraged and is believed to be rare.
5633 Applications that parse structures containing CHOICE or ANY DEFINED BY
5634 components may be affected. Certificate parsing (d2i_X509 and related
5635 functions) are however not affected. OpenSSL clients and servers are
5641 * PKCS7 NULL pointer dereferences fix
5643 The PKCS#7 parsing code does not handle missing outer ContentInfo
5644 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5645 missing content and trigger a NULL pointer dereference on parsing.
5647 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5648 otherwise parse PKCS#7 structures from untrusted sources are
5649 affected. OpenSSL clients and servers are not affected.
5651 This issue was reported to OpenSSL by Michal Zalewski (Google).
5656 * DoS via reachable assert in SSLv2 servers fix
5658 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5659 servers that both support SSLv2 and enable export cipher suites by sending
5660 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5662 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5663 (OpenSSL development team).
5668 * Empty CKE with client auth and DHE fix
5670 If client auth is used then a server can seg fault in the event of a DHE
5671 ciphersuite being selected and a zero length ClientKeyExchange message
5672 being sent by the client. This could be exploited in a DoS attack.
5677 * Handshake with unseeded PRNG fix
5679 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5680 with an unseeded PRNG. The conditions are:
5681 - The client is on a platform where the PRNG has not been seeded
5682 automatically, and the user has not seeded manually
5683 - A protocol specific client method version has been used (i.e. not
5684 SSL_client_methodv23)
5685 - A ciphersuite is used that does not require additional random data from
5686 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5688 If the handshake succeeds then the client random that has been used will
5689 have been generated from a PRNG with insufficient entropy and therefore the
5690 output may be predictable.
5692 For example using the following command with an unseeded openssl will
5693 succeed on an unpatched platform:
5695 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5700 * Use After Free following d2i_ECPrivatekey error fix
5702 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5703 could cause a use after free condition. This, in turn, could cause a double
5704 free in several private key parsing functions (such as d2i_PrivateKey
5705 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5706 for applications that receive EC private keys from untrusted
5707 sources. This scenario is considered rare.
5709 This issue was discovered by the BoringSSL project and fixed in their
5715 * X509_to_X509_REQ NULL pointer deref fix
5717 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5718 the certificate key is invalid. This function is rarely used in practice.
5720 This issue was discovered by Brian Carpenter.
5725 * Removed the export ciphers from the DEFAULT ciphers
5729 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5731 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5732 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5733 So far those who have to target multiple platforms would compromise
5734 and argue that binary targeting say ARMv5 would still execute on
5735 ARMv8. "Universal" build resolves this compromise by providing
5736 near-optimal performance even on newer platforms.
5740 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5741 (other platforms pending).
5743 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5745 * Add support for the SignedCertificateTimestampList certificate and
5746 OCSP response extensions from RFC6962.
5750 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5751 for corner cases. (Certain input points at infinity could lead to
5752 bogus results, with non-infinity inputs mapped to infinity too.)
5756 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5757 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5758 common cases are optimized and there still is room for further
5759 improvements. Vector Permutation AES for Altivec is also added.
5763 * Add support for little-endian ppc64 Linux target.
5765 *Marcelo Cerri (IBM)*
5767 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5768 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5769 are optimized and there still is room for further improvements.
5770 Both 32- and 64-bit modes are supported.
5772 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5774 * Improved ARMv7 NEON support.
5778 * Support for SPARC Architecture 2011 crypto extensions, first
5779 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5780 SHA256/512, MD5, GHASH and modular exponentiation.
5782 *Andy Polyakov, David Miller*
5784 * Accelerated modular exponentiation for Intel processors, a.k.a.
5787 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5789 * Support for new and upcoming Intel processors, including AVX2,
5790 BMI and SHA ISA extensions. This includes additional "stitched"
5791 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5794 This work was sponsored by Intel Corp.
5798 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5799 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5800 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5804 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5805 this fixes a limitation in previous versions of OpenSSL.
5809 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5810 MGF1 digest and OAEP label.
5814 * Add EVP support for key wrapping algorithms, to avoid problems with
5815 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5816 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5817 algorithms and include tests cases.
5821 * Add functions to allocate and set the fields of an ECDSA_METHOD
5824 *Douglas E. Engert, Steve Henson*
5826 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5827 difference in days and seconds between two tm or ASN1_TIME structures.
5831 * Add -rev test option to s_server to just reverse order of characters
5832 received by client and send back to server. Also prints an abbreviated
5833 summary of the connection parameters.
5837 * New option -brief for s_client and s_server to print out a brief summary
5838 of connection parameters.
5842 * Add callbacks for arbitrary TLS extensions.
5844 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5846 * New option -crl_download in several openssl utilities to download CRLs
5847 from CRLDP extension in certificates.
5851 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5855 * New function X509_CRL_diff to generate a delta CRL from the difference
5856 of two full CRLs. Add support to "crl" utility.
5860 * New functions to set lookup_crls function and to retrieve
5861 X509_STORE from X509_STORE_CTX.
5865 * Print out deprecated issuer and subject unique ID fields in
5870 * Extend OCSP I/O functions so they can be used for simple general purpose
5871 HTTP as well as OCSP. New wrapper function which can be used to download
5872 CRLs using the OCSP API.
5876 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5880 * `SSL_CONF*` functions. These provide a common framework for application
5881 configuration using configuration files or command lines.
5885 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5886 message callback and prints the results. Needs compile time option
5887 "enable-ssl-trace". New options to s_client and s_server to enable
5892 * New ctrl and macro to retrieve supported points extensions.
5893 Print out extension in s_server and s_client.
5897 * New functions to retrieve certificate signature and signature
5902 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5907 * New Suite B modes for TLS code. These use and enforce the requirements
5908 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5909 only use Suite B curves. The Suite B modes can be set by using the
5910 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5914 * New chain verification flags for Suite B levels of security. Check
5915 algorithms are acceptable when flags are set in X509_verify_cert.
5919 * Make tls1_check_chain return a set of flags indicating checks passed
5920 by a certificate chain. Add additional tests to handle client
5921 certificates: checks for matching certificate type and issuer name
5926 * If an attempt is made to use a signature algorithm not in the peer
5927 preference list abort the handshake. If client has no suitable
5928 signature algorithms in response to a certificate request do not
5929 use the certificate.
5933 * If server EC tmp key is not in client preference list abort handshake.
5937 * Add support for certificate stores in CERT structure. This makes it
5938 possible to have different stores per SSL structure or one store in
5939 the parent SSL_CTX. Include distinct stores for certificate chain
5940 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5941 to build and store a certificate chain in CERT structure: returning
5942 an error if the chain cannot be built: this will allow applications
5943 to test if a chain is correctly configured.
5945 Note: if the CERT based stores are not set then the parent SSL_CTX
5946 store is used to retain compatibility with existing behaviour.
5950 * New function ssl_set_client_disabled to set a ciphersuite disabled
5951 mask based on the current session, check mask when sending client
5952 hello and checking the requested ciphersuite.
5956 * New ctrls to retrieve and set certificate types in a certificate
5957 request message. Print out received values in s_client. If certificate
5958 types is not set with custom values set sensible values based on
5959 supported signature algorithms.
5963 * Support for distinct client and server supported signature algorithms.
5967 * Add certificate callback. If set this is called whenever a certificate
5968 is required by client or server. An application can decide which
5969 certificate chain to present based on arbitrary criteria: for example
5970 supported signature algorithms. Add very simple example to s_server.
5971 This fixes many of the problems and restrictions of the existing client
5972 certificate callback: for example you can now clear an existing
5973 certificate and specify the whole chain.
5977 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5978 the certificate can be used for (if anything). Set valid_flags field
5979 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5980 to have similar checks in it.
5982 Add new "cert_flags" field to CERT structure and include a "strict mode".
5983 This enforces some TLS certificate requirements (such as only permitting
5984 certificate signature algorithms contained in the supported algorithms
5985 extension) which some implementations ignore: this option should be used
5986 with caution as it could cause interoperability issues.
5990 * Update and tidy signature algorithm extension processing. Work out
5991 shared signature algorithms based on preferences and peer algorithms
5992 and print them out in s_client and s_server. Abort handshake if no
5993 shared signature algorithms.
5997 * Add new functions to allow customised supported signature algorithms
5998 for SSL and SSL_CTX structures. Add options to s_client and s_server
6003 * New function SSL_certs_clear() to delete all references to certificates
6004 from an SSL structure. Before this once a certificate had been added
6005 it couldn't be removed.
6009 * Integrate hostname, email address and IP address checking with certificate
6010 verification. New verify options supporting checking in openssl utility.
6014 * Fixes and wildcard matching support to hostname and email checking
6015 functions. Add manual page.
6017 *Florian Weimer (Red Hat Product Security Team)*
6019 * New functions to check a hostname email or IP address against a
6020 certificate. Add options x509 utility to print results of checks against
6025 * Fix OCSP checking.
6027 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6029 * Initial experimental support for explicitly trusted non-root CAs.
6030 OpenSSL still tries to build a complete chain to a root but if an
6031 intermediate CA has a trust setting included that is used. The first
6032 setting is used: whether to trust (e.g., -addtrust option to the x509
6037 * Add -trusted_first option which attempts to find certificates in the
6038 trusted store even if an untrusted chain is also supplied.
6042 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6043 platform support for Linux and Android.
6047 * Support for linux-x32, ILP32 environment in x86_64 framework.
6051 * Experimental multi-implementation support for FIPS capable OpenSSL.
6052 When in FIPS mode the approved implementations are used as normal,
6053 when not in FIPS mode the internal unapproved versions are used instead.
6054 This means that the FIPS capable OpenSSL isn't forced to use the
6055 (often lower performance) FIPS implementations outside FIPS mode.
6059 * Transparently support X9.42 DH parameters when calling
6060 PEM_read_bio_DHparameters. This means existing applications can handle
6061 the new parameter format automatically.
6065 * Initial experimental support for X9.42 DH parameter format: mainly
6066 to support use of 'q' parameter for RFC5114 parameters.
6070 * Add DH parameters from RFC5114 including test data to dhtest.
6074 * Support for automatic EC temporary key parameter selection. If enabled
6075 the most preferred EC parameters are automatically used instead of
6076 hardcoded fixed parameters. Now a server just has to call:
6077 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6078 support ECDH and use the most appropriate parameters.
6082 * Enhance and tidy EC curve and point format TLS extension code. Use
6083 static structures instead of allocation if default values are used.
6084 New ctrls to set curves we wish to support and to retrieve shared curves.
6085 Print out shared curves in s_server. New options to s_server and s_client
6086 to set list of supported curves.
6090 * New ctrls to retrieve supported signature algorithms and
6091 supported curve values as an array of NIDs. Extend openssl utility
6092 to print out received values.
6096 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6097 between NIDs and the more common NIST names such as "P-256". Enhance
6098 ecparam utility and ECC method to recognise the NIST names for curves.
6102 * Enhance SSL/TLS certificate chain handling to support different
6103 chains for each certificate instead of one chain in the parent SSL_CTX.
6107 * Support for fixed DH ciphersuite client authentication: where both
6108 server and client use DH certificates with common parameters.
6112 * Support for fixed DH ciphersuites: those requiring DH server
6117 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6119 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6120 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6121 X509_CINF_get_signature were reverted post internal team review.
6126 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6128 * OCSP Status Request extension unbounded memory growth
6130 A malicious client can send an excessively large OCSP Status Request
6131 extension. If that client continually requests renegotiation, sending a
6132 large OCSP Status Request extension each time, then there will be unbounded
6133 memory growth on the server. This will eventually lead to a Denial Of
6134 Service attack through memory exhaustion. Servers with a default
6135 configuration are vulnerable even if they do not support OCSP. Builds using
6136 the "no-ocsp" build time option are not affected.
6138 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6143 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6146 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6152 * OOB write in MDC2_Update()
6154 An overflow can occur in MDC2_Update() either if called directly or
6155 through the EVP_DigestUpdate() function using MDC2. If an attacker
6156 is able to supply very large amounts of input data after a previous
6157 call to EVP_EncryptUpdate() with a partial block then a length check
6158 can overflow resulting in a heap corruption.
6160 The amount of data needed is comparable to SIZE_MAX which is impractical
6163 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6168 * Malformed SHA512 ticket DoS
6170 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6171 DoS attack where a malformed ticket will result in an OOB read which will
6174 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6175 a custom server callback and ticket lookup mechanism.
6177 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6182 * OOB write in BN_bn2dec()
6184 The function BN_bn2dec() does not check the return value of BN_div_word().
6185 This can cause an OOB write if an application uses this function with an
6186 overly large BIGNUM. This could be a problem if an overly large certificate
6187 or CRL is printed out from an untrusted source. TLS is not affected because
6188 record limits will reject an oversized certificate before it is parsed.
6190 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6195 * OOB read in TS_OBJ_print_bio()
6197 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6198 the total length the OID text representation would use and not the amount
6199 of data written. This will result in OOB reads when large OIDs are
6202 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6207 * Pointer arithmetic undefined behaviour
6209 Avoid some undefined pointer arithmetic
6211 A common idiom in the codebase is to check limits in the following manner:
6214 Where "p" points to some malloc'd data of SIZE bytes and
6217 "len" here could be from some externally supplied data (e.g. from a TLS
6220 The rules of C pointer arithmetic are such that "p + len" is only well
6221 defined where len <= SIZE. Therefore the above idiom is actually
6222 undefined behaviour.
6224 For example this could cause problems if some malloc implementation
6225 provides an address for "p" such that "p + len" actually overflows for
6226 values of len that are too big and therefore p + len < limit.
6228 This issue was reported to OpenSSL by Guido Vranken
6233 * Constant time flag not preserved in DSA signing
6235 Operations in the DSA signing algorithm should run in constant time in
6236 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6237 implementation means that a non-constant time codepath is followed for
6238 certain operations. This has been demonstrated through a cache-timing
6239 attack to be sufficient for an attacker to recover the private DSA key.
6241 This issue was reported by César Pereida (Aalto University), Billy Brumley
6242 (Tampere University of Technology), and Yuval Yarom (The University of
6243 Adelaide and NICTA).
6248 * DTLS buffered message DoS
6250 In a DTLS connection where handshake messages are delivered out-of-order
6251 those messages that OpenSSL is not yet ready to process will be buffered
6252 for later use. Under certain circumstances, a flaw in the logic means that
6253 those messages do not get removed from the buffer even though the handshake
6254 has been completed. An attacker could force up to approx. 15 messages to
6255 remain in the buffer when they are no longer required. These messages will
6256 be cleared when the DTLS connection is closed. The default maximum size for
6257 a message is 100k. Therefore the attacker could force an additional 1500k
6258 to be consumed per connection. By opening many simulataneous connections an
6259 attacker could cause a DoS attack through memory exhaustion.
6261 This issue was reported to OpenSSL by Quan Luo.
6266 * DTLS replay protection DoS
6268 A flaw in the DTLS replay attack protection mechanism means that records
6269 that arrive for future epochs update the replay protection "window" before
6270 the MAC for the record has been validated. This could be exploited by an
6271 attacker by sending a record for the next epoch (which does not have to
6272 decrypt or have a valid MAC), with a very large sequence number. This means
6273 that all subsequent legitimate packets are dropped causing a denial of
6274 service for a specific DTLS connection.
6276 This issue was reported to OpenSSL by the OCAP audit team.
6281 * Certificate message OOB reads
6283 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6284 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6285 theoretical DoS risk but this has not been observed in practice on common
6288 The messages affected are client certificate, client certificate request
6289 and server certificate. As a result the attack can only be performed
6290 against a client or a server which enables client authentication.
6292 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6297 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6299 * Prevent padding oracle in AES-NI CBC MAC check
6301 A MITM attacker can use a padding oracle attack to decrypt traffic
6302 when the connection uses an AES CBC cipher and the server support
6305 This issue was introduced as part of the fix for Lucky 13 padding
6306 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6307 constant time by making sure that always the same bytes are read and
6308 compared against either the MAC or padding bytes. But it no longer
6309 checked that there was enough data to have both the MAC and padding
6312 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6317 * Fix EVP_EncodeUpdate overflow
6319 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6320 Base64 encoding of binary data. If an attacker is able to supply very large
6321 amounts of input data then a length check can overflow resulting in a heap
6324 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6325 the `PEM_write_bio*` family of functions. These are mainly used within the
6326 OpenSSL command line applications, so any application which processes data
6327 from an untrusted source and outputs it as a PEM file should be considered
6328 vulnerable to this issue. User applications that call these APIs directly
6329 with large amounts of untrusted data may also be vulnerable.
6331 This issue was reported by Guido Vranken.
6336 * Fix EVP_EncryptUpdate overflow
6338 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6339 is able to supply very large amounts of input data after a previous call to
6340 EVP_EncryptUpdate() with a partial block then a length check can overflow
6341 resulting in a heap corruption. Following an analysis of all OpenSSL
6342 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6343 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6344 the first called function after an EVP_EncryptInit(), and therefore that
6345 specific call must be safe. The second form is where the length passed to
6346 EVP_EncryptUpdate() can be seen from the code to be some small value and
6347 therefore there is no possibility of an overflow. Since all instances are
6348 one of these two forms, it is believed that there can be no overflows in
6349 internal code due to this problem. It should be noted that
6350 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6351 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6352 of these calls have also been analysed too and it is believed there are no
6353 instances in internal usage where an overflow could occur.
6355 This issue was reported by Guido Vranken.
6360 * Prevent ASN.1 BIO excessive memory allocation
6362 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6363 a short invalid encoding can casuse allocation of large amounts of memory
6364 potentially consuming excessive resources or exhausting memory.
6366 Any application parsing untrusted data through d2i BIO functions is
6367 affected. The memory based functions such as d2i_X509() are *not* affected.
6368 Since the memory based functions are used by the TLS library, TLS
6369 applications are not affected.
6371 This issue was reported by Brian Carpenter.
6378 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6379 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6380 in arbitrary stack data being returned in the buffer.
6382 This issue was reported by Guido Vranken.
6387 * Modify behavior of ALPN to invoke callback after SNI/servername
6388 callback, such that updates to the SSL_CTX affect ALPN.
6392 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6397 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6398 methods are enabled and ssl2 is disabled the methods return NULL.
6402 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6404 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6405 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6406 provide any "EXPORT" or "LOW" strength ciphers.
6410 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6411 is by default disabled at build-time. Builds that are not configured with
6412 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6413 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6414 will need to explicitly call either of:
6416 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6418 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6420 as appropriate. Even if either of those is used, or the application
6421 explicitly uses the version-specific SSLv2_method() or its client and
6422 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6423 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6424 ciphers, and SSLv2 56-bit DES are no longer available.
6429 * Fix a double-free in DSA code
6431 A double free bug was discovered when OpenSSL parses malformed DSA private
6432 keys and could lead to a DoS attack or memory corruption for applications
6433 that receive DSA private keys from untrusted sources. This scenario is
6436 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6442 * Disable SRP fake user seed to address a server memory leak.
6444 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6446 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6447 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6448 was changed to ignore the "fake user" SRP seed, even if the seed
6451 Users should use SRP_VBASE_get1_by_user instead. Note that in
6452 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6453 also that even though configuring the SRP seed attempts to hide
6454 invalid usernames by continuing the handshake with fake
6455 credentials, this behaviour is not constant time and no strong
6456 guarantees are made that the handshake is indistinguishable from
6457 that of a valid user.
6462 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6464 In the BN_hex2bn function the number of hex digits is calculated using an
6465 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6466 large values of `i` this can result in `bn_expand` not allocating any
6467 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6468 field as NULL leading to a subsequent NULL ptr deref. For very large values
6469 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6470 In this case memory is allocated to the internal BIGNUM data field, but it
6471 is insufficiently sized leading to heap corruption. A similar issue exists
6472 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6473 is ever called by user applications with very large untrusted hex/dec data.
6474 This is anticipated to be a rare occurrence.
6476 All OpenSSL internal usage of these functions use data that is not expected
6477 to be untrusted, e.g. config file data or application command line
6478 arguments. If user developed applications generate config file data based
6479 on untrusted data then it is possible that this could also lead to security
6480 consequences. This is also anticipated to be rare.
6482 This issue was reported to OpenSSL by Guido Vranken.
6487 * Fix memory issues in `BIO_*printf` functions
6489 The internal `fmtstr` function used in processing a "%s" format string in
6490 the `BIO_*printf` functions could overflow while calculating the length of a
6491 string and cause an OOB read when printing very long strings.
6493 Additionally the internal `doapr_outch` function can attempt to write to an
6494 OOB memory location (at an offset from the NULL pointer) in the event of a
6495 memory allocation failure. In 1.0.2 and below this could be caused where
6496 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6497 could be in processing a very long "%s" format string. Memory leaks can
6500 The first issue may mask the second issue dependent on compiler behaviour.
6501 These problems could enable attacks where large amounts of untrusted data
6502 is passed to the `BIO_*printf` functions. If applications use these functions
6503 in this way then they could be vulnerable. OpenSSL itself uses these
6504 functions when printing out human-readable dumps of ASN.1 data. Therefore
6505 applications that print this data could be vulnerable if the data is from
6506 untrusted sources. OpenSSL command line applications could also be
6507 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6508 as command line arguments.
6510 Libssl is not considered directly vulnerable. Additionally certificates etc
6511 received via remote connections via libssl are also unlikely to be able to
6512 trigger these issues because of message size limits enforced within libssl.
6514 This issue was reported to OpenSSL Guido Vranken.
6519 * Side channel attack on modular exponentiation
6521 A side-channel attack was found which makes use of cache-bank conflicts on
6522 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6523 of RSA keys. The ability to exploit this issue is limited as it relies on
6524 an attacker who has control of code in a thread running on the same
6525 hyper-threaded core as the victim thread which is performing decryptions.
6527 This issue was reported to OpenSSL by Yuval Yarom, The University of
6528 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6529 Nadia Heninger, University of Pennsylvania with more information at
6530 <http://cachebleed.info>.
6535 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6536 if no keysize is specified with default_bits. This fixes an
6537 omission in an earlier change that changed all RSA/DSA key generation
6538 commands to use 2048 bits by default.
6542 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6544 * Protection for DH small subgroup attacks
6546 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6547 switched on by default and cannot be disabled. This could have some
6552 * SSLv2 doesn't block disabled ciphers
6554 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6555 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6556 been disabled, provided that the SSLv2 protocol was not also disabled via
6559 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6560 and Sebastian Schinzel.
6565 * Reject DH handshakes with parameters shorter than 1024 bits.
6569 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6571 * Certificate verify crash with missing PSS parameter
6573 The signature verification routines will crash with a NULL pointer
6574 dereference if presented with an ASN.1 signature using the RSA PSS
6575 algorithm and absent mask generation function parameter. Since these
6576 routines are used to verify certificate signature algorithms this can be
6577 used to crash any certificate verification operation and exploited in a
6578 DoS attack. Any application which performs certificate verification is
6579 vulnerable including OpenSSL clients and servers which enable client
6582 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6587 * X509_ATTRIBUTE memory leak
6589 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6590 memory. This structure is used by the PKCS#7 and CMS routines so any
6591 application which reads PKCS#7 or CMS data from untrusted sources is
6592 affected. SSL/TLS is not affected.
6594 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6600 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6601 This changes the decoding behaviour for some invalid messages,
6602 though the change is mostly in the more lenient direction, and
6603 legacy behaviour is preserved as much as possible.
6607 * In DSA_generate_parameters_ex, if the provided seed is too short,
6608 use a random seed, as already documented.
6610 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6612 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6614 * Alternate chains certificate forgery
6616 During certificate verfification, OpenSSL will attempt to find an
6617 alternative certificate chain if the first attempt to build such a chain
6618 fails. An error in the implementation of this logic can mean that an
6619 attacker could cause certain checks on untrusted certificates to be
6620 bypassed, such as the CA flag, enabling them to use a valid leaf
6621 certificate to act as a CA and "issue" an invalid certificate.
6623 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6629 * Race condition handling PSK identify hint
6631 If PSK identity hints are received by a multi-threaded client then
6632 the values are wrongly updated in the parent SSL_CTX structure. This can
6633 result in a race condition potentially leading to a double free of the
6639 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6641 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6642 incompatibility in the handling of HMAC. The previous ABI has now been
6645 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6647 * Malformed ECParameters causes infinite loop
6649 When processing an ECParameters structure OpenSSL enters an infinite loop
6650 if the curve specified is over a specially malformed binary polynomial
6653 This can be used to perform denial of service against any
6654 system which processes public keys, certificate requests or
6655 certificates. This includes TLS clients and TLS servers with
6656 client authentication enabled.
6658 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6663 * Exploitable out-of-bounds read in X509_cmp_time
6665 X509_cmp_time does not properly check the length of the ASN1_TIME
6666 string and can read a few bytes out of bounds. In addition,
6667 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6670 An attacker can use this to craft malformed certificates and CRLs of
6671 various sizes and potentially cause a segmentation fault, resulting in
6672 a DoS on applications that verify certificates or CRLs. TLS clients
6673 that verify CRLs are affected. TLS clients and servers with client
6674 authentication enabled may be affected if they use custom verification
6677 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6678 independently by Hanno Böck.
6683 * PKCS7 crash with missing EnvelopedContent
6685 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6686 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6687 with missing content and trigger a NULL pointer dereference on parsing.
6689 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6690 structures from untrusted sources are affected. OpenSSL clients and
6691 servers are not affected.
6693 This issue was reported to OpenSSL by Michal Zalewski (Google).
6698 * CMS verify infinite loop with unknown hash function
6700 When verifying a signedData message the CMS code can enter an infinite loop
6701 if presented with an unknown hash function OID. This can be used to perform
6702 denial of service against any system which verifies signedData messages using
6704 This issue was reported to OpenSSL by Johannes Bauer.
6709 * Race condition handling NewSessionTicket
6711 If a NewSessionTicket is received by a multi-threaded client when attempting to
6712 reuse a previous ticket then a race condition can occur potentially leading to
6713 a double free of the ticket data.
6718 * Reject DH handshakes with parameters shorter than 768 bits.
6720 *Kurt Roeckx and Emilia Kasper*
6722 * dhparam: generate 2048-bit parameters by default.
6724 *Kurt Roeckx and Emilia Kasper*
6726 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6728 * Segmentation fault in ASN1_TYPE_cmp fix
6730 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6731 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6732 certificate signature algorithm consistency this can be used to crash any
6733 certificate verification operation and exploited in a DoS attack. Any
6734 application which performs certificate verification is vulnerable including
6735 OpenSSL clients and servers which enable client authentication.
6740 * ASN.1 structure reuse memory corruption fix
6742 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6743 memory corruption via an invalid write. Such reuse is and has been
6744 strongly discouraged and is believed to be rare.
6746 Applications that parse structures containing CHOICE or ANY DEFINED BY
6747 components may be affected. Certificate parsing (d2i_X509 and related
6748 functions) are however not affected. OpenSSL clients and servers are
6754 * PKCS7 NULL pointer dereferences fix
6756 The PKCS#7 parsing code does not handle missing outer ContentInfo
6757 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6758 missing content and trigger a NULL pointer dereference on parsing.
6760 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6761 otherwise parse PKCS#7 structures from untrusted sources are
6762 affected. OpenSSL clients and servers are not affected.
6764 This issue was reported to OpenSSL by Michal Zalewski (Google).
6769 * DoS via reachable assert in SSLv2 servers fix
6771 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6772 servers that both support SSLv2 and enable export cipher suites by sending
6773 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6775 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6776 (OpenSSL development team).
6781 * Use After Free following d2i_ECPrivatekey error fix
6783 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6784 could cause a use after free condition. This, in turn, could cause a double
6785 free in several private key parsing functions (such as d2i_PrivateKey
6786 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6787 for applications that receive EC private keys from untrusted
6788 sources. This scenario is considered rare.
6790 This issue was discovered by the BoringSSL project and fixed in their
6796 * X509_to_X509_REQ NULL pointer deref fix
6798 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6799 the certificate key is invalid. This function is rarely used in practice.
6801 This issue was discovered by Brian Carpenter.
6806 * Removed the export ciphers from the DEFAULT ciphers
6810 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6812 * Build fixes for the Windows and OpenVMS platforms
6814 *Matt Caswell and Richard Levitte*
6816 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6818 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6819 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6820 dereference. This could lead to a Denial Of Service attack. Thanks to
6821 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6826 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6827 dtls1_buffer_record function under certain conditions. In particular this
6828 could occur if an attacker sent repeated DTLS records with the same
6829 sequence number but for the next epoch. The memory leak could be exploited
6830 by an attacker in a Denial of Service attack through memory exhaustion.
6831 Thanks to Chris Mueller for reporting this issue.
6836 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6837 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6838 method would be set to NULL which could later result in a NULL pointer
6839 dereference. Thanks to Frank Schmirler for reporting this issue.
6844 * Abort handshake if server key exchange message is omitted for ephemeral
6847 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6848 reporting this issue.
6853 * Remove non-export ephemeral RSA code on client and server. This code
6854 violated the TLS standard by allowing the use of temporary RSA keys in
6855 non-export ciphersuites and could be used by a server to effectively
6856 downgrade the RSA key length used to a value smaller than the server
6857 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6858 INRIA or reporting this issue.
6863 * Fixed issue where DH client certificates are accepted without verification.
6864 An OpenSSL server will accept a DH certificate for client authentication
6865 without the certificate verify message. This effectively allows a client to
6866 authenticate without the use of a private key. This only affects servers
6867 which trust a client certificate authority which issues certificates
6868 containing DH keys: these are extremely rare and hardly ever encountered.
6869 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6875 * Ensure that the session ID context of an SSL is updated when its
6876 SSL_CTX is updated via SSL_set_SSL_CTX.
6878 The session ID context is typically set from the parent SSL_CTX,
6879 and can vary with the CTX.
6883 * Fix various certificate fingerprint issues.
6885 By using non-DER or invalid encodings outside the signed portion of a
6886 certificate the fingerprint can be changed without breaking the signature.
6887 Although no details of the signed portion of the certificate can be changed
6888 this can cause problems with some applications: e.g. those using the
6889 certificate fingerprint for blacklists.
6891 1. Reject signatures with non zero unused bits.
6893 If the BIT STRING containing the signature has non zero unused bits reject
6894 the signature. All current signature algorithms require zero unused bits.
6896 2. Check certificate algorithm consistency.
6898 Check the AlgorithmIdentifier inside TBS matches the one in the
6899 certificate signature. NB: this will result in signature failure
6900 errors for some broken certificates.
6902 Thanks to Konrad Kraszewski from Google for reporting this issue.
6904 3. Check DSA/ECDSA signatures use DER.
6906 Re-encode DSA/ECDSA signatures and compare with the original received
6907 signature. Return an error if there is a mismatch.
6909 This will reject various cases including garbage after signature
6910 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6911 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6912 (negative or with leading zeroes).
6914 Further analysis was conducted and fixes were developed by Stephen Henson
6915 of the OpenSSL core team.
6921 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6922 results on some platforms, including x86_64. This bug occurs at random
6923 with a very low probability, and is not known to be exploitable in any
6924 way, though its exact impact is difficult to determine. Thanks to Pieter
6925 Wuille (Blockstream) who reported this issue and also suggested an initial
6926 fix. Further analysis was conducted by the OpenSSL development team and
6927 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6928 the OpenSSL core team.
6933 * Do not resume sessions on the server if the negotiated protocol
6934 version does not match the session's version. Resuming with a different
6935 version, while not strictly forbidden by the RFC, is of questionable
6936 sanity and breaks all known clients.
6938 *David Benjamin, Emilia Käsper*
6940 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6941 early CCS messages during renegotiation. (Note that because
6942 renegotiation is encrypted, this early CCS was not exploitable.)
6946 * Tighten client-side session ticket handling during renegotiation:
6947 ensure that the client only accepts a session ticket if the server sends
6948 the extension anew in the ServerHello. Previously, a TLS client would
6949 reuse the old extension state and thus accept a session ticket if one was
6950 announced in the initial ServerHello.
6952 Similarly, ensure that the client requires a session ticket if one
6953 was advertised in the ServerHello. Previously, a TLS client would
6954 ignore a missing NewSessionTicket message.
6958 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6962 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6963 sends a carefully crafted handshake message, to cause OpenSSL to fail
6964 to free up to 64k of memory causing a memory leak. This could be
6965 exploited in a Denial Of Service attack. This issue affects OpenSSL
6966 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6967 whether SRTP is used or configured. Implementations of OpenSSL that
6968 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6970 The fix was developed by the OpenSSL team.
6975 * Session Ticket Memory Leak.
6977 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6978 integrity of that ticket is first verified. In the event of a session
6979 ticket integrity check failing, OpenSSL will fail to free memory
6980 causing a memory leak. By sending a large number of invalid session
6981 tickets an attacker could exploit this issue in a Denial Of Service
6987 * Build option no-ssl3 is incomplete.
6989 When OpenSSL is configured with "no-ssl3" as a build option, servers
6990 could accept and complete a SSL 3.0 handshake, and clients could be
6991 configured to send them.
6994 *Akamai and the OpenSSL team*
6996 * Add support for TLS_FALLBACK_SCSV.
6997 Client applications doing fallback retries should call
6998 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7001 *Adam Langley, Bodo Moeller*
7003 * Add additional DigestInfo checks.
7005 Re-encode DigestInto in DER and check against the original when
7006 verifying RSA signature: this will reject any improperly encoded
7007 DigestInfo structures.
7009 Note: this is a precautionary measure and no attacks are currently known.
7013 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7015 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7016 SRP code can be overrun an internal buffer. Add sanity check that
7017 g, A, B < N to SRP code.
7019 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7020 Group for discovering this issue.
7025 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7026 TLS 1.0 instead of higher protocol versions when the ClientHello message
7027 is badly fragmented. This allows a man-in-the-middle attacker to force a
7028 downgrade to TLS 1.0 even if both the server and the client support a
7029 higher protocol version, by modifying the client's TLS records.
7031 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7032 researching this issue.
7037 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7038 to a denial of service attack. A malicious server can crash the client
7039 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7040 ciphersuite and sending carefully crafted handshake messages.
7042 Thanks to Felix Gröbert (Google) for discovering and researching this
7048 * By sending carefully crafted DTLS packets an attacker could cause openssl
7049 to leak memory. This can be exploited through a Denial of Service attack.
7050 Thanks to Adam Langley for discovering and researching this issue.
7055 * An attacker can force openssl to consume large amounts of memory whilst
7056 processing DTLS handshake messages. This can be exploited through a
7057 Denial of Service attack.
7058 Thanks to Adam Langley for discovering and researching this issue.
7063 * An attacker can force an error condition which causes openssl to crash
7064 whilst processing DTLS packets due to memory being freed twice. This
7065 can be exploited through a Denial of Service attack.
7066 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7072 * If a multithreaded client connects to a malicious server using a resumed
7073 session and the server sends an ec point format extension it could write
7074 up to 255 bytes to freed memory.
7076 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7082 * A malicious server can crash an OpenSSL client with a null pointer
7083 dereference (read) by specifying an SRP ciphersuite even though it was not
7084 properly negotiated with the client. This can be exploited through a
7085 Denial of Service attack.
7087 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7088 discovering and researching this issue.
7093 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7094 X509_name_oneline, X509_name_print_ex et al. to leak some information
7095 from the stack. Applications may be affected if they echo pretty printing
7096 output to the attacker.
7098 Thanks to Ivan Fratric (Google) for discovering this issue.
7101 *Emilia Käsper, and Steve Henson*
7103 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7104 for corner cases. (Certain input points at infinity could lead to
7105 bogus results, with non-infinity inputs mapped to infinity too.)
7109 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7111 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7112 handshake can force the use of weak keying material in OpenSSL
7113 SSL/TLS clients and servers.
7115 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7116 researching this issue. ([CVE-2014-0224])
7118 *KIKUCHI Masashi, Steve Henson*
7120 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7121 OpenSSL DTLS client the code can be made to recurse eventually crashing
7124 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7127 *Imre Rad, Steve Henson*
7129 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7130 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7131 client or server. This is potentially exploitable to run arbitrary
7132 code on a vulnerable client or server.
7134 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7136 *Jüri Aedla, Steve Henson*
7138 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7139 are subject to a denial of service attack.
7141 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7142 this issue. ([CVE-2014-3470])
7144 *Felix Gröbert, Ivan Fratric, Steve Henson*
7146 * Harmonize version and its documentation. -f flag is used to display
7149 *mancha <mancha1@zoho.com>*
7151 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7152 in i2d_ECPrivateKey.
7154 *mancha <mancha1@zoho.com>*
7156 * Fix some double frees. These are not thought to be exploitable.
7158 *mancha <mancha1@zoho.com>*
7160 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7162 * A missing bounds check in the handling of the TLS heartbeat extension
7163 can be used to reveal up to 64k of memory to a connected client or
7166 Thanks for Neel Mehta of Google Security for discovering this bug and to
7167 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7168 preparing the fix ([CVE-2014-0160])
7170 *Adam Langley, Bodo Moeller*
7172 * Fix for the attack described in the paper "Recovering OpenSSL
7173 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7174 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7175 <http://eprint.iacr.org/2014/140>
7177 Thanks to Yuval Yarom and Naomi Benger for discovering this
7178 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7180 *Yuval Yarom and Naomi Benger*
7182 * TLS pad extension: draft-agl-tls-padding-03
7184 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7185 TLS client Hello record length value would otherwise be > 255 and
7186 less that 512 pad with a dummy extension containing zeroes so it
7187 is at least 512 bytes long.
7189 *Adam Langley, Steve Henson*
7191 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7193 * Fix for TLS record tampering bug. A carefully crafted invalid
7194 handshake could crash OpenSSL with a NULL pointer exception.
7195 Thanks to Anton Johansson for reporting this issues.
7198 * Keep original DTLS digest and encryption contexts in retransmission
7199 structures so we can use the previous session parameters if they need
7200 to be resent. ([CVE-2013-6450])
7204 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7205 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7206 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7207 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7208 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7209 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7211 *Rob Stradling, Adam Langley*
7213 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7215 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7216 supporting platforms or when small records were transferred.
7218 *Andy Polyakov, Steve Henson*
7220 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7222 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7224 This addresses the flaw in CBC record processing discovered by
7225 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7226 at: <http://www.isg.rhul.ac.uk/tls/>
7228 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7229 Security Group at Royal Holloway, University of London
7230 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7231 Emilia Käsper for the initial patch.
7234 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7236 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7237 ciphersuites which can be exploited in a denial of service attack.
7238 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7239 and detecting this bug and to Wolfgang Ettlinger
7240 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7245 * Return an error when checking OCSP signatures when key is NULL.
7246 This fixes a DoS attack. ([CVE-2013-0166])
7250 * Make openssl verify return errors.
7252 *Chris Palmer <palmer@google.com> and Ben Laurie*
7254 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7255 the right response is stapled. Also change SSL_get_certificate()
7256 so it returns the certificate actually sent.
7257 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7259 *Rob Stradling <rob.stradling@comodo.com>*
7261 * Fix possible deadlock when decoding public keys.
7265 * Don't use TLS 1.0 record version number in initial client hello
7270 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7272 * Sanity check record length before skipping explicit IV in TLS
7273 1.2, 1.1 and DTLS to fix DoS attack.
7275 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7276 fuzzing as a service testing platform.
7281 * Initialise tkeylen properly when encrypting CMS messages.
7282 Thanks to Solar Designer of Openwall for reporting this issue.
7286 * In FIPS mode don't try to use composite ciphers as they are not
7291 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7293 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7294 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7295 mean any application compiled against OpenSSL 1.0.0 headers setting
7296 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7297 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7298 0x10000000L Any application which was previously compiled against
7299 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7300 will need to be recompiled as a result. Letting be results in
7301 inability to disable specifically TLS 1.1 and in client context,
7302 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7306 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7307 disable just protocol X, but all protocols above X *if* there are
7308 protocols *below* X still enabled. In more practical terms it means
7309 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7310 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7311 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7316 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7318 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7319 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7320 in CRYPTO_realloc_clean.
7322 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7323 issue and to Adam Langley <agl@chromium.org> for fixing it.
7326 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7328 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7332 * Workarounds for some broken servers that "hang" if a client hello
7333 record length exceeds 255 bytes.
7335 1. Do not use record version number > TLS 1.0 in initial client
7336 hello: some (but not all) hanging servers will now work.
7337 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7338 the number of ciphers sent in the client hello. This should be
7339 set to an even number, such as 50, for example by passing:
7340 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7341 Most broken servers should now work.
7342 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7343 TLS 1.2 client support entirely.
7347 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7351 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7353 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7354 STRING form instead of a DigestInfo.
7358 * The format used for MDC2 RSA signatures is inconsistent between EVP
7359 and the RSA_sign/RSA_verify functions. This was made more apparent when
7360 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7361 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7362 the correct format in RSA_verify so both forms transparently work.
7366 * Some servers which support TLS 1.0 can choke if we initially indicate
7367 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7368 encrypted premaster secret. As a workaround use the maximum permitted
7369 client version in client hello, this should keep such servers happy
7370 and still work with previous versions of OpenSSL.
7374 * Add support for TLS/DTLS heartbeats.
7376 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7378 * Add support for SCTP.
7380 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7382 * Improved PRNG seeding for VOS.
7384 *Paul Green <Paul.Green@stratus.com>*
7386 * Extensive assembler packs updates, most notably:
7388 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7389 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7390 - x86_64: bit-sliced AES implementation;
7391 - ARM: NEON support, contemporary platforms optimizations;
7392 - s390x: z196 support;
7393 - `*`: GHASH and GF(2^m) multiplication implementations;
7397 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7398 (removal of unnecessary code)
7400 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7402 * Add TLS key material exporter from RFC 5705.
7406 * Add DTLS-SRTP negotiation from RFC 5764.
7410 * Add Next Protocol Negotiation,
7411 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7412 disabled with a no-npn flag to config or Configure. Code donated
7415 *Adam Langley <agl@google.com> and Ben Laurie*
7417 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7418 NIST-P256, NIST-P521, with constant-time single point multiplication on
7419 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7420 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7421 Code made available under Apache License version 2.0.
7423 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7424 line to include this in your build of OpenSSL, and run "make depend" (or
7425 "make update"). This enables the following EC_METHODs:
7427 EC_GFp_nistp224_method()
7428 EC_GFp_nistp256_method()
7429 EC_GFp_nistp521_method()
7431 EC_GROUP_new_by_curve_name() will automatically use these (while
7432 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7435 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7437 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7438 all platforms. Move ssize_t definition from e_os.h to the public
7439 header file e_os2.h as it now appears in public header file cms.h
7443 * New -sigopt option to the ca, req and x509 utilities. Additional
7444 signature parameters can be passed using this option and in
7449 * Add RSA PSS signing function. This will generate and set the
7450 appropriate AlgorithmIdentifiers for PSS based on those in the
7451 corresponding EVP_MD_CTX structure. No application support yet.
7455 * Support for companion algorithm specific ASN1 signing routines.
7456 New function ASN1_item_sign_ctx() signs a pre-initialised
7457 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7458 the appropriate parameters.
7462 * Add new algorithm specific ASN1 verification initialisation function
7463 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7464 handling will be the same no matter what EVP_PKEY_METHOD is used.
7465 Add a PSS handler to support verification of PSS signatures: checked
7466 against a number of sample certificates.
7470 * Add signature printing for PSS. Add PSS OIDs.
7472 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7474 * Add algorithm specific signature printing. An individual ASN1 method
7475 can now print out signatures instead of the standard hex dump.
7477 More complex signatures (e.g. PSS) can print out more meaningful
7478 information. Include DSA version that prints out the signature
7483 * Password based recipient info support for CMS library: implementing
7488 * Split password based encryption into PBES2 and PBKDF2 functions. This
7489 neatly separates the code into cipher and PBE sections and is required
7490 for some algorithms that split PBES2 into separate pieces (such as
7491 password based CMS).
7495 * Session-handling fixes:
7496 - Fix handling of connections that are resuming with a session ID,
7497 but also support Session Tickets.
7498 - Fix a bug that suppressed issuing of a new ticket if the client
7499 presented a ticket with an expired session.
7500 - Try to set the ticket lifetime hint to something reasonable.
7501 - Make tickets shorter by excluding irrelevant information.
7502 - On the client side, don't ignore renewed tickets.
7504 *Adam Langley, Bodo Moeller (Google)*
7506 * Fix PSK session representation.
7510 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7512 This work was sponsored by Intel.
7516 * Add GCM support to TLS library. Some custom code is needed to split
7517 the IV between the fixed (from PRF) and explicit (from TLS record)
7518 portions. This adds all GCM ciphersuites supported by RFC5288 and
7519 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7520 add a special AESGCM string for GCM only.
7524 * Expand range of ctrls for AES GCM. Permit setting invocation
7525 field on decrypt and retrieval of invocation field only on encrypt.
7529 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7530 As required by RFC5289 these ciphersuites cannot be used if for
7531 versions of TLS earlier than 1.2.
7535 * For FIPS capable OpenSSL interpret a NULL default public key method
7536 as unset and return the appropriate default but do *not* set the default.
7537 This means we can return the appropriate method in applications that
7538 switch between FIPS and non-FIPS modes.
7542 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7543 ENGINE is used then we cannot handle that in the FIPS module so we
7544 keep original code iff non-FIPS operations are allowed.
7548 * Add -attime option to openssl utilities.
7550 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7552 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7556 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7557 FIPS EC methods unconditionally for now.
7561 * New build option no-ec2m to disable characteristic 2 code.
7565 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7566 all cases can be covered as some introduce binary incompatibilities.
7570 * Redirect RSA operations to FIPS module including keygen,
7571 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7575 * Add similar low level API blocking to ciphers.
7579 * Low level digest APIs are not approved in FIPS mode: any attempt
7580 to use these will cause a fatal error. Applications that *really* want
7581 to use them can use the `private_*` version instead.
7585 * Redirect cipher operations to FIPS module for FIPS builds.
7589 * Redirect digest operations to FIPS module for FIPS builds.
7593 * Update build system to add "fips" flag which will link in fipscanister.o
7594 for static and shared library builds embedding a signature if needed.
7598 * Output TLS supported curves in preference order instead of numerical
7599 order. This is currently hardcoded for the highest order curves first.
7600 This should be configurable so applications can judge speed vs strength.
7604 * Add TLS v1.2 server support for client authentication.
7608 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7613 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7614 FIPS modules versions.
7618 * Add TLS v1.2 client side support for client authentication. Keep cache
7619 of handshake records longer as we don't know the hash algorithm to use
7620 until after the certificate request message is received.
7624 * Initial TLS v1.2 client support. Add a default signature algorithms
7625 extension including all the algorithms we support. Parse new signature
7626 format in client key exchange. Relax some ECC signing restrictions for
7627 TLS v1.2 as indicated in RFC5246.
7631 * Add server support for TLS v1.2 signature algorithms extension. Switch
7632 to new signature format when needed using client digest preference.
7633 All server ciphersuites should now work correctly in TLS v1.2. No client
7634 support yet and no support for client certificates.
7638 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7639 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7640 ciphersuites. At present only RSA key exchange ciphersuites work with
7641 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7642 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7643 and version checking.
7647 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7648 with this defined it will not be affected by any changes to ssl internal
7649 structures. Add several utility functions to allow openssl application
7650 to work with OPENSSL_NO_SSL_INTERN defined.
7654 * A long standing patch to add support for SRP from EdelWeb (Peter
7655 Sylvester and Christophe Renou) was integrated.
7656 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7657 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7660 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7664 * Permit abbreviated handshakes when renegotiating using the function
7665 SSL_renegotiate_abbreviated().
7667 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7669 * Add call to ENGINE_register_all_complete() to
7670 ENGINE_load_builtin_engines(), so some implementations get used
7671 automatically instead of needing explicit application support.
7675 * Add support for TLS key exporter as described in RFC5705.
7677 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7679 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7680 a few changes are required:
7682 Add SSL_OP_NO_TLSv1_1 flag.
7683 Add TLSv1_1 methods.
7684 Update version checking logic to handle version 1.1.
7685 Add explicit IV handling (ported from DTLS code).
7686 Add command line options to s_client/s_server.
7693 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7695 * X509_ATTRIBUTE memory leak
7697 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7698 memory. This structure is used by the PKCS#7 and CMS routines so any
7699 application which reads PKCS#7 or CMS data from untrusted sources is
7700 affected. SSL/TLS is not affected.
7702 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7708 * Race condition handling PSK identify hint
7710 If PSK identity hints are received by a multi-threaded client then
7711 the values are wrongly updated in the parent SSL_CTX structure. This can
7712 result in a race condition potentially leading to a double free of the
7718 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7720 * Malformed ECParameters causes infinite loop
7722 When processing an ECParameters structure OpenSSL enters an infinite loop
7723 if the curve specified is over a specially malformed binary polynomial
7726 This can be used to perform denial of service against any
7727 system which processes public keys, certificate requests or
7728 certificates. This includes TLS clients and TLS servers with
7729 client authentication enabled.
7731 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7736 * Exploitable out-of-bounds read in X509_cmp_time
7738 X509_cmp_time does not properly check the length of the ASN1_TIME
7739 string and can read a few bytes out of bounds. In addition,
7740 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7743 An attacker can use this to craft malformed certificates and CRLs of
7744 various sizes and potentially cause a segmentation fault, resulting in
7745 a DoS on applications that verify certificates or CRLs. TLS clients
7746 that verify CRLs are affected. TLS clients and servers with client
7747 authentication enabled may be affected if they use custom verification
7750 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7751 independently by Hanno Böck.
7756 * PKCS7 crash with missing EnvelopedContent
7758 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7759 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7760 with missing content and trigger a NULL pointer dereference on parsing.
7762 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7763 structures from untrusted sources are affected. OpenSSL clients and
7764 servers are not affected.
7766 This issue was reported to OpenSSL by Michal Zalewski (Google).
7771 * CMS verify infinite loop with unknown hash function
7773 When verifying a signedData message the CMS code can enter an infinite loop
7774 if presented with an unknown hash function OID. This can be used to perform
7775 denial of service against any system which verifies signedData messages using
7777 This issue was reported to OpenSSL by Johannes Bauer.
7782 * Race condition handling NewSessionTicket
7784 If a NewSessionTicket is received by a multi-threaded client when attempting to
7785 reuse a previous ticket then a race condition can occur potentially leading to
7786 a double free of the ticket data.
7791 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7793 * Segmentation fault in ASN1_TYPE_cmp fix
7795 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7796 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7797 certificate signature algorithm consistency this can be used to crash any
7798 certificate verification operation and exploited in a DoS attack. Any
7799 application which performs certificate verification is vulnerable including
7800 OpenSSL clients and servers which enable client authentication.
7805 * ASN.1 structure reuse memory corruption fix
7807 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7808 memory corruption via an invalid write. Such reuse is and has been
7809 strongly discouraged and is believed to be rare.
7811 Applications that parse structures containing CHOICE or ANY DEFINED BY
7812 components may be affected. Certificate parsing (d2i_X509 and related
7813 functions) are however not affected. OpenSSL clients and servers are
7819 * PKCS7 NULL pointer dereferences fix
7821 The PKCS#7 parsing code does not handle missing outer ContentInfo
7822 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7823 missing content and trigger a NULL pointer dereference on parsing.
7825 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7826 otherwise parse PKCS#7 structures from untrusted sources are
7827 affected. OpenSSL clients and servers are not affected.
7829 This issue was reported to OpenSSL by Michal Zalewski (Google).
7834 * DoS via reachable assert in SSLv2 servers fix
7836 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7837 servers that both support SSLv2 and enable export cipher suites by sending
7838 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7840 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7841 (OpenSSL development team).
7846 * Use After Free following d2i_ECPrivatekey error fix
7848 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7849 could cause a use after free condition. This, in turn, could cause a double
7850 free in several private key parsing functions (such as d2i_PrivateKey
7851 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7852 for applications that receive EC private keys from untrusted
7853 sources. This scenario is considered rare.
7855 This issue was discovered by the BoringSSL project and fixed in their
7861 * X509_to_X509_REQ NULL pointer deref fix
7863 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7864 the certificate key is invalid. This function is rarely used in practice.
7866 This issue was discovered by Brian Carpenter.
7871 * Removed the export ciphers from the DEFAULT ciphers
7875 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7877 * Build fixes for the Windows and OpenVMS platforms
7879 *Matt Caswell and Richard Levitte*
7881 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7883 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7884 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7885 dereference. This could lead to a Denial Of Service attack. Thanks to
7886 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7891 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7892 dtls1_buffer_record function under certain conditions. In particular this
7893 could occur if an attacker sent repeated DTLS records with the same
7894 sequence number but for the next epoch. The memory leak could be exploited
7895 by an attacker in a Denial of Service attack through memory exhaustion.
7896 Thanks to Chris Mueller for reporting this issue.
7901 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7902 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7903 method would be set to NULL which could later result in a NULL pointer
7904 dereference. Thanks to Frank Schmirler for reporting this issue.
7909 * Abort handshake if server key exchange message is omitted for ephemeral
7912 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7913 reporting this issue.
7918 * Remove non-export ephemeral RSA code on client and server. This code
7919 violated the TLS standard by allowing the use of temporary RSA keys in
7920 non-export ciphersuites and could be used by a server to effectively
7921 downgrade the RSA key length used to a value smaller than the server
7922 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7923 INRIA or reporting this issue.
7928 * Fixed issue where DH client certificates are accepted without verification.
7929 An OpenSSL server will accept a DH certificate for client authentication
7930 without the certificate verify message. This effectively allows a client to
7931 authenticate without the use of a private key. This only affects servers
7932 which trust a client certificate authority which issues certificates
7933 containing DH keys: these are extremely rare and hardly ever encountered.
7934 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7940 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7941 results on some platforms, including x86_64. This bug occurs at random
7942 with a very low probability, and is not known to be exploitable in any
7943 way, though its exact impact is difficult to determine. Thanks to Pieter
7944 Wuille (Blockstream) who reported this issue and also suggested an initial
7945 fix. Further analysis was conducted by the OpenSSL development team and
7946 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7947 the OpenSSL core team.
7952 * Fix various certificate fingerprint issues.
7954 By using non-DER or invalid encodings outside the signed portion of a
7955 certificate the fingerprint can be changed without breaking the signature.
7956 Although no details of the signed portion of the certificate can be changed
7957 this can cause problems with some applications: e.g. those using the
7958 certificate fingerprint for blacklists.
7960 1. Reject signatures with non zero unused bits.
7962 If the BIT STRING containing the signature has non zero unused bits reject
7963 the signature. All current signature algorithms require zero unused bits.
7965 2. Check certificate algorithm consistency.
7967 Check the AlgorithmIdentifier inside TBS matches the one in the
7968 certificate signature. NB: this will result in signature failure
7969 errors for some broken certificates.
7971 Thanks to Konrad Kraszewski from Google for reporting this issue.
7973 3. Check DSA/ECDSA signatures use DER.
7975 Reencode DSA/ECDSA signatures and compare with the original received
7976 signature. Return an error if there is a mismatch.
7978 This will reject various cases including garbage after signature
7979 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7980 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7981 (negative or with leading zeroes).
7983 Further analysis was conducted and fixes were developed by Stephen Henson
7984 of the OpenSSL core team.
7990 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7992 * Session Ticket Memory Leak.
7994 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7995 integrity of that ticket is first verified. In the event of a session
7996 ticket integrity check failing, OpenSSL will fail to free memory
7997 causing a memory leak. By sending a large number of invalid session
7998 tickets an attacker could exploit this issue in a Denial Of Service
8004 * Build option no-ssl3 is incomplete.
8006 When OpenSSL is configured with "no-ssl3" as a build option, servers
8007 could accept and complete a SSL 3.0 handshake, and clients could be
8008 configured to send them.
8011 *Akamai and the OpenSSL team*
8013 * Add support for TLS_FALLBACK_SCSV.
8014 Client applications doing fallback retries should call
8015 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8018 *Adam Langley, Bodo Moeller*
8020 * Add additional DigestInfo checks.
8022 Reencode DigestInto in DER and check against the original when
8023 verifying RSA signature: this will reject any improperly encoded
8024 DigestInfo structures.
8026 Note: this is a precautionary measure and no attacks are currently known.
8030 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8032 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8033 to a denial of service attack. A malicious server can crash the client
8034 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8035 ciphersuite and sending carefully crafted handshake messages.
8037 Thanks to Felix Gröbert (Google) for discovering and researching this
8043 * By sending carefully crafted DTLS packets an attacker could cause openssl
8044 to leak memory. This can be exploited through a Denial of Service attack.
8045 Thanks to Adam Langley for discovering and researching this issue.
8050 * An attacker can force openssl to consume large amounts of memory whilst
8051 processing DTLS handshake messages. This can be exploited through a
8052 Denial of Service attack.
8053 Thanks to Adam Langley for discovering and researching this issue.
8058 * An attacker can force an error condition which causes openssl to crash
8059 whilst processing DTLS packets due to memory being freed twice. This
8060 can be exploited through a Denial of Service attack.
8061 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8067 * If a multithreaded client connects to a malicious server using a resumed
8068 session and the server sends an ec point format extension it could write
8069 up to 255 bytes to freed memory.
8071 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8077 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8078 X509_name_oneline, X509_name_print_ex et al. to leak some information
8079 from the stack. Applications may be affected if they echo pretty printing
8080 output to the attacker.
8082 Thanks to Ivan Fratric (Google) for discovering this issue.
8085 *Emilia Käsper, and Steve Henson*
8087 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8088 for corner cases. (Certain input points at infinity could lead to
8089 bogus results, with non-infinity inputs mapped to infinity too.)
8093 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8095 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8096 handshake can force the use of weak keying material in OpenSSL
8097 SSL/TLS clients and servers.
8099 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8100 researching this issue. ([CVE-2014-0224])
8102 *KIKUCHI Masashi, Steve Henson*
8104 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8105 OpenSSL DTLS client the code can be made to recurse eventually crashing
8108 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8111 *Imre Rad, Steve Henson*
8113 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8114 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8115 client or server. This is potentially exploitable to run arbitrary
8116 code on a vulnerable client or server.
8118 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8120 *Jüri Aedla, Steve Henson*
8122 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8123 are subject to a denial of service attack.
8125 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8126 this issue. ([CVE-2014-3470])
8128 *Felix Gröbert, Ivan Fratric, Steve Henson*
8130 * Harmonize version and its documentation. -f flag is used to display
8133 *mancha <mancha1@zoho.com>*
8135 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8136 in i2d_ECPrivateKey.
8138 *mancha <mancha1@zoho.com>*
8140 * Fix some double frees. These are not thought to be exploitable.
8142 *mancha <mancha1@zoho.com>*
8144 * Fix for the attack described in the paper "Recovering OpenSSL
8145 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8146 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8147 <http://eprint.iacr.org/2014/140>
8149 Thanks to Yuval Yarom and Naomi Benger for discovering this
8150 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8152 *Yuval Yarom and Naomi Benger*
8154 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8156 * Keep original DTLS digest and encryption contexts in retransmission
8157 structures so we can use the previous session parameters if they need
8158 to be resent. ([CVE-2013-6450])
8162 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8163 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8164 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8165 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8166 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8167 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8169 *Rob Stradling, Adam Langley*
8171 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8173 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8175 This addresses the flaw in CBC record processing discovered by
8176 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8177 at: <http://www.isg.rhul.ac.uk/tls/>
8179 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8180 Security Group at Royal Holloway, University of London
8181 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8182 Emilia Käsper for the initial patch.
8185 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8187 * Return an error when checking OCSP signatures when key is NULL.
8188 This fixes a DoS attack. ([CVE-2013-0166])
8192 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8193 the right response is stapled. Also change SSL_get_certificate()
8194 so it returns the certificate actually sent.
8195 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8196 (This is a backport)
8198 *Rob Stradling <rob.stradling@comodo.com>*
8200 * Fix possible deadlock when decoding public keys.
8204 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8206 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8209 * Sanity check record length before skipping explicit IV in DTLS
8212 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8213 fuzzing as a service testing platform.
8218 * Initialise tkeylen properly when encrypting CMS messages.
8219 Thanks to Solar Designer of Openwall for reporting this issue.
8223 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8225 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8226 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8227 in CRYPTO_realloc_clean.
8229 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8230 issue and to Adam Langley <agl@chromium.org> for fixing it.
8233 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8235 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8237 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8238 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8239 content decryption and always return the same error. Note: this attack
8240 needs on average 2^20 messages so it only affects automated senders. The
8241 old behaviour can be re-enabled in the CMS code by setting the
8242 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8243 an MMA defence is not necessary.
8244 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8245 this issue. ([CVE-2012-0884])
8249 * Fix CVE-2011-4619: make sure we really are receiving a
8250 client hello before rejecting multiple SGC restarts. Thanks to
8251 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8255 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8257 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8258 Thanks to Antonio Martin, Enterprise Secure Access Research and
8259 Development, Cisco Systems, Inc. for discovering this bug and
8260 preparing a fix. ([CVE-2012-0050])
8264 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8266 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8267 of the Vaudenay padding oracle attack on CBC mode encryption
8268 which enables an efficient plaintext recovery attack against
8269 the OpenSSL implementation of DTLS. Their attack exploits timing
8270 differences arising during decryption processing. A research
8271 paper describing this attack can be found at:
8272 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8273 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8274 Security Group at Royal Holloway, University of London
8275 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8276 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8277 for preparing the fix. ([CVE-2011-4108])
8279 *Robin Seggelmann, Michael Tuexen*
8281 * Clear bytes used for block padding of SSL 3.0 records.
8284 *Adam Langley (Google)*
8286 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8287 Kadianakis <desnacked@gmail.com> for discovering this issue and
8288 Adam Langley for preparing the fix. ([CVE-2011-4619])
8290 *Adam Langley (Google)*
8292 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8294 *Andrey Kulikov <amdeich@gmail.com>*
8296 * Prevent malformed RFC3779 data triggering an assertion failure.
8297 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8298 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8300 *Rob Austein <sra@hactrn.net>*
8302 * Improved PRNG seeding for VOS.
8304 *Paul Green <Paul.Green@stratus.com>*
8306 * Fix ssl_ciph.c set-up race.
8308 *Adam Langley (Google)*
8310 * Fix spurious failures in ecdsatest.c.
8312 *Emilia Käsper (Google)*
8314 * Fix the BIO_f_buffer() implementation (which was mixing different
8315 interpretations of the `..._len` fields).
8317 *Adam Langley (Google)*
8319 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8320 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8321 threads won't reuse the same blinding coefficients.
8323 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8324 lock to call BN_BLINDING_invert_ex, and avoids one use of
8325 BN_BLINDING_update for each BN_BLINDING structure (previously,
8326 the last update always remained unused).
8328 *Emilia Käsper (Google)*
8330 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8332 *Bob Buckholz (Google)*
8334 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8336 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8337 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8339 *Kaspar Brand <ossl@velox.ch>*
8341 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8342 for multi-threaded use of ECDH. ([CVE-2011-3210])
8344 *Adam Langley (Google)*
8346 * Fix x509_name_ex_d2i memory leak on bad inputs.
8350 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8351 signature public key algorithm by using OID xref utilities instead.
8352 Before this you could only use some ECC ciphersuites with SHA1 only.
8356 * Add protection against ECDSA timing attacks as mentioned in the paper
8357 by Billy Bob Brumley and Nicola Tuveri, see:
8358 <http://eprint.iacr.org/2011/232.pdf>
8360 *Billy Bob Brumley and Nicola Tuveri*
8362 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8364 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8366 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8368 * Fix bug in string printing code: if *any* escaping is enabled we must
8369 escape the escape character (backslash) or the resulting string is
8374 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8376 * Disable code workaround for ancient and obsolete Netscape browsers
8377 and servers: an attacker can use it in a ciphersuite downgrade attack.
8378 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8382 * Fixed J-PAKE implementation error, originally discovered by
8383 Sebastien Martini, further info and confirmation from Stefan
8384 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8388 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8390 * Fix extension code to avoid race conditions which can result in a buffer
8391 overrun vulnerability: resumed sessions must not be modified as they can
8392 be shared by multiple threads. CVE-2010-3864
8396 * Fix WIN32 build system to correctly link an ENGINE directory into
8401 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8403 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8406 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8408 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8410 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8411 context. The operation can be customised via the ctrl mechanism in
8412 case ENGINEs want to include additional functionality.
8416 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8420 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8421 output hashes compatible with older versions of OpenSSL.
8423 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8425 * Fix compression algorithm handling: if resuming a session use the
8426 compression algorithm of the resumed session instead of determining
8427 it from client hello again. Don't allow server to change algorithm.
8431 * Add load_crls() function to commands tidying load_certs() too. Add option
8432 to verify utility to allow additional CRLs to be included.
8436 * Update OCSP request code to permit adding custom headers to the request:
8437 some responders need this.
8441 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8444 *Julia Lawall <julia@diku.dk>*
8446 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8447 needlessly dereferenced structures, used obsolete functions and
8448 didn't handle all updated verify codes correctly.
8452 * Disable MD2 in the default configuration.
8456 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8457 indicate the initial BIO being pushed or popped. This makes it possible
8458 to determine whether the BIO is the one explicitly called or as a result
8459 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8460 it handles reference counts correctly and doesn't zero out the I/O bio
8461 when it is not being explicitly popped. WARNING: applications which
8462 included workarounds for the old buggy behaviour will need to be modified
8463 or they could free up already freed BIOs.
8467 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8468 renaming to all platforms (within the 0.9.8 branch, this was
8469 done conditionally on Netware platforms to avoid a name clash).
8471 *Guenter <lists@gknw.net>*
8473 * Add ECDHE and PSK support to DTLS.
8475 *Michael Tuexen <tuexen@fh-muenster.de>*
8477 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8482 * Add "missing" function EVP_MD_flags() (without this the only way to
8483 retrieve a digest flags is by accessing the structure directly. Update
8484 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8485 or cipher is registered as in the "from" argument. Print out all
8486 registered digests in the dgst usage message instead of manually
8487 attempting to work them out.
8491 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8492 this allows the use of compression and extensions. Change default cipher
8493 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8494 by default unless an application cipher string requests it.
8498 * Alter match criteria in PKCS12_parse(). It used to try to use local
8499 key ids to find matching certificates and keys but some PKCS#12 files
8500 don't follow the (somewhat unwritten) rules and this strategy fails.
8501 Now just gather all certificates together and the first private key
8502 then look for the first certificate that matches the key.
8506 * Support use of registered digest and cipher names for dgst and cipher
8507 commands instead of having to add each one as a special case. So now
8514 openssl dgst -sha256 foo
8516 and this works for ENGINE based algorithms too.
8520 * Update Gost ENGINE to support parameter files.
8522 *Victor B. Wagner <vitus@cryptocom.ru>*
8524 * Support GeneralizedTime in ca utility.
8526 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8528 * Enhance the hash format used for certificate directory links. The new
8529 form uses the canonical encoding (meaning equivalent names will work
8530 even if they aren't identical) and uses SHA1 instead of MD5. This form
8531 is incompatible with the older format and as a result c_rehash should
8532 be used to rebuild symbolic links.
8536 * Make PKCS#8 the default write format for private keys, replacing the
8537 traditional format. This form is standardised, more secure and doesn't
8538 include an implicit MD5 dependency.
8542 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8543 committed to OpenSSL should pass this lot as a minimum.
8547 * Add session ticket override functionality for use by EAP-FAST.
8549 *Jouni Malinen <j@w1.fi>*
8551 * Modify HMAC functions to return a value. Since these can be implemented
8552 in an ENGINE errors can occur.
8556 * Type-checked OBJ_bsearch_ex.
8560 * Type-checked OBJ_bsearch. Also some constification necessitated
8561 by type-checking. Still to come: TXT_DB, bsearch(?),
8562 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8567 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8568 seconds to a tm structure directly, instead of going through OS
8569 specific date routines. This avoids any issues with OS routines such
8570 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8571 and X509_time_adj_ex() to cover the extended range. The existing
8572 X509_time_adj() is still usable and will no longer have any date issues.
8576 * Delta CRL support. New use deltas option which will attempt to locate
8577 and search any appropriate delta CRLs available.
8579 This work was sponsored by Google.
8583 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8584 code and add additional score elements. Validate alternate CRL paths
8585 as part of the CRL checking and indicate a new error "CRL path validation
8586 error" in this case. Applications wanting additional details can use
8587 the verify callback and check the new "parent" field. If this is not
8588 NULL CRL path validation is taking place. Existing applications won't
8589 see this because it requires extended CRL support which is off by
8592 This work was sponsored by Google.
8596 * Support for freshest CRL extension.
8598 This work was sponsored by Google.
8602 * Initial indirect CRL support. Currently only supported in the CRLs
8603 passed directly and not via lookup. Process certificate issuer
8604 CRL entry extension and lookup CRL entries by bother issuer name
8605 and serial number. Check and process CRL issuer entry in IDP extension.
8607 This work was sponsored by Google.
8611 * Add support for distinct certificate and CRL paths. The CRL issuer
8612 certificate is validated separately in this case. Only enabled if
8613 an extended CRL support flag is set: this flag will enable additional
8614 CRL functionality in future.
8616 This work was sponsored by Google.
8620 * Add support for policy mappings extension.
8622 This work was sponsored by Google.
8626 * Fixes to pathlength constraint, self issued certificate handling,
8627 policy processing to align with RFC3280 and PKITS tests.
8629 This work was sponsored by Google.
8633 * Support for name constraints certificate extension. DN, email, DNS
8634 and URI types are currently supported.
8636 This work was sponsored by Google.
8640 * To cater for systems that provide a pointer-based thread ID rather
8641 than numeric, deprecate the current numeric thread ID mechanism and
8642 replace it with a structure and associated callback type. This
8643 mechanism allows a numeric "hash" to be extracted from a thread ID in
8644 either case, and on platforms where pointers are larger than 'long',
8645 mixing is done to help ensure the numeric 'hash' is usable even if it
8646 can't be guaranteed unique. The default mechanism is to use "&errno"
8647 as a pointer-based thread ID to distinguish between threads.
8649 Applications that want to provide their own thread IDs should now use
8650 CRYPTO_THREADID_set_callback() to register a callback that will call
8651 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8653 Note that ERR_remove_state() is now deprecated, because it is tied
8654 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8655 to free the current thread's error state should be replaced by
8656 ERR_remove_thread_state(NULL).
8658 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8659 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8660 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8661 application was previously providing a numeric thread callback that
8662 was inappropriate for distinguishing threads, then uniqueness might
8663 have been obtained with &errno that happened immediately in the
8664 intermediate development versions of OpenSSL; this is no longer the
8665 case, the numeric thread callback will now override the automatic use
8668 *Geoff Thorpe, with help from Bodo Moeller*
8670 * Initial support for different CRL issuing certificates. This covers a
8671 simple case where the self issued certificates in the chain exist and
8672 the real CRL issuer is higher in the existing chain.
8674 This work was sponsored by Google.
8678 * Removed effectively defunct crypto/store from the build.
8682 * Revamp of STACK to provide stronger type-checking. Still to come:
8683 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8684 ASN1_STRING, CONF_VALUE.
8688 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8689 RAM on SSL connections. This option can save about 34k per idle SSL.
8693 * Revamp of LHASH to provide stronger type-checking. Still to come:
8694 STACK, TXT_DB, bsearch, qsort.
8698 * Initial support for Cryptographic Message Syntax (aka CMS) based
8699 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8700 support for data, signedData, compressedData, digestedData and
8701 encryptedData, envelopedData types included. Scripts to check against
8702 RFC4134 examples draft and interop and consistency checks of many
8703 content types and variants.
8707 * Add options to enc utility to support use of zlib compression BIO.
8711 * Extend mk1mf to support importing of options and assembly language
8712 files from Configure script, currently only included in VC-WIN32.
8713 The assembly language rules can now optionally generate the source
8714 files from the associated perl scripts.
8718 * Implement remaining functionality needed to support GOST ciphersuites.
8719 Interop testing has been performed using CryptoPro implementations.
8721 *Victor B. Wagner <vitus@cryptocom.ru>*
8723 * s390x assembler pack.
8727 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8732 * Implement Opaque PRF Input TLS extension as specified in
8733 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8734 official specification yet and no extension type assignment by
8735 IANA exists, this extension (for now) will have to be explicitly
8736 enabled when building OpenSSL by providing the extension number
8737 to use. For example, specify an option
8739 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8741 to the "config" or "Configure" script to enable the extension,
8742 assuming extension number 0x9527 (which is a completely arbitrary
8743 and unofficial assignment based on the MD5 hash of the Internet
8744 Draft). Note that by doing so, you potentially lose
8745 interoperability with other TLS implementations since these might
8746 be using the same extension number for other purposes.
8748 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8749 opaque PRF input value to use in the handshake. This will create
8750 an internal copy of the length-'len' string at 'src', and will
8751 return non-zero for success.
8753 To get more control and flexibility, provide a callback function
8756 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8757 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8761 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8764 Callback function 'cb' will be called in handshakes, and is
8765 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8766 Argument 'arg' is for application purposes (the value as given to
8767 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8768 be provided to the callback function). The callback function
8769 has to return non-zero to report success: usually 1 to use opaque
8770 PRF input just if possible, or 2 to enforce use of the opaque PRF
8771 input. In the latter case, the library will abort the handshake
8772 if opaque PRF input is not successfully negotiated.
8774 Arguments 'peerinput' and 'len' given to the callback function
8775 will always be NULL and 0 in the case of a client. A server will
8776 see the client's opaque PRF input through these variables if
8777 available (NULL and 0 otherwise). Note that if the server
8778 provides an opaque PRF input, the length must be the same as the
8779 length of the client's opaque PRF input.
8781 Note that the callback function will only be called when creating
8782 a new session (session resumption can resume whatever was
8783 previously negotiated), and will not be called in SSL 2.0
8784 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8785 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8786 for applications that need to enforce opaque PRF input.
8790 * Update ssl code to support digests other than SHA1+MD5 for handshake
8793 *Victor B. Wagner <vitus@cryptocom.ru>*
8795 * Add RFC4507 support to OpenSSL. This includes the corrections in
8796 RFC4507bis. The encrypted ticket format is an encrypted encoded
8797 SSL_SESSION structure, that way new session features are automatically
8800 If a client application caches session in an SSL_SESSION structure
8801 support is transparent because tickets are now stored in the encoded
8804 The SSL_CTX structure automatically generates keys for ticket
8805 protection in servers so again support should be possible
8806 with no application modification.
8808 If a client or server wishes to disable RFC4507 support then the option
8809 SSL_OP_NO_TICKET can be set.
8811 Add a TLS extension debugging callback to allow the contents of any client
8812 or server extensions to be examined.
8814 This work was sponsored by Google.
8818 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8819 OpenSSL should now compile cleanly on gcc 4.2
8821 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8823 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8824 support including streaming MAC support: this is required for GOST
8825 ciphersuite support.
8827 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8829 * Add option -stream to use PKCS#7 streaming in smime utility. New
8830 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8831 to output in BER and PEM format.
8835 * Experimental support for use of HMAC via EVP_PKEY interface. This
8836 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8837 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8838 ENGINE support for HMAC keys which are unextractable. New -mac and
8839 -macopt options to dgst utility.
8843 * New option -sigopt to dgst utility. Update dgst to use
8844 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8845 alternative signing parameters such as X9.31 or PSS in the dgst
8850 * Change ssl_cipher_apply_rule(), the internal function that does
8851 the work each time a ciphersuite string requests enabling
8852 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8853 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8854 the order of disabled ciphersuites such that those ciphersuites
8855 that most recently went from enabled to disabled not only stay
8856 in order with respect to each other, but also have higher priority
8857 than other disabled ciphersuites the next time ciphersuites are
8860 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8861 the same ciphersuites as with "HIGH" alone, but in a specific
8862 order where the PSK ciphersuites come first (since they are the
8863 most recently disabled ciphersuites when "HIGH" is parsed).
8865 Also, change ssl_create_cipher_list() (using this new
8866 functionality) such that between otherwise identical
8867 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8872 * Change ssl_create_cipher_list() so that it automatically
8873 arranges the ciphersuites in reasonable order before starting
8874 to process the rule string. Thus, the definition for "DEFAULT"
8875 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8876 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8877 This makes it much easier to arrive at a reasonable default order
8878 in applications for which anonymous ciphers are OK (meaning
8879 that you can't actually use DEFAULT).
8881 *Bodo Moeller; suggested by Victor Duchovni*
8883 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8884 processing) into multiple integers instead of setting
8885 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8886 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8887 (These masks as well as the individual bit definitions are hidden
8888 away into the non-exported interface ssl/ssl_locl.h, so this
8889 change to the definition of the SSL_CIPHER structure shouldn't
8890 affect applications.) This give us more bits for each of these
8891 categories, so there is no longer a need to coagulate AES128 and
8892 AES256 into a single algorithm bit, and to coagulate Camellia128
8893 and Camellia256 into a single algorithm bit, which has led to all
8896 Thus, among other things, the kludge introduced in 0.9.7m and
8897 0.9.8e for masking out AES256 independently of AES128 or masking
8898 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8900 With the change, we also introduce new ciphersuite aliases that
8901 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8906 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8907 Use the leftmost N bytes of the signature input if the input is
8908 larger than the prime q (with N being the size in bytes of q).
8912 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8913 it yet and it is largely untested.
8917 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8921 * Initial incomplete changes to avoid need for function casts in OpenSSL
8922 some compilers (gcc 4.2 and later) reject their use. Safestack is
8923 reimplemented. Update ASN1 to avoid use of legacy functions.
8927 * Win32/64 targets are linked with Winsock2.
8931 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8932 to external functions. This can be used to increase CRL handling
8933 efficiency especially when CRLs are very large by (for example) storing
8934 the CRL revoked certificates in a database.
8938 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8939 new CRLs added to a directory can be used. New command line option
8940 -verify_return_error to s_client and s_server. This causes real errors
8941 to be returned by the verify callback instead of carrying on no matter
8942 what. This reflects the way a "real world" verify callback would behave.
8946 * GOST engine, supporting several GOST algorithms and public key formats.
8947 Kindly donated by Cryptocom.
8951 * Partial support for Issuing Distribution Point CRL extension. CRLs
8952 partitioned by DP are handled but no indirect CRL or reason partitioning
8953 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8954 selected via a scoring technique which handles IDP and AKID in CRLs.
8958 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8959 will ultimately be used for all verify operations: this will remove the
8960 X509_STORE dependency on certificate verification and allow alternative
8961 lookup methods. X509_STORE based implementations of these two callbacks.
8965 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8966 Modify get_crl() to find a valid (unexpired) CRL if possible.
8970 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8971 this would be called X509_CRL_cmp() but that name is already used by
8972 a function that just compares CRL issuer names. Cache several CRL
8973 extensions in X509_CRL structure and cache CRLDP in X509.
8977 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8978 this maps equivalent X509_NAME structures into a consistent structure.
8979 Name comparison can then be performed rapidly using memcmp().
8983 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8988 * Allow digests to supply their own micalg string for S/MIME type using
8989 the ctrl EVP_MD_CTRL_MICALG.
8993 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8994 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8995 ctrl. It can then customise the structure before and/or after signing
9000 * New function OBJ_add_sigid() to allow application defined signature OIDs
9001 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9002 to free up any added signature OIDs.
9006 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9007 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9008 digest and cipher tables. New options added to openssl utility:
9009 list-message-digest-algorithms and list-cipher-algorithms.
9013 * Change the array representation of binary polynomials: the list
9014 of degrees of non-zero coefficients is now terminated with -1.
9015 Previously it was terminated with 0, which was also part of the
9016 value; thus, the array representation was not applicable to
9017 polynomials where t^0 has coefficient zero. This change makes
9018 the array representation useful in a more general context.
9022 * Various modifications and fixes to SSL/TLS cipher string
9023 handling. For ECC, the code now distinguishes between fixed ECDH
9024 with RSA certificates on the one hand and with ECDSA certificates
9025 on the other hand, since these are separate ciphersuites. The
9026 unused code for Fortezza ciphersuites has been removed.
9028 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9029 (not "ECDHE"). For consistency with the code for DH
9030 certificates, use of ECDH certificates is now considered ECDH
9031 authentication, not RSA or ECDSA authentication (the latter is
9032 merely the CA's signing algorithm and not actively used in the
9035 The temporary ciphersuite alias "ECCdraft" is no longer
9036 available, and ECC ciphersuites are no longer excluded from "ALL"
9037 and "DEFAULT". The following aliases now exist for RFC 4492
9038 ciphersuites, most of these by analogy with the DH case:
9040 kECDHr - ECDH cert, signed with RSA
9041 kECDHe - ECDH cert, signed with ECDSA
9042 kECDH - ECDH cert (signed with either RSA or ECDSA)
9043 kEECDH - ephemeral ECDH
9044 ECDH - ECDH cert or ephemeral ECDH
9050 AECDH - anonymous ECDH
9051 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9055 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9056 Use correct micalg parameters depending on digest(s) in signed message.
9060 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9061 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9065 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9066 an engine to register a method. Add ENGINE lookups for methods and
9067 functional reference processing.
9071 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9072 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9077 * New -resign option to smime utility. This adds one or more signers
9078 to an existing PKCS#7 signedData structure. Also -md option to use an
9079 alternative message digest algorithm for signing.
9083 * Tidy up PKCS#7 routines and add new functions to make it easier to
9084 create PKCS7 structures containing multiple signers. Update smime
9085 application to support multiple signers.
9089 * New -macalg option to pkcs12 utility to allow setting of an alternative
9094 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9095 Reorganize PBE internals to lookup from a static table using NIDs,
9096 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9097 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9098 PRF which will be automatically used with PBES2.
9102 * Replace the algorithm specific calls to generate keys in "req" with the
9107 * Update PKCS#7 enveloped data routines to use new API. This is now
9108 supported by any public key method supporting the encrypt operation. A
9109 ctrl is added to allow the public key algorithm to examine or modify
9110 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9115 * Add a ctrl to asn1 method to allow a public key algorithm to express
9116 a default digest type to use. In most cases this will be SHA1 but some
9117 algorithms (such as GOST) need to specify an alternative digest. The
9118 return value indicates how strong the preference is 1 means optional and
9119 2 is mandatory (that is it is the only supported type). Modify
9120 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9121 use the default md. Update openssl utilities to use the default digest
9122 type for signing if it is not explicitly indicated.
9126 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9127 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9128 signing method from the key type. This effectively removes the link
9129 between digests and public key types.
9133 * Add an OID cross reference table and utility functions. Its purpose is to
9134 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9135 rsaEncryption. This will allow some of the algorithm specific hackery
9136 needed to use the correct OID to be removed.
9140 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9141 structures for PKCS7_sign(). They are now set up by the relevant public
9146 * Add provisional EC pkey method with support for ECDSA and ECDH.
9150 * Add support for key derivation (agreement) in the API, DH method and
9155 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9156 public and private key formats. As a side effect these add additional
9157 command line functionality not previously available: DSA signatures can be
9158 generated and verified using pkeyutl and DH key support and generation in
9165 *Oliver Tappe <zooey@hirschkaefer.de>*
9167 * New make target "install_html_docs" installs HTML renditions of the
9170 *Oliver Tappe <zooey@hirschkaefer.de>*
9172 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9173 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9174 support key and parameter generation and add initial key generation
9175 functionality for RSA.
9179 * Add functions for main EVP_PKEY_method operations. The undocumented
9180 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9181 `EVP_PKEY_{encrypt,decrypt}_old`.
9185 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9186 key API, doesn't do much yet.
9190 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9191 public key algorithms. New option to openssl utility:
9192 "list-public-key-algorithms" to print out info.
9196 * Implement the Supported Elliptic Curves Extension for
9197 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9201 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9202 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9206 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9207 utilities such as rsa, dsa, dsaparam etc except they process any key
9212 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9213 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9214 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9219 * Initial support for pluggable public key ASN1.
9220 De-spaghettify the public key ASN1 handling. Move public and private
9221 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9222 algorithm specific handling to a single module within the relevant
9223 algorithm directory. Add functions to allow (near) opaque processing
9224 of public and private key structures.
9228 * Implement the Supported Point Formats Extension for
9229 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9233 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9234 for the psk identity [hint] and the psk callback functions to the
9235 SSL_SESSION, SSL and SSL_CTX structure.
9238 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9242 SSL_CTX_use_psk_identity_hint
9243 SSL_get_psk_identity_hint
9244 SSL_get_psk_identity
9245 SSL_use_psk_identity_hint
9247 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9249 * Add RFC 3161 compliant time stamp request creation, response generation
9250 and response verification functionality.
9252 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9254 * Add initial support for TLS extensions, specifically for the server_name
9255 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9256 have new members for a host name. The SSL data structure has an
9257 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9258 stored in that context to allow for session resumption, even after the
9259 SSL has been switched to a new SSL_CTX in reaction to a client's
9260 server_name extension.
9262 New functions (subject to change):
9264 SSL_get_servername()
9265 SSL_get_servername_type()
9268 New CTRL codes and macros (subject to change):
9270 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9271 - SSL_CTX_set_tlsext_servername_callback()
9272 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9273 - SSL_CTX_set_tlsext_servername_arg()
9274 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9276 openssl s_client has a new '-servername ...' option.
9278 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9279 '-key2 ...', '-servername_fatal' (subject to change). This allows
9280 testing the HostName extension for a specific single host name ('-cert'
9281 and '-key' remain fallbacks for handshakes without HostName
9282 negotiation). If the unrecognized_name alert has to be sent, this by
9283 default is a warning; it becomes fatal with the '-servername_fatal'
9286 *Peter Sylvester, Remy Allais, Christophe Renou*
9288 * Whirlpool hash implementation is added.
9292 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9293 bn(64,32). Because of instruction set limitations it doesn't have
9294 any negative impact on performance. This was done mostly in order
9295 to make it possible to share assembler modules, such as bn_mul_mont
9296 implementations, between 32- and 64-bit builds without hassle.
9300 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9301 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9306 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9307 dedicated Montgomery multiplication procedure, is introduced.
9308 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9309 "64-bit" performance on certain 32-bit targets.
9313 * New option SSL_OP_NO_COMP to disable use of compression selectively
9314 in SSL structures. New SSL ctrl to set maximum send fragment size.
9315 Save memory by setting the I/O buffer sizes dynamically instead of
9316 using the maximum available value.
9320 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9321 in addition to the text details.
9325 * Very, very preliminary EXPERIMENTAL support for printing of general
9326 ASN1 structures. This currently produces rather ugly output and doesn't
9327 handle several customised structures at all.
9331 * Integrated support for PVK file format and some related formats such
9332 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9333 these in the 'rsa' and 'dsa' utilities.
9337 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9341 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9342 place for the (very old) "NETSCAPE" format certificates which are now
9343 handled using new ASN1 code equivalents.
9347 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9348 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9349 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9353 * Modify CRL distribution points extension code to print out previously
9354 unsupported fields. Enhance extension setting code to allow setting of
9359 * Add print and set support for Issuing Distribution Point CRL extension.
9363 * Change 'Configure' script to enable Camellia by default.
9370 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9372 * When rejecting SSL/TLS records due to an incorrect version number, never
9373 update s->server with a new major version number. As of
9374 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9375 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9376 the previous behavior could result in a read attempt at NULL when
9377 receiving specific incorrect SSL/TLS records once record payload
9378 protection is active. ([CVE-2010-0740])
9380 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9382 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9383 could be crashed if the relevant tables were not present (e.g. chrooted).
9385 *Tomas Hoger <thoger@redhat.com>*
9387 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9389 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9391 *Martin Olsson, Neel Mehta*
9393 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9394 accommodate for stack sorting, always a write lock!).
9398 * On some versions of WIN32 Heap32Next is very slow. This can cause
9399 excessive delays in the RAND_poll(): over a minute. As a workaround
9400 include a time check in the inner Heap32Next loop too.
9404 * The code that handled flushing of data in SSL/TLS originally used the
9405 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9406 the problem outlined in PR#1949. The fix suggested there however can
9407 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9408 of Apache). So instead simplify the code to flush unconditionally.
9409 This should be fine since flushing with no data to flush is a no op.
9413 * Handle TLS versions 2.0 and later properly and correctly use the
9414 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9415 off ancient servers have a habit of sticking around for a while...
9419 * Modify compression code so it frees up structures without using the
9420 ex_data callbacks. This works around a problem where some applications
9421 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9422 restarting) then use compression (e.g. SSL with compression) later.
9423 This results in significant per-connection memory leaks and
9424 has caused some security issues including CVE-2008-1678 and
9429 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9430 change when encrypting or decrypting.
9434 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9435 connect and renegotiate with servers which do not support RI.
9436 Until RI is more widely deployed this option is enabled by default.
9440 * Add "missing" ssl ctrls to clear options and mode.
9444 * If client attempts to renegotiate and doesn't support RI respond with
9445 a no_renegotiation alert as required by RFC5746. Some renegotiating
9446 TLS clients will continue a connection gracefully when they receive
9447 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9448 waiting for a server hello which it will never receive. Now we treat a
9449 received no_renegotiation alert as a fatal error. This is because
9450 applications requesting a renegotiation might well expect it to succeed
9451 and would have no code in place to handle the server denying it so the
9452 only safe thing to do is to terminate the connection.
9456 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9457 peer supports secure renegotiation and 0 otherwise. Print out peer
9458 renegotiation support in s_client/s_server.
9462 * Replace the highly broken and deprecated SPKAC certification method with
9463 the updated NID creation version. This should correctly handle UTF8.
9467 * Implement RFC5746. Re-enable renegotiation but require the extension
9468 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9469 turns out to be a bad idea. It has been replaced by
9470 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9471 SSL_CTX_set_options(). This is really not recommended unless you
9472 know what you are doing.
9474 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9476 * Fixes to stateless session resumption handling. Use initial_ctx when
9477 issuing and attempting to decrypt tickets in case it has changed during
9478 servername handling. Use a non-zero length session ID when attempting
9479 stateless session resumption: this makes it possible to determine if
9480 a resumption has occurred immediately after receiving server hello
9481 (several places in OpenSSL subtly assume this) instead of later in
9486 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9487 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9488 fixes for a few places where the return code is not checked
9491 *Julia Lawall <julia@diku.dk>*
9493 * Add --strict-warnings option to Configure script to include devteam
9494 warnings in other configurations.
9498 * Add support for --libdir option and LIBDIR variable in makefiles. This
9499 makes it possible to install openssl libraries in locations which
9500 have names other than "lib", for example "/usr/lib64" which some
9503 *Steve Henson, based on patch from Jeremy Utley*
9505 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9506 X690 8.9.12 and can produce some misleading textual output of OIDs.
9508 *Steve Henson, reported by Dan Kaminsky*
9510 * Delete MD2 from algorithm tables. This follows the recommendation in
9511 several standards that it is not used in new applications due to
9512 several cryptographic weaknesses. For binary compatibility reasons
9513 the MD2 API is still compiled in by default.
9517 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9522 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9523 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9526 *Guenter <lists@gknw.net>*
9528 * Fix the server certificate chain building code to use X509_verify_cert(),
9529 it used to have an ad-hoc builder which was unable to cope with anything
9530 other than a simple chain.
9532 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9534 * Don't check self signed certificate signatures in X509_verify_cert()
9535 by default (a flag can override this): it just wastes time without
9536 adding any security. As a useful side effect self signed root CAs
9537 with non-FIPS digests are now usable in FIPS mode.
9541 * In dtls1_process_out_of_seq_message() the check if the current message
9542 is already buffered was missing. For every new message was memory
9543 allocated, allowing an attacker to perform an denial of service attack
9544 with sending out of seq handshake messages until there is no memory
9545 left. Additionally every future message was buffered, even if the
9546 sequence number made no sense and would be part of another handshake.
9547 So only messages with sequence numbers less than 10 in advance will be
9548 buffered. ([CVE-2009-1378])
9550 *Robin Seggelmann, discovered by Daniel Mentz*
9552 * Records are buffered if they arrive with a future epoch to be
9553 processed after finishing the corresponding handshake. There is
9554 currently no limitation to this buffer allowing an attacker to perform
9555 a DOS attack with sending records with future epochs until there is no
9556 memory left. This patch adds the pqueue_size() function to determine
9557 the size of a buffer and limits the record buffer to 100 entries.
9560 *Robin Seggelmann, discovered by Daniel Mentz*
9562 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9563 parent structure is freed. ([CVE-2009-1379])
9567 * Handle non-blocking I/O properly in SSL_shutdown() call.
9569 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9571 * Add `2.5.4.*` OIDs
9573 *Ilya O. <vrghost@gmail.com>*
9575 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9577 * Disable renegotiation completely - this fixes a severe security
9578 problem ([CVE-2009-3555]) at the cost of breaking all
9579 renegotiation. Renegotiation can be re-enabled by setting
9580 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9581 run-time. This is really not recommended unless you know what
9586 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9588 * Don't set val to NULL when freeing up structures, it is freed up by
9589 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9590 zeroing past the valid field. ([CVE-2009-0789])
9592 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9594 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9595 checked correctly. This would allow some invalid signed attributes to
9596 appear to verify correctly. ([CVE-2009-0591])
9598 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9600 * Reject UniversalString and BMPString types with invalid lengths. This
9601 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9602 a legal length. ([CVE-2009-0590])
9606 * Set S/MIME signing as the default purpose rather than setting it
9607 unconditionally. This allows applications to override it at the store
9612 * Permit restricted recursion of ASN1 strings. This is needed in practice
9613 to handle some structures.
9617 * Improve efficiency of mem_gets: don't search whole buffer each time
9620 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9622 * New -hex option for openssl rand.
9626 * Print out UTF8String and NumericString when parsing ASN1.
9630 * Support NumericString type for name components.
9634 * Allow CC in the environment to override the automatically chosen
9635 compiler. Note that nothing is done to ensure flags work with the
9640 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9642 * Properly check EVP_VerifyFinal() and similar return values
9645 *Ben Laurie, Bodo Moeller, Google Security Team*
9647 * Enable TLS extensions by default.
9651 * Allow the CHIL engine to be loaded, whether the application is
9652 multithreaded or not. (This does not release the developer from the
9653 obligation to set up the dynamic locking callbacks.)
9655 *Sander Temme <sander@temme.net>*
9657 * Use correct exit code if there is an error in dgst command.
9659 *Steve Henson; problem pointed out by Roland Dirlewanger*
9661 * Tweak Configure so that you need to say "experimental-jpake" to enable
9662 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9666 * Add experimental JPAKE support, including demo authentication in
9667 s_client and s_server.
9671 * Set the comparison function in v3_addr_canonize().
9673 *Rob Austein <sra@hactrn.net>*
9675 * Add support for XMPP STARTTLS in s_client.
9677 *Philip Paeps <philip@freebsd.org>*
9679 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9680 to ensure that even with this option, only ciphersuites in the
9681 server's preference list will be accepted. (Note that the option
9682 applies only when resuming a session, so the earlier behavior was
9683 just about the algorithm choice for symmetric cryptography.)
9687 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9689 * Fix NULL pointer dereference if a DTLS server received
9690 ChangeCipherSpec as first record ([CVE-2009-1386]).
9694 * Fix a state transition in s3_srvr.c and d1_srvr.c
9695 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9699 * The fix in 0.9.8c that supposedly got rid of unsafe
9700 double-checked locking was incomplete for RSA blinding,
9701 addressing just one layer of what turns out to have been
9702 doubly unsafe triple-checked locking.
9704 So now fix this for real by retiring the MONT_HELPER macro
9705 in crypto/rsa/rsa_eay.c.
9707 *Bodo Moeller; problem pointed out by Marius Schilder*
9709 * Various precautionary measures:
9711 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9713 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9714 (NB: This would require knowledge of the secret session ticket key
9715 to exploit, in which case you'd be SOL either way.)
9717 - Change bn_nist.c so that it will properly handle input BIGNUMs
9718 outside the expected range.
9720 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9723 *Neel Mehta, Bodo Moeller*
9725 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9726 the load fails. Useful for distros.
9728 *Ben Laurie and the FreeBSD team*
9730 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9734 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9738 * Expand ENGINE to support engine supplied SSL client certificate functions.
9740 This work was sponsored by Logica.
9744 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9745 keystores. Support for SSL/TLS client authentication too.
9746 Not compiled unless enable-capieng specified to Configure.
9748 This work was sponsored by Logica.
9752 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9753 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9754 attribute creation routines such as certificate requests and PKCS#12
9759 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9761 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9762 handshake which could lead to a client crash as found using the
9763 Codenomicon TLS test suite ([CVE-2008-1672])
9765 *Steve Henson, Mark Cox*
9767 * Fix double free in TLS server name extensions which could lead to
9768 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9772 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9774 Clear the error queue to ensure that error entries left from
9775 older function calls do not interfere with the correct operation.
9777 *Lutz Jaenicke, Erik de Castro Lopo*
9779 * Remove root CA certificates of commercial CAs:
9781 The OpenSSL project does not recommend any specific CA and does not
9782 have any policy with respect to including or excluding any CA.
9783 Therefore it does not make any sense to ship an arbitrary selection
9784 of root CA certificates with the OpenSSL software.
9788 * RSA OAEP patches to fix two separate invalid memory reads.
9789 The first one involves inputs when 'lzero' is greater than
9790 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9791 before the beginning of from). The second one involves inputs where
9792 the 'db' section contains nothing but zeroes (there is a one-byte
9793 invalid read after the end of 'db').
9795 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9797 * Partial backport from 0.9.9-dev:
9799 Introduce bn_mul_mont (dedicated Montgomery multiplication
9800 procedure) as a candidate for BIGNUM assembler implementation.
9801 While 0.9.9-dev uses assembler for various architectures, only
9802 x86_64 is available by default here in the 0.9.8 branch, and
9803 32-bit x86 is available through a compile-time setting.
9805 To try the 32-bit x86 assembler implementation, use Configure
9806 option "enable-montasm" (which exists only for this backport).
9808 As "enable-montasm" for 32-bit x86 disclaims code stability
9809 anyway, in this constellation we activate additional code
9810 backported from 0.9.9-dev for further performance improvements,
9811 namely BN_from_montgomery_word. (To enable this otherwise,
9812 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9814 *Andy Polyakov (backport partially by Bodo Moeller)*
9816 * Add TLS session ticket callback. This allows an application to set
9817 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9818 values. This is useful for key rollover for example where several key
9819 sets may exist with different names.
9823 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9824 This was broken until now in 0.9.8 releases, such that the only way
9825 a registered ENGINE could be used (assuming it initialises
9826 successfully on the host) was to explicitly set it as the default
9827 for the relevant algorithms. This is in contradiction with 0.9.7
9828 behaviour and the documentation. With this fix, when an ENGINE is
9829 registered into a given algorithm's table of implementations, the
9830 'uptodate' flag is reset so that auto-discovery will be used next
9831 time a new context for that algorithm attempts to select an
9834 *Ian Lister (tweaked by Geoff Thorpe)*
9836 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9837 implementation in the following ways:
9839 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9842 Lack of BER streaming support means one pass streaming processing is
9843 only supported if data is detached: setting the streaming flag is
9844 ignored for embedded content.
9846 CMS support is disabled by default and must be explicitly enabled
9847 with the enable-cms configuration option.
9851 * Update the GMP engine glue to do direct copies between BIGNUM and
9852 mpz_t when openssl and GMP use the same limb size. Otherwise the
9853 existing "conversion via a text string export" trick is still used.
9855 *Paul Sheer <paulsheer@gmail.com>*
9857 * Zlib compression BIO. This is a filter BIO which compressed and
9858 uncompresses any data passed through it.
9862 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9863 RFC3394 compatible AES key wrapping.
9867 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9868 sets string data without copying. X509_ALGOR_set0() and
9869 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9870 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9871 from an X509_ATTRIBUTE structure optionally checking it occurs only
9872 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9877 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9878 to get the expected BN_FLG_CONSTTIME behavior.
9880 *Bodo Moeller (Google)*
9884 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9885 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9886 - added some more tests to do_tests.pl
9887 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9888 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9889 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9890 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9891 - various changes to netware.pl to enable gcc-cross builds on Win32
9893 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9894 - various changes to fix missing prototype warnings
9895 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9896 - added AES, WHIRLPOOL and CPUID assembler code to build files
9897 - added missing AES assembler make rules to mk1mf.pl
9898 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9900 *Guenter Knauf <eflash@gmx.net>*
9902 * Implement certificate status request TLS extension defined in RFC3546.
9903 A client can set the appropriate parameters and receive the encoded
9904 OCSP response via a callback. A server can query the supplied parameters
9905 and set the encoded OCSP response in the callback. Add simplified examples
9906 to s_client and s_server.
9910 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9913 + Binary incompatibility of ssl_ctx_st structure
9914 + DTLS interoperation with non-compliant servers
9915 + Don't call get_session_cb() without proposed session
9916 + Fix ia64 assembler code
9918 *Andy Polyakov, Steve Henson*
9920 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9922 * DTLS Handshake overhaul. There were longstanding issues with
9923 OpenSSL DTLS implementation, which were making it impossible for
9924 RFC 4347 compliant client to communicate with OpenSSL server.
9925 Unfortunately just fixing these incompatibilities would "cut off"
9926 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9927 server keeps tolerating non RFC compliant syntax. The opposite is
9928 not true, 0.9.8f client can not communicate with earlier server.
9929 This update even addresses CVE-2007-4995.
9933 * Changes to avoid need for function casts in OpenSSL: some compilers
9934 (gcc 4.2 and later) reject their use.
9935 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9938 * Add RFC4507 support to OpenSSL. This includes the corrections in
9939 RFC4507bis. The encrypted ticket format is an encrypted encoded
9940 SSL_SESSION structure, that way new session features are automatically
9943 If a client application caches session in an SSL_SESSION structure
9944 support is transparent because tickets are now stored in the encoded
9947 The SSL_CTX structure automatically generates keys for ticket
9948 protection in servers so again support should be possible
9949 with no application modification.
9951 If a client or server wishes to disable RFC4507 support then the option
9952 SSL_OP_NO_TICKET can be set.
9954 Add a TLS extension debugging callback to allow the contents of any client
9955 or server extensions to be examined.
9957 This work was sponsored by Google.
9961 * Add initial support for TLS extensions, specifically for the server_name
9962 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9963 have new members for a host name. The SSL data structure has an
9964 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9965 stored in that context to allow for session resumption, even after the
9966 SSL has been switched to a new SSL_CTX in reaction to a client's
9967 server_name extension.
9969 New functions (subject to change):
9971 SSL_get_servername()
9972 SSL_get_servername_type()
9975 New CTRL codes and macros (subject to change):
9977 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9978 - SSL_CTX_set_tlsext_servername_callback()
9979 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9980 - SSL_CTX_set_tlsext_servername_arg()
9981 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9983 openssl s_client has a new '-servername ...' option.
9985 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9986 '-key2 ...', '-servername_fatal' (subject to change). This allows
9987 testing the HostName extension for a specific single host name ('-cert'
9988 and '-key' remain fallbacks for handshakes without HostName
9989 negotiation). If the unrecognized_name alert has to be sent, this by
9990 default is a warning; it becomes fatal with the '-servername_fatal'
9993 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9995 * Add AES and SSE2 assembly language support to VC++ build.
9999 * Mitigate attack on final subtraction in Montgomery reduction.
10003 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10004 (which previously caused an internal error).
10008 * Squeeze another 10% out of IGE mode when in != out.
10012 * AES IGE mode speedup.
10014 *Dean Gaudet (Google)*
10016 * Add the Korean symmetric 128-bit cipher SEED (see
10017 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10018 add SEED ciphersuites from RFC 4162:
10020 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10021 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10022 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10023 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10025 To minimize changes between patchlevels in the OpenSSL 0.9.8
10026 series, SEED remains excluded from compilation unless OpenSSL
10027 is configured with 'enable-seed'.
10029 *KISA, Bodo Moeller*
10031 * Mitigate branch prediction attacks, which can be practical if a
10032 single processor is shared, allowing a spy process to extract
10033 information. For detailed background information, see
10034 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10035 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10036 and Necessary Software Countermeasures"). The core of the change
10037 are new versions BN_div_no_branch() and
10038 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10039 respectively, which are slower, but avoid the security-relevant
10040 conditional branches. These are automatically called by BN_div()
10041 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10042 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10043 remove a conditional branch.
10045 BN_FLG_CONSTTIME is the new name for the previous
10046 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10047 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10048 in the exponent causes BN_mod_exp_mont() to use the alternative
10049 implementation in BN_mod_exp_mont_consttime().) The old name
10050 remains as a deprecated alias.
10052 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10053 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10054 constant-time implementations for more than just exponentiation.
10055 Here too the old name is kept as a deprecated alias.
10057 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10058 the BN_BLINDING structure gets an independent copy of the
10059 modulus. This means that the previous `BIGNUM *m` argument to
10060 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10061 essentially becomes `const BIGNUM *m`, although we can't actually
10062 change this in the header file before 0.9.9. It allows
10063 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10064 enable BN_FLG_CONSTTIME.
10066 *Matthew D Wood (Intel Corp)*
10068 * In the SSL/TLS server implementation, be strict about session ID
10069 context matching (which matters if an application uses a single
10070 external cache for different purposes). Previously,
10071 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10072 set. This did ensure strict client verification, but meant that,
10073 with applications using a single external cache for quite
10074 different requirements, clients could circumvent ciphersuite
10075 restrictions for a given session ID context by starting a session
10076 in a different context.
10080 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10081 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10082 authentication-only ciphersuites.
10086 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10087 not complete and could lead to a possible single byte overflow
10088 ([CVE-2007-5135]) [Ben Laurie]
10090 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10092 * Since AES128 and AES256 (and similarly Camellia128 and
10093 Camellia256) share a single mask bit in the logic of
10094 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10095 kludge to work properly if AES128 is available and AES256 isn't
10096 (or if Camellia128 is available and Camellia256 isn't).
10100 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10101 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10102 When a point or a seed is encoded in a BIT STRING, we need to
10103 prevent the removal of trailing zero bits to get the proper DER
10104 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10105 of a NamedBitList, for which trailing 0 bits need to be removed.)
10109 * Have SSL/TLS server implementation tolerate "mismatched" record
10110 protocol version while receiving ClientHello even if the
10111 ClientHello is fragmented. (The server can't insist on the
10112 particular protocol version it has chosen before the ServerHello
10113 message has informed the client about his choice.)
10117 * Add RFC 3779 support.
10119 *Rob Austein for ARIN, Ben Laurie*
10121 * Load error codes if they are not already present instead of using a
10122 static variable. This allows them to be cleanly unloaded and reloaded.
10123 Improve header file function name parsing.
10127 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10128 or CAPABILITY handshake as required by RFCs.
10130 *Goetz Babin-Ebell*
10132 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10134 * Introduce limits to prevent malicious keys being able to
10135 cause a denial of service. ([CVE-2006-2940])
10137 *Steve Henson, Bodo Moeller*
10139 * Fix ASN.1 parsing of certain invalid structures that can result
10140 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10142 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10143 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10145 * Fix SSL client code which could crash if connecting to a
10146 malicious SSLv2 server. ([CVE-2006-4343])
10148 *Tavis Ormandy and Will Drewry, Google Security Team*
10150 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10151 match only those. Before that, "AES256-SHA" would be interpreted
10152 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10153 the same strength classification in 0.9.7h) as we currently only
10154 have a single AES bit in the ciphersuite description bitmap.
10155 That change, however, also applied to ciphersuite strings such as
10156 "RC4-MD5" that intentionally matched multiple ciphersuites --
10157 namely, SSL 2.0 ciphersuites in addition to the more common ones
10158 from SSL 3.0/TLS 1.0.
10160 So we change the selection algorithm again: Naming an explicit
10161 ciphersuite selects this one ciphersuite, and any other similar
10162 ciphersuite (same bitmap) from *other* protocol versions.
10163 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10164 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10166 Since SSL 2.0 does not have any ciphersuites for which the
10167 128/256 bit distinction would be relevant, this works for now.
10168 The proper fix will be to use different bits for AES128 and
10169 AES256, which would have avoided the problems from the beginning;
10170 however, bits are scarce, so we can only do this in a new release
10171 (not just a patchlevel) when we can change the SSL_CIPHER
10172 definition to split the single 'unsigned long mask' bitmap into
10173 multiple values to extend the available space.
10177 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10179 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10180 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10182 * Add AES IGE and biIGE modes.
10186 * Change the Unix randomness entropy gathering to use poll() when
10187 possible instead of select(), since the latter has some
10188 undesirable limitations.
10190 *Darryl Miles via Richard Levitte and Bodo Moeller*
10192 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10193 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10194 cannot be implicitly activated as part of, e.g., the "AES" alias.
10195 However, please upgrade to OpenSSL 0.9.9[-dev] for
10196 non-experimental use of the ECC ciphersuites to get TLS extension
10197 support, which is required for curve and point format negotiation
10198 to avoid potential handshake problems.
10202 * Disable rogue ciphersuites:
10204 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10205 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10206 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10208 The latter two were purportedly from
10209 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10212 Also deactivate the remaining ciphersuites from
10213 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10214 unofficial, and the ID has long expired.
10218 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10219 dual-core machines) and other potential thread-safety issues.
10223 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10224 versions), which is now available for royalty-free use
10225 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10226 Also, add Camellia TLS ciphersuites from RFC 4132.
10228 To minimize changes between patchlevels in the OpenSSL 0.9.8
10229 series, Camellia remains excluded from compilation unless OpenSSL
10230 is configured with 'enable-camellia'.
10234 * Disable the padding bug check when compression is in use. The padding
10235 bug check assumes the first packet is of even length, this is not
10236 necessarily true if compression is enabled and can result in false
10237 positives causing handshake failure. The actual bug test is ancient
10238 code so it is hoped that implementations will either have fixed it by
10239 now or any which still have the bug do not support compression.
10243 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10245 * When applying a cipher rule check to see if string match is an explicit
10246 cipher suite and only match that one cipher suite if it is.
10250 * Link in manifests for VC++ if needed.
10252 *Austin Ziegler <halostatue@gmail.com>*
10254 * Update support for ECC-based TLS ciphersuites according to
10255 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10256 TLS extensions, which are supported starting with the 0.9.9
10257 branch, not in the OpenSSL 0.9.8 branch).
10261 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10262 opaque EVP_CIPHER_CTX handling.
10266 * Fixes and enhancements to zlib compression code. We now only use
10267 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10268 to conform with the standards mentioned here:
10269 <http://www.zlib.net/DLL_FAQ.txt>
10270 Static zlib linking now works on Windows and the new --with-zlib-include
10271 --with-zlib-lib options to Configure can be used to supply the location
10272 of the headers and library. Gracefully handle case where zlib library
10277 * Several fixes and enhancements to the OID generation code. The old code
10278 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10279 handle numbers larger than ULONG_MAX, truncated printing and had a
10280 non standard OBJ_obj2txt() behaviour.
10284 * Add support for building of engines under engine/ as shared libraries
10285 under VC++ build system.
10289 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10290 Hopefully, we will not see any false combination of paths any more.
10294 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10296 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10297 (part of SSL_OP_ALL). This option used to disable the
10298 countermeasure against man-in-the-middle protocol-version
10299 rollback in the SSL 2.0 server implementation, which is a bad
10300 idea. ([CVE-2005-2969])
10302 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10303 for Information Security, National Institute of Advanced Industrial
10304 Science and Technology [AIST], Japan)*
10306 * Add two function to clear and return the verify parameter flags.
10310 * Keep cipherlists sorted in the source instead of sorting them at
10311 runtime, thus removing the need for a lock.
10315 * Avoid some small subgroup attacks in Diffie-Hellman.
10317 *Nick Mathewson and Ben Laurie*
10319 * Add functions for well-known primes.
10323 * Extended Windows CE support.
10325 *Satoshi Nakamura and Andy Polyakov*
10327 * Initialize SSL_METHOD structures at compile time instead of during
10328 runtime, thus removing the need for a lock.
10332 * Make PKCS7_decrypt() work even if no certificate is supplied by
10333 attempting to decrypt each encrypted key in turn. Add support to
10338 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10340 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10343 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10347 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10348 key into the same file any more.
10352 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10356 * Add -utf8 command line and config file option to 'ca'.
10358 *Stefan <stf@udoma.org*
10360 * Removed the macro des_crypt(), as it seems to conflict with some
10361 libraries. Use DES_crypt().
10365 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10366 involves renaming the source and generated shared-libs for
10367 both. The engines will accept the corrected or legacy ids
10368 ('ncipher' and '4758_cca' respectively) when binding. NB,
10369 this only applies when building 'shared'.
10371 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10373 * Add attribute functions to EVP_PKEY structure. Modify
10374 PKCS12_create() to recognize a CSP name attribute and
10375 use it. Make -CSP option work again in pkcs12 utility.
10379 * Add new functionality to the bn blinding code:
10380 - automatic re-creation of the BN_BLINDING parameters after
10381 a fixed number of uses (currently 32)
10382 - add new function for parameter creation
10383 - introduce flags to control the update behaviour of the
10384 BN_BLINDING parameters
10385 - hide BN_BLINDING structure
10386 Add a second BN_BLINDING slot to the RSA structure to improve
10387 performance when a single RSA object is shared among several
10392 * Add support for DTLS.
10394 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10396 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10397 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10401 * Remove buggy and incomplete DH cert support from
10402 ssl/ssl_rsa.c and ssl/s3_both.c
10406 * Use SHA-1 instead of MD5 as the default digest algorithm for
10407 the `apps/openssl` commands.
10411 * Compile clean with "-Wall -Wmissing-prototypes
10412 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10413 DEBUG_SAFESTACK must also be set.
10417 * Change ./Configure so that certain algorithms can be disabled by default.
10418 The new counterpiece to "no-xxx" is "enable-xxx".
10420 The patented RC5 and MDC2 algorithms will now be disabled unless
10421 "enable-rc5" and "enable-mdc2", respectively, are specified.
10423 (IDEA remains enabled despite being patented. This is because IDEA
10424 is frequently required for interoperability, and there is no license
10425 fee for non-commercial use. As before, "no-idea" can be used to
10426 avoid this algorithm.)
10430 * Add processing of proxy certificates (see RFC 3820). This work was
10431 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10432 EGEE (Enabling Grids for E-science in Europe).
10436 * RC4 performance overhaul on modern architectures/implementations, such
10437 as Intel P4, IA-64 and AMD64.
10441 * New utility extract-section.pl. This can be used specify an alternative
10442 section number in a pod file instead of having to treat each file as
10443 a separate case in Makefile. This can be done by adding two lines to the
10446 =for comment openssl_section:XXX
10448 The blank line is mandatory.
10452 * New arguments -certform, -keyform and -pass for s_client and s_server
10453 to allow alternative format key and certificate files and passphrase
10458 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10459 update associated structures and add various utility functions.
10461 Add new policy related verify parameters, include policy checking in
10462 standard verify code. Enhance 'smime' application with extra parameters
10463 to support policy checking and print out.
10467 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10468 Nehemiah processors. These extensions support AES encryption in hardware
10469 as well as RNG (though RNG support is currently disabled).
10471 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10473 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10477 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10479 *Andy Polyakov and a number of other people*
10481 * Improved PowerPC platform support. Most notably BIGNUM assembler
10482 implementation contributed by IBM.
10484 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10486 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10487 exponent rather than 'unsigned long'. There is a corresponding change to
10488 the new 'rsa_keygen' element of the RSA_METHOD structure.
10490 *Jelte Jansen, Geoff Thorpe*
10492 * Functionality for creating the initial serial number file is now
10493 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10495 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10496 number file to 1, which is bound to cause problems. To avoid
10497 the problems while respecting compatibility between different 0.9.7
10498 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10499 CA.pl for serial number initialization. With the new release 0.9.8,
10500 we can fix the problem directly in the 'ca' utility.)
10504 * Reduced header interdependencies by declaring more opaque objects in
10505 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10506 give fewer recursive includes, which could break lazy source code - so
10507 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10508 developers should define this symbol when building and using openssl to
10509 ensure they track the recommended behaviour, interfaces, [etc], but
10510 backwards-compatible behaviour prevails when this isn't defined.
10514 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10518 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10519 This will generate a random key of the appropriate length based on the
10520 cipher context. The EVP_CIPHER can provide its own random key generation
10521 routine to support keys of a specific form. This is used in the des and
10522 3des routines to generate a key of the correct parity. Update S/MIME
10523 code to use new functions and hence generate correct parity DES keys.
10524 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10525 valid (weak or incorrect parity).
10529 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10530 as looking them up. This is useful when the verified structure may contain
10531 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10532 present unless the new PKCS7_NO_CRL flag is asserted.
10536 * Extend ASN1 oid configuration module. It now additionally accepts the
10539 shortName = some long name, 1.2.3.4
10543 * Reimplemented the BN_CTX implementation. There is now no more static
10544 limitation on the number of variables it can handle nor the depth of the
10545 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10546 information can now expand as required, and rather than having a single
10547 static array of bignums, BN_CTX now uses a linked-list of such arrays
10548 allowing it to expand on demand whilst maintaining the usefulness of
10549 BN_CTX's "bundling".
10553 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10554 to allow all RSA operations to function using a single BN_CTX.
10558 * Preliminary support for certificate policy evaluation and checking. This
10559 is initially intended to pass the tests outlined in "Conformance Testing
10560 of Relying Party Client Certificate Path Processing Logic" v1.07.
10564 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10565 remained unused and not that useful. A variety of other little bignum
10566 tweaks and fixes have also been made continuing on from the audit (see
10571 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10572 associated ASN1, EVP and SSL functions and old ASN1 macros.
10576 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10577 and this should never fail. So the return value from the use of
10578 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10579 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10583 * BN_CTX_get() should return zero-valued bignums, providing the same
10584 initialised value as BN_new().
10586 *Geoff Thorpe, suggested by Ulf Möller*
10588 * Support for inhibitAnyPolicy certificate extension.
10592 * An audit of the BIGNUM code is underway, for which debugging code is
10593 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10594 is considered valid when processing BIGNUMs, and causes execution to
10595 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10596 further steps are taken to deliberately pollute unused data in BIGNUM
10597 structures to try and expose faulty code further on. For now, openssl will
10598 (in its default mode of operation) continue to tolerate the inconsistent
10599 forms that it has tolerated in the past, but authors and packagers should
10600 consider trying openssl and their own applications when compiled with
10601 these debugging symbols defined. It will help highlight potential bugs in
10602 their own code, and will improve the test coverage for OpenSSL itself. At
10603 some point, these tighter rules will become openssl's default to improve
10604 maintainability, though the assert()s and other overheads will remain only
10605 in debugging configurations. See bn.h for more details.
10607 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10609 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10610 that can only be obtained through BN_CTX_new() (which implicitly
10611 initialises it). The presence of this function only made it possible
10612 to overwrite an existing structure (and cause memory leaks).
10616 * Because of the callback-based approach for implementing LHASH as a
10617 template type, lh_insert() adds opaque objects to hash-tables and
10618 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10619 to clean up those corresponding objects before destroying the hash table
10620 (and losing the object pointers). So some over-zealous constifications in
10621 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10622 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10623 prototyped to have "const" restrictions on the object pointers they are
10624 given (and so aren't required to cast them away any more).
10628 * The tmdiff.h API was so ugly and minimal that our own timing utility
10629 (speed) prefers to use its own implementation. The two implementations
10630 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10631 its object type properly exposed (MS_TM) instead of casting to/from
10632 `char *`. This may still change yet if someone realises MS_TM and
10634 aren't necessarily the greatest nomenclatures - but this is what was used
10635 internally to the implementation so I've used that for now.
10639 * Ensure that deprecated functions do not get compiled when
10640 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10641 the self-tests were still using deprecated key-generation functions so
10642 these have been updated also.
10646 * Reorganise PKCS#7 code to separate the digest location functionality
10647 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10648 New function PKCS7_set_digest() to set the digest type for PKCS#7
10649 digestedData type. Add additional code to correctly generate the
10650 digestedData type and add support for this type in PKCS7 initialization
10655 * New function PKCS7_set0_type_other() this initializes a PKCS7
10656 structure of type "other".
10660 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10661 sure the loop does correctly stop and breaking ("division by zero")
10662 modulus operations are not performed. The (pre-generated) prime
10663 table crypto/bn/bn_prime.h was already correct, but it could not be
10664 re-generated on some platforms because of the "division by zero"
10665 situation in the script.
10667 *Ralf S. Engelschall*
10669 * Update support for ECC-based TLS ciphersuites according to
10670 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10671 SHA-1 now is only used for "small" curves (where the
10672 representation of a field element takes up to 24 bytes); for
10673 larger curves, the field element resulting from ECDH is directly
10674 used as premaster secret.
10676 *Douglas Stebila (Sun Microsystems Laboratories)*
10678 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10679 curve secp160r1 to the tests.
10681 *Douglas Stebila (Sun Microsystems Laboratories)*
10683 * Add the possibility to load symbols globally with DSO.
10685 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10687 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10688 control of the error stack.
10692 * Add support for STORE in ENGINE.
10696 * Add the STORE type. The intention is to provide a common interface
10697 to certificate and key stores, be they simple file-based stores, or
10698 HSM-type store, or LDAP stores, or...
10699 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10703 * Add a generic structure called OPENSSL_ITEM. This can be used to
10704 pass a list of arguments to any function as well as provide a way
10705 for a function to pass data back to the caller.
10709 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10710 works like BUF_strdup() but can be used to duplicate a portion of
10711 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10716 * Add the function sk_find_ex() which works like sk_find(), but will
10717 return an index to an element even if an exact match couldn't be
10718 found. The index is guaranteed to point at the element where the
10719 searched-for key would be inserted to preserve sorting order.
10723 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10724 takes an extra flags argument for optional functionality. Currently,
10725 the following flags are defined:
10727 OBJ_BSEARCH_VALUE_ON_NOMATCH
10728 This one gets OBJ_bsearch_ex() to return a pointer to the first
10729 element where the comparing function returns a negative or zero
10732 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10733 This one gets OBJ_bsearch_ex() to return a pointer to the first
10734 element where the comparing function returns zero. This is useful
10735 if there are more than one element where the comparing function
10740 * Make it possible to create self-signed certificates with 'openssl ca'
10741 in such a way that the self-signed certificate becomes part of the
10742 CA database and uses the same mechanisms for serial number generation
10743 as all other certificate signing. The new flag '-selfsign' enables
10744 this functionality. Adapt CA.sh and CA.pl.in.
10748 * Add functionality to check the public key of a certificate request
10749 against a given private. This is useful to check that a certificate
10750 request can be signed by that key (self-signing).
10754 * Make it possible to have multiple active certificates with the same
10755 subject in the CA index file. This is done only if the keyword
10756 'unique_subject' is set to 'no' in the main CA section (default
10757 if 'CA_default') of the configuration file. The value is saved
10758 with the database itself in a separate index attribute file,
10759 named like the index file with '.attr' appended to the name.
10763 * Generate multi-valued AVAs using '+' notation in config files for
10768 * Support for nameConstraints certificate extension.
10772 * Support for policyConstraints certificate extension.
10776 * Support for policyMappings certificate extension.
10780 * Make sure the default DSA_METHOD implementation only uses its
10781 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10782 and change its own handlers to be NULL so as to remove unnecessary
10783 indirection. This lets alternative implementations fallback to the
10784 default implementation more easily.
10788 * Support for directoryName in GeneralName related extensions
10793 * Make it possible to link applications using Makefile.shared.
10794 Make that possible even when linking against static libraries!
10798 * Support for single pass processing for S/MIME signing. This now
10799 means that S/MIME signing can be done from a pipe, in addition
10800 cleartext signing (multipart/signed type) is effectively streaming
10801 and the signed data does not need to be all held in memory.
10803 This is done with a new flag PKCS7_STREAM. When this flag is set
10804 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10805 is done after the data is output (and digests calculated) in
10806 SMIME_write_PKCS7().
10810 * Add full support for -rpath/-R, both in shared libraries and
10811 applications, at least on the platforms where it's known how
10816 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10817 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10818 will now compute a table of multiples of the generator that
10819 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10820 faster (notably in the case of a single point multiplication,
10821 scalar * generator).
10823 *Nils Larsch, Bodo Moeller*
10825 * IPv6 support for certificate extensions. The various extensions
10826 which use the IP:a.b.c.d can now take IPv6 addresses using the
10827 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10832 * Added an ENGINE that implements RSA by performing private key
10833 exponentiations with the GMP library. The conversions to and from
10834 GMP's mpz_t format aren't optimised nor are any montgomery forms
10835 cached, and on x86 it appears OpenSSL's own performance has caught up.
10836 However there are likely to be other architectures where GMP could
10837 provide a boost. This ENGINE is not built in by default, but it can be
10838 specified at Configure time and should be accompanied by the necessary
10839 linker additions, eg;
10840 ./config -DOPENSSL_USE_GMP -lgmp
10844 * "openssl engine" will not display ENGINE/DSO load failure errors when
10845 testing availability of engines with "-t" - the old behaviour is
10846 produced by increasing the feature's verbosity with "-tt".
10850 * ECDSA routines: under certain error conditions uninitialized BN objects
10851 could be freed. Solution: make sure initialization is performed early
10852 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10857 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10858 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10859 software implementations. For DSA and DH, parameter generation can
10860 also be overridden by providing the appropriate method callbacks.
10864 * Change the "progress" mechanism used in key-generation and
10865 primality testing to functions that take a new BN_GENCB pointer in
10866 place of callback/argument pairs. The new API functions have `_ex`
10867 postfixes and the older functions are reimplemented as wrappers for
10868 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10869 declarations of the old functions to help (graceful) attempts to
10870 migrate to the new functions. Also, the new key-generation API
10871 functions operate on a caller-supplied key-structure and return
10872 success/failure rather than returning a key or NULL - this is to
10873 help make "keygen" another member function of RSA_METHOD etc.
10875 Example for using the new callback interface:
10877 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10878 void *my_arg = ...;
10881 BN_GENCB_set(&my_cb, my_callback, my_arg);
10883 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10884 /* For the meaning of a, b in calls to my_callback(), see the
10885 * documentation of the function that calls the callback.
10886 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10887 * my_callback should return 1 if it wants BN_is_prime_ex()
10888 * to continue, or 0 to stop.
10893 * Change the ZLIB compression method to be stateful, and make it
10894 available to TLS with the number defined in
10895 draft-ietf-tls-compression-04.txt.
10899 * Add the ASN.1 structures and functions for CertificatePair, which
10900 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10902 CertificatePair ::= SEQUENCE {
10903 forward [0] Certificate OPTIONAL,
10904 reverse [1] Certificate OPTIONAL,
10905 -- at least one of the pair shall be present -- }
10907 Also implement the PEM functions to read and write certificate
10908 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10910 This needed to be defined, mostly for the sake of the LDAP
10911 attribute crossCertificatePair, but may prove useful elsewhere as
10916 * Make it possible to inhibit symlinking of shared libraries in
10917 Makefile.shared, for Cygwin's sake.
10921 * Extend the BIGNUM API by creating a function
10922 void BN_set_negative(BIGNUM *a, int neg);
10923 and a macro that behave like
10924 int BN_is_negative(const BIGNUM *a);
10926 to avoid the need to access 'a->neg' directly in applications.
10930 * Implement fast modular reduction for pseudo-Mersenne primes
10931 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10932 EC_GROUP_new_curve_GFp() will now automatically use this
10935 *Nils Larsch <nla@trustcenter.de>*
10937 * Add new lock type (CRYPTO_LOCK_BN).
10941 * Change the ENGINE framework to automatically load engines
10942 dynamically from specific directories unless they could be
10943 found to already be built in or loaded. Move all the
10944 current engines except for the cryptodev one to a new
10945 directory engines/.
10946 The engines in engines/ are built as shared libraries if
10947 the "shared" options was given to ./Configure or ./config.
10948 Otherwise, they are inserted in libcrypto.a.
10949 /usr/local/ssl/engines is the default directory for dynamic
10950 engines, but that can be overridden at configure time through
10951 the usual use of --prefix and/or --openssldir, and at run
10952 time with the environment variable OPENSSL_ENGINES.
10954 *Geoff Thorpe and Richard Levitte*
10956 * Add Makefile.shared, a helper makefile to build shared
10957 libraries. Adapt Makefile.org.
10961 * Add version info to Win32 DLLs.
10963 *Peter 'Luna' Runestig" <peter@runestig.com>*
10965 * Add new 'medium level' PKCS#12 API. Certificates and keys
10966 can be added using this API to created arbitrary PKCS#12
10967 files while avoiding the low level API.
10969 New options to PKCS12_create(), key or cert can be NULL and
10970 will then be omitted from the output file. The encryption
10971 algorithm NIDs can be set to -1 for no encryption, the mac
10972 iteration count can be set to 0 to omit the mac.
10974 Enhance pkcs12 utility by making the -nokeys and -nocerts
10975 options work when creating a PKCS#12 file. New option -nomac
10976 to omit the mac, NONE can be set for an encryption algorithm.
10977 New code is modified to use the enhanced PKCS12_create()
10978 instead of the low level API.
10982 * Extend ASN1 encoder to support indefinite length constructed
10983 encoding. This can output sequences tags and octet strings in
10984 this form. Modify pk7_asn1.c to support indefinite length
10985 encoding. This is experimental and needs additional code to
10986 be useful, such as an ASN1 bio and some enhanced streaming
10989 Extend template encode functionality so that tagging is passed
10990 down to the template encoder.
10994 * Let 'openssl req' fail if an argument to '-newkey' is not
10995 recognized instead of using RSA as a default.
10999 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11000 As these are not official, they are not included in "ALL";
11001 the "ECCdraft" ciphersuite group alias can be used to select them.
11003 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11005 * Add ECDH engine support.
11007 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11009 * Add ECDH in new directory crypto/ecdh/.
11011 *Douglas Stebila (Sun Microsystems Laboratories)*
11013 * Let BN_rand_range() abort with an error after 100 iterations
11014 without success (which indicates a broken PRNG).
11018 * Change BN_mod_sqrt() so that it verifies that the input value
11019 is really the square of the return value. (Previously,
11020 BN_mod_sqrt would show GIGO behaviour.)
11024 * Add named elliptic curves over binary fields from X9.62, SECG,
11025 and WAP/WTLS; add OIDs that were still missing.
11027 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11029 * Extend the EC library for elliptic curves over binary fields
11030 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11033 EC_GF2m_simple_method
11037 EC_GROUP_new_curve_GF2m
11038 EC_GROUP_set_curve_GF2m
11039 EC_GROUP_get_curve_GF2m
11040 EC_POINT_set_affine_coordinates_GF2m
11041 EC_POINT_get_affine_coordinates_GF2m
11042 EC_POINT_set_compressed_coordinates_GF2m
11044 Point compression for binary fields is disabled by default for
11045 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11048 As binary polynomials are represented as BIGNUMs, various members
11049 of the EC_GROUP and EC_POINT data structures can be shared
11050 between the implementations for prime fields and binary fields;
11051 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11052 are essentially identical to their `..._GFp` counterparts.
11053 (For simplicity, the `..._GFp` prefix has been dropped from
11054 various internal method names.)
11056 An internal 'field_div' method (similar to 'field_mul' and
11057 'field_sqr') has been added; this is used only for binary fields.
11059 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11061 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11062 through methods ('mul', 'precompute_mult').
11064 The generic implementations (now internally called 'ec_wNAF_mul'
11065 and 'ec_wNAF_precomputed_mult') remain the default if these
11066 methods are undefined.
11068 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11070 * New function EC_GROUP_get_degree, which is defined through
11071 EC_METHOD. For curves over prime fields, this returns the bit
11072 length of the modulus.
11074 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11076 * New functions EC_GROUP_dup, EC_POINT_dup.
11077 (These simply call ..._new and ..._copy).
11079 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11081 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11082 Polynomials are represented as BIGNUMs (where the sign bit is not
11083 used) in the following functions [macros]:
11086 BN_GF2m_sub [= BN_GF2m_add]
11087 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11088 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11089 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11091 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11092 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11093 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11094 BN_GF2m_cmp [= BN_ucmp]
11096 (Note that only the 'mod' functions are actually for fields GF(2^m).
11097 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11099 For some functions, an the irreducible polynomial defining a
11100 field can be given as an 'unsigned int[]' with strictly
11101 decreasing elements giving the indices of those bits that are set;
11102 i.e., p[] represents the polynomial
11103 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11105 p[0] > p[1] > ... > p[k] = 0.
11106 This applies to the following functions:
11109 BN_GF2m_mod_mul_arr
11110 BN_GF2m_mod_sqr_arr
11111 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11112 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11113 BN_GF2m_mod_exp_arr
11114 BN_GF2m_mod_sqrt_arr
11115 BN_GF2m_mod_solve_quad_arr
11119 Conversion can be performed by the following functions:
11124 bntest.c has additional tests for binary polynomial arithmetic.
11126 Two implementations for BN_GF2m_mod_div() are available.
11127 The default algorithm simply uses BN_GF2m_mod_inv() and
11128 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11129 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11130 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11132 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11134 * Add new error code 'ERR_R_DISABLED' that can be used when some
11135 functionality is disabled at compile-time.
11137 *Douglas Stebila <douglas.stebila@sun.com>*
11139 * Change default behaviour of 'openssl asn1parse' so that more
11140 information is visible when viewing, e.g., a certificate:
11142 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11143 mode the content of non-printable OCTET STRINGs is output in a
11144 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11145 avoid the appearance of a printable string.
11147 *Nils Larsch <nla@trustcenter.de>*
11149 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11151 EC_GROUP_set_asn1_flag()
11152 EC_GROUP_get_asn1_flag()
11153 EC_GROUP_set_point_conversion_form()
11154 EC_GROUP_get_point_conversion_form()
11155 These control ASN1 encoding details:
11156 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11157 has been set to OPENSSL_EC_NAMED_CURVE.
11158 - Points are encoded in uncompressed form by default; options for
11159 asn1_for are as for point2oct, namely
11160 POINT_CONVERSION_COMPRESSED
11161 POINT_CONVERSION_UNCOMPRESSED
11162 POINT_CONVERSION_HYBRID
11164 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11166 EC_GROUP_set_seed()
11167 EC_GROUP_get0_seed()
11168 EC_GROUP_get_seed_len()
11169 This is used only for ASN1 purposes (so far).
11171 *Nils Larsch <nla@trustcenter.de>*
11173 * Add 'field_type' member to EC_METHOD, which holds the NID
11174 of the appropriate field type OID. The new function
11175 EC_METHOD_get_field_type() returns this value.
11177 *Nils Larsch <nla@trustcenter.de>*
11180 EC_POINT_point2bn()
11181 EC_POINT_bn2point()
11182 EC_POINT_point2hex()
11183 EC_POINT_hex2point()
11184 providing useful interfaces to EC_POINT_point2oct() and
11185 EC_POINT_oct2point().
11187 *Nils Larsch <nla@trustcenter.de>*
11189 * Change internals of the EC library so that the functions
11190 EC_GROUP_set_generator()
11191 EC_GROUP_get_generator()
11192 EC_GROUP_get_order()
11193 EC_GROUP_get_cofactor()
11194 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11195 to methods, which would lead to unnecessary code duplication when
11196 adding different types of curves.
11198 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11200 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11201 arithmetic, and such that modified wNAFs are generated
11202 (which avoid length expansion in many cases).
11206 * Add a function EC_GROUP_check_discriminant() (defined via
11207 EC_METHOD) that verifies that the curve discriminant is non-zero.
11209 Add a function EC_GROUP_check() that makes some sanity tests
11210 on a EC_GROUP, its generator and order. This includes
11211 EC_GROUP_check_discriminant().
11213 *Nils Larsch <nla@trustcenter.de>*
11215 * Add ECDSA in new directory crypto/ecdsa/.
11217 Add applications 'openssl ecparam' and 'openssl ecdsa'
11218 (these are based on 'openssl dsaparam' and 'openssl dsa').
11220 ECDSA support is also included in various other files across the
11221 library. Most notably,
11222 - 'openssl req' now has a '-newkey ecdsa:file' option;
11223 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11224 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11225 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11226 them suitable for ECDSA where domain parameters must be
11227 extracted before the specific public key;
11228 - ECDSA engine support has been added.
11230 *Nils Larsch <nla@trustcenter.de>*
11232 * Include some named elliptic curves, and add OIDs from X9.62,
11233 SECG, and WAP/WTLS. Each curve can be obtained from the new
11235 EC_GROUP_new_by_curve_name(),
11236 and the list of available named curves can be obtained with
11237 EC_get_builtin_curves().
11238 Also add a 'curve_name' member to EC_GROUP objects, which can be
11240 EC_GROUP_set_curve_name()
11241 EC_GROUP_get_curve_name()
11243 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11245 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11246 was actually never needed) and in BN_mul(). The removal in BN_mul()
11247 required a small change in bn_mul_part_recursive() and the addition
11248 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11249 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11250 bn_sub_words() and bn_add_words() except they take arrays with
11255 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11257 * Cleanse PEM buffers before freeing them since they may contain
11260 *Benjamin Bennett <ben@psc.edu>*
11262 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11263 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11264 authentication-only ciphersuites.
11268 * Since AES128 and AES256 share a single mask bit in the logic of
11269 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11270 kludge to work properly if AES128 is available and AES256 isn't.
11274 * Expand security boundary to match 1.1.1 module.
11278 * Remove redundant features: hash file source, editing of test vectors
11279 modify fipsld to use external fips_premain.c signature.
11283 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11284 run algorithm test programs.
11288 * Make algorithm test programs more tolerant of whitespace.
11292 * Have SSL/TLS server implementation tolerate "mismatched" record
11293 protocol version while receiving ClientHello even if the
11294 ClientHello is fragmented. (The server can't insist on the
11295 particular protocol version it has chosen before the ServerHello
11296 message has informed the client about his choice.)
11300 * Load error codes if they are not already present instead of using a
11301 static variable. This allows them to be cleanly unloaded and reloaded.
11305 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11307 * Introduce limits to prevent malicious keys being able to
11308 cause a denial of service. ([CVE-2006-2940])
11310 *Steve Henson, Bodo Moeller*
11312 * Fix ASN.1 parsing of certain invalid structures that can result
11313 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11315 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11316 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11318 * Fix SSL client code which could crash if connecting to a
11319 malicious SSLv2 server. ([CVE-2006-4343])
11321 *Tavis Ormandy and Will Drewry, Google Security Team*
11323 * Change ciphersuite string processing so that an explicit
11324 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11325 will no longer include "AES128-SHA"), and any other similar
11326 ciphersuite (same bitmap) from *other* protocol versions (so that
11327 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11328 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11329 changes from 0.9.8b and 0.9.8d.
11333 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11335 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11336 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11338 * Change the Unix randomness entropy gathering to use poll() when
11339 possible instead of select(), since the latter has some
11340 undesirable limitations.
11342 *Darryl Miles via Richard Levitte and Bodo Moeller*
11344 * Disable rogue ciphersuites:
11346 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11347 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11348 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11350 The latter two were purportedly from
11351 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11354 Also deactivate the remaining ciphersuites from
11355 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11356 unofficial, and the ID has long expired.
11360 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11361 dual-core machines) and other potential thread-safety issues.
11365 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11367 * Adapt fipsld and the build system to link against the validated FIPS
11368 module in FIPS mode.
11372 * Fixes for VC++ 2005 build under Windows.
11376 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11377 from a Windows bash shell such as MSYS. It is autodetected from the
11378 "config" script when run from a VC++ environment. Modify standard VC++
11379 build to use fipscanister.o from the GNU make build.
11383 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11385 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11386 The value now differs depending on if you build for FIPS or not.
11387 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11388 safely run with a non-FIPSed libcrypto, as it may crash because of
11389 the difference induced by this change.
11393 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11395 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11396 (part of SSL_OP_ALL). This option used to disable the
11397 countermeasure against man-in-the-middle protocol-version
11398 rollback in the SSL 2.0 server implementation, which is a bad
11399 idea. ([CVE-2005-2969])
11401 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11402 for Information Security, National Institute of Advanced Industrial
11403 Science and Technology [AIST, Japan)]*
11405 * Minimal support for X9.31 signatures and PSS padding modes. This is
11406 mainly for FIPS compliance and not fully integrated at this stage.
11410 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11411 the exponentiation using a fixed-length exponent. (Otherwise,
11412 the information leaked through timing could expose the secret key
11413 after many signatures; cf. Bleichenbacher's attack on DSA with
11418 * Make a new fixed-window mod_exp implementation the default for
11419 RSA, DSA, and DH private-key operations so that the sequence of
11420 squares and multiplies and the memory access pattern are
11421 independent of the particular secret key. This will mitigate
11422 cache-timing and potential related attacks.
11424 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11425 and this is automatically used by BN_mod_exp_mont() if the new flag
11426 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11427 will use this BN flag for private exponents unless the flag
11428 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11429 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11431 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11433 * Change the client implementation for SSLv23_method() and
11434 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11435 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11436 (Previously, the SSL 2.0 backwards compatible Client Hello
11437 message format would be used even with SSL_OP_NO_SSLv2.)
11441 * Add support for smime-type MIME parameter in S/MIME messages which some
11446 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11447 a threadsafe manner. Modify rsa code to use new function and add calls
11448 to dsa and dh code (which had race conditions before).
11452 * Include the fixed error library code in the C error file definitions
11453 instead of fixing them up at runtime. This keeps the error code
11454 structures constant.
11458 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11460 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11463 * Fixes for newer kerberos headers. NB: the casts are needed because
11464 the 'length' field is signed on one version and unsigned on another
11465 with no (?) obvious way to tell the difference, without these VC++
11466 complains. Also the "definition" of FAR (blank) is no longer included
11467 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11468 some needed definitions.
11472 * Undo Cygwin change.
11476 * Added support for proxy certificates according to RFC 3820.
11477 Because they may be a security thread to unaware applications,
11478 they must be explicitly allowed in run-time. See
11479 docs/HOWTO/proxy_certificates.txt for further information.
11483 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11485 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11486 server and client random values. Previously
11487 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11488 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11490 This change has negligible security impact because:
11492 1. Server and client random values still have 24 bytes of pseudo random
11495 2. Server and client random values are sent in the clear in the initial
11498 3. The master secret is derived using the premaster secret (48 bytes in
11499 size for static RSA ciphersuites) as well as client server and random
11502 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11505 *Stephen Henson, reported by UK NISCC*
11507 * Use Windows randomness collection on Cygwin.
11511 * Fix hang in EGD/PRNGD query when communication socket is closed
11512 prematurely by EGD/PRNGD.
11514 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11516 * Prompt for pass phrases when appropriate for PKCS12 input format.
11520 * Back-port of selected performance improvements from development
11521 branch, as well as improved support for PowerPC platforms.
11525 * Add lots of checks for memory allocation failure, error codes to indicate
11526 failure and freeing up memory if a failure occurs.
11528 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11530 * Add new -passin argument to dgst.
11534 * Perform some character comparisons of different types in X509_NAME_cmp:
11535 this is needed for some certificates that re-encode DNs into UTF8Strings
11536 (in violation of RFC3280) and can't or won't issue name rollover
11541 * Make an explicit check during certificate validation to see that
11542 the CA setting in each certificate on the chain is correct. As a
11543 side effect always do the following basic checks on extensions,
11544 not just when there's an associated purpose to the check:
11546 - if there is an unhandled critical extension (unless the user
11547 has chosen to ignore this fault)
11548 - if the path length has been exceeded (if one is set at all)
11549 - that certain extensions fit the associated purpose (if one has
11554 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11556 * Avoid a race condition when CRLs are checked in a multi threaded
11557 environment. This would happen due to the reordering of the revoked
11558 entries during signature checking and serial number lookup. Now the
11559 encoding is cached and the serial number sort performed under a lock.
11560 Add new STACK function sk_is_sorted().
11564 * Add Delta CRL to the extension code.
11568 * Various fixes to s3_pkt.c so alerts are sent properly.
11570 *David Holmes <d.holmes@f5.com>*
11572 * Reduce the chances of duplicate issuer name and serial numbers (in
11573 violation of RFC3280) using the OpenSSL certificate creation utilities.
11574 This is done by creating a random 64 bit value for the initial serial
11575 number when a serial number file is created or when a self signed
11576 certificate is created using 'openssl req -x509'. The initial serial
11577 number file is created using 'openssl x509 -next_serial' in CA.pl
11578 rather than being initialized to 1.
11582 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11584 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11585 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11587 *Joe Orton, Steve Henson*
11589 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11592 *Joe Orton, Steve Henson*
11594 * Make it possible to have multiple active certificates with the same
11595 subject in the CA index file. This is done only if the keyword
11596 'unique_subject' is set to 'no' in the main CA section (default
11597 if 'CA_default') of the configuration file. The value is saved
11598 with the database itself in a separate index attribute file,
11599 named like the index file with '.attr' appended to the name.
11603 * X509 verify fixes. Disable broken certificate workarounds when
11604 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11605 keyUsage extension present. Don't accept CRLs with unhandled critical
11606 extensions: since verify currently doesn't process CRL extensions this
11607 rejects a CRL with *any* critical extensions. Add new verify error codes
11612 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11613 A clarification of RFC2560 will require the use of OCTET STRINGs and
11614 some implementations cannot handle the current raw format. Since OpenSSL
11615 copies and compares OCSP nonces as opaque blobs without any attempt at
11616 parsing them this should not create any compatibility issues.
11620 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11621 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11622 this HMAC (and other) operations are several times slower than OpenSSL
11627 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11629 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11631 * Use the correct content when signing type "other".
11635 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11637 * Fix various bugs revealed by running the NISCC test suite:
11639 Stop out of bounds reads in the ASN1 code when presented with
11640 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11642 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11644 If verify callback ignores invalid public key errors don't try to check
11645 certificate signature with the NULL public key.
11649 * New -ignore_err option in ocsp application to stop the server
11650 exiting on the first error in a request.
11654 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11655 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11660 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11661 extra data after the compression methods not only for TLS 1.0
11662 but also for SSL 3.0 (as required by the specification).
11664 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11666 * Change X509_certificate_type() to mark the key as exported/exportable
11667 when it's 512 *bits* long, not 512 bytes.
11671 * Change AES_cbc_encrypt() so it outputs exact multiple of
11672 blocks during encryption.
11676 * Various fixes to base64 BIO and non blocking I/O. On write
11677 flushes were not handled properly if the BIO retried. On read
11678 data was not being buffered properly and had various logic bugs.
11679 This also affects blocking I/O when the data being decoded is a
11684 * Various S/MIME bugfixes and compatibility changes:
11685 output correct application/pkcs7 MIME type if
11686 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11687 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11688 of files as .eml work). Correctly handle very long lines in MIME
11693 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11695 * Countermeasure against the Klima-Pokorny-Rosa extension of
11696 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11697 a protocol version number mismatch like a decryption error
11698 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11702 * Turn on RSA blinding by default in the default implementation
11703 to avoid a timing attack. Applications that don't want it can call
11704 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11705 They would be ill-advised to do so in most cases.
11707 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11709 * Change RSA blinding code so that it works when the PRNG is not
11710 seeded (in this case, the secret RSA exponent is abused as
11711 an unpredictable seed -- if it is not unpredictable, there
11712 is no point in blinding anyway). Make RSA blinding thread-safe
11713 by remembering the creator's thread ID in rsa->blinding and
11714 having all other threads use local one-time blinding factors
11715 (this requires more computation than sharing rsa->blinding, but
11716 avoids excessive locking; and if an RSA object is not shared
11717 between threads, blinding will still be very fast).
11721 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11722 ENGINE as defaults for all supported algorithms irrespective of
11723 the 'flags' parameter. 'flags' is now honoured, so applications
11724 should make sure they are passing it correctly.
11728 * Target "mingw" now allows native Windows code to be generated in
11729 the Cygwin environment as well as with the MinGW compiler.
11733 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11735 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11736 via timing by performing a MAC computation even if incorrect
11737 block cipher padding has been found. This is a countermeasure
11738 against active attacks where the attacker has to distinguish
11739 between bad padding and a MAC verification error. ([CVE-2003-0078])
11741 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11742 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11743 Martin Vuagnoux (EPFL, Ilion)*
11745 * Make the no-err option work as intended. The intention with no-err
11746 is not to have the whole error stack handling routines removed from
11747 libcrypto, it's only intended to remove all the function name and
11748 reason texts, thereby removing some of the footprint that may not
11749 be interesting if those errors aren't displayed anyway.
11751 NOTE: it's still possible for any application or module to have its
11752 own set of error texts inserted. The routines are there, just not
11753 used by default when no-err is given.
11757 * Add support for FreeBSD on IA64.
11759 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11761 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11762 Kerberos function mit_des_cbc_cksum(). Before this change,
11763 the value returned by DES_cbc_cksum() was like the one from
11764 mit_des_cbc_cksum(), except the bytes were swapped.
11766 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11768 * Allow an application to disable the automatic SSL chain building.
11769 Before this a rather primitive chain build was always performed in
11770 ssl3_output_cert_chain(): an application had no way to send the
11771 correct chain if the automatic operation produced an incorrect result.
11773 Now the chain builder is disabled if either:
11775 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11777 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11779 The reasoning behind this is that an application would not want the
11780 auto chain building to take place if extra chain certificates are
11781 present and it might also want a means of sending no additional
11782 certificates (for example the chain has two certificates and the
11787 * Add the possibility to build without the ENGINE framework.
11789 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11791 * Under Win32 gmtime() can return NULL: check return value in
11792 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11796 * DSA routines: under certain error conditions uninitialized BN objects
11797 could be freed. Solution: make sure initialization is performed early
11798 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11799 Nils Larsch <nla@trustcenter.de> via PR#459)
11803 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11804 checked on reconnect on the client side, therefore session resumption
11805 could still fail with a "ssl session id is different" error. This
11806 behaviour is masked when SSL_OP_ALL is used due to
11807 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11808 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11809 followup to PR #377.
11813 * IA-32 assembler support enhancements: unified ELF targets, support
11814 for SCO/Caldera platforms, fix for Cygwin shared build.
11818 * Add support for FreeBSD on sparc64. As a consequence, support for
11819 FreeBSD on non-x86 processors is separate from x86 processors on
11820 the config script, much like the NetBSD support.
11822 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11824 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11826 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11829 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11830 code (06) was taken as the first octet of the session ID and the last
11831 octet was ignored consequently. As a result SSLv2 client side session
11832 caching could not have worked due to the session ID mismatch between
11834 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11839 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11840 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11845 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11846 seems that in spite of existing for more than a year, many application
11847 author have done nothing to provide the necessary callbacks, which
11848 means that this particular engine will not work properly anywhere.
11849 This is a very unfortunate situation which forces us, in the name
11850 of usability, to give the hw_ncipher.c a static lock, which is part
11852 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11853 appear in 0.9.8 or later. We EXPECT application authors to have
11854 dealt properly with this when 0.9.8 is released (unless we actually
11855 make such changes in the libcrypto locking code that changes will
11856 have to be made anyway).
11860 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11861 octets have been read, EOF or an error occurs. Without this change
11862 some truncated ASN1 structures will not produce an error.
11866 * Disable Heimdal support, since it hasn't been fully implemented.
11867 Still give the possibility to force the use of Heimdal, but with
11868 warnings and a request that patches get sent to openssl-dev.
11872 * Add the VC-CE target, introduce the WINCE sysname, and add
11873 INSTALL.WCE and appropriate conditionals to make it build.
11875 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11877 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11878 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11879 edit numbers of the version.
11881 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11883 * Introduce safe string copy and catenation functions
11884 (BUF_strlcpy() and BUF_strlcat()).
11886 *Ben Laurie (CHATS) and Richard Levitte*
11888 * Avoid using fixed-size buffers for one-line DNs.
11890 *Ben Laurie (CHATS)*
11892 * Add BUF_MEM_grow_clean() to avoid information leakage when
11893 resizing buffers containing secrets, and use where appropriate.
11895 *Ben Laurie (CHATS)*
11897 * Avoid using fixed size buffers for configuration file location.
11899 *Ben Laurie (CHATS)*
11901 * Avoid filename truncation for various CA files.
11903 *Ben Laurie (CHATS)*
11905 * Use sizeof in preference to magic numbers.
11907 *Ben Laurie (CHATS)*
11909 * Avoid filename truncation in cert requests.
11911 *Ben Laurie (CHATS)*
11913 * Add assertions to check for (supposedly impossible) buffer
11916 *Ben Laurie (CHATS)*
11918 * Don't cache truncated DNS entries in the local cache (this could
11919 potentially lead to a spoofing attack).
11921 *Ben Laurie (CHATS)*
11923 * Fix various buffers to be large enough for hex/decimal
11924 representations in a platform independent manner.
11926 *Ben Laurie (CHATS)*
11928 * Add CRYPTO_realloc_clean() to avoid information leakage when
11929 resizing buffers containing secrets, and use where appropriate.
11931 *Ben Laurie (CHATS)*
11933 * Add BIO_indent() to avoid much slightly worrying code to do
11936 *Ben Laurie (CHATS)*
11938 * Convert sprintf()/BIO_puts() to BIO_printf().
11940 *Ben Laurie (CHATS)*
11942 * buffer_gets() could terminate with the buffer only half
11945 *Ben Laurie (CHATS)*
11947 * Add assertions to prevent user-supplied crypto functions from
11948 overflowing internal buffers by having large block sizes, etc.
11950 *Ben Laurie (CHATS)*
11952 * New OPENSSL_assert() macro (similar to assert(), but enabled
11955 *Ben Laurie (CHATS)*
11957 * Eliminate unused copy of key in RC4.
11959 *Ben Laurie (CHATS)*
11961 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11963 *Ben Laurie (CHATS)*
11965 * Fix off-by-one error in EGD path.
11967 *Ben Laurie (CHATS)*
11969 * If RANDFILE path is too long, ignore instead of truncating.
11971 *Ben Laurie (CHATS)*
11973 * Eliminate unused and incorrectly sized X.509 structure
11976 *Ben Laurie (CHATS)*
11978 * Eliminate unused and dangerous function knumber().
11980 *Ben Laurie (CHATS)*
11982 * Eliminate unused and dangerous structure, KSSL_ERR.
11984 *Ben Laurie (CHATS)*
11986 * Protect against overlong session ID context length in an encoded
11987 session object. Since these are local, this does not appear to be
11990 *Ben Laurie (CHATS)*
11992 * Change from security patch (see 0.9.6e below) that did not affect
11993 the 0.9.6 release series:
11995 Remote buffer overflow in SSL3 protocol - an attacker could
11996 supply an oversized master key in Kerberos-enabled versions.
11999 *Ben Laurie (CHATS)*
12001 * Change the SSL kerb5 codes to match RFC 2712.
12005 * Make -nameopt work fully for req and add -reqopt switch.
12007 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12009 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12011 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12013 * Make sure tests can be performed even if the corresponding algorithms
12014 have been removed entirely. This was also the last step to make
12015 OpenSSL compilable with DJGPP under all reasonable conditions.
12017 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12019 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12020 to allow version independent disabling of normally unselected ciphers,
12021 which may be activated as a side-effect of selecting a single cipher.
12023 (E.g., cipher list string "RSA" enables ciphersuites that are left
12024 out of "ALL" because they do not provide symmetric encryption.
12025 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12027 *Lutz Jaenicke, Bodo Moeller*
12029 * Add appropriate support for separate platform-dependent build
12030 directories. The recommended way to make a platform-dependent
12031 build directory is the following (tested on Linux), maybe with
12034 # Place yourself outside of the OpenSSL source tree. In
12035 # this example, the environment variable OPENSSL_SOURCE
12036 # is assumed to contain the absolute OpenSSL source directory.
12037 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12038 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12039 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12040 mkdir -p `dirname $F`
12041 ln -s $OPENSSL_SOURCE/$F $F
12044 To be absolutely sure not to disturb the source tree, a "make clean"
12045 is a good thing. If it isn't successful, don't worry about it,
12046 it probably means the source directory is very clean.
12050 * Make sure any ENGINE control commands make local copies of string
12051 pointers passed to them whenever necessary. Otherwise it is possible
12052 the caller may have overwritten (or deallocated) the original string
12053 data when a later ENGINE operation tries to use the stored values.
12055 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12057 * Improve diagnostics in file reading and command-line digests.
12059 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12061 * Add AES modes CFB and OFB to the object database. Correct an
12062 error in AES-CFB decryption.
12066 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12067 allows existing EVP_CIPHER_CTX structures to be reused after
12068 calling `EVP_*Final()`. This behaviour is used by encryption
12069 BIOs and some applications. This has the side effect that
12070 applications must explicitly clean up cipher contexts with
12071 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12075 * Check the values of dna and dnb in bn_mul_recursive before calling
12076 bn_mul_comba (a non zero value means the a or b arrays do not contain
12077 n2 elements) and fallback to bn_mul_normal if either is not zero.
12081 * Fix escaping of non-ASCII characters when using the -subj option
12082 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12086 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12087 form for "surname", serialNumber has no short form.
12088 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12089 therefore remove "mail" short name for "internet 7".
12090 The OID for unique identifiers in X509 certificates is
12091 x500UniqueIdentifier, not uniqueIdentifier.
12092 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12096 * Add an "init" command to the ENGINE config module and auto initialize
12097 ENGINEs. Without any "init" command the ENGINE will be initialized
12098 after all ctrl commands have been executed on it. If init=1 the
12099 ENGINE is initialized at that point (ctrls before that point are run
12100 on the uninitialized ENGINE and after on the initialized one). If
12101 init=0 then the ENGINE will not be initialized at all.
12105 * Fix the 'app_verify_callback' interface so that the user-defined
12106 argument is actually passed to the callback: In the
12107 SSL_CTX_set_cert_verify_callback() prototype, the callback
12108 declaration has been changed from
12111 int (*cb)(X509_STORE_CTX *,void *);
12112 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12113 i=s->ctx->app_verify_callback(&ctx)
12114 has been changed into
12115 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12117 To update applications using SSL_CTX_set_cert_verify_callback(),
12118 a dummy argument can be added to their callback functions.
12120 *D. K. Smetters <smetters@parc.xerox.com>*
12122 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12124 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12126 * Add and OPENSSL_LOAD_CONF define which will cause
12127 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12128 This allows older applications to transparently support certain
12129 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12130 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12131 load the config file and OPENSSL_add_all_algorithms_conf() which will
12132 always load it have also been added.
12136 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12137 Adjust NIDs and EVP layer.
12139 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12141 * Config modules support in openssl utility.
12143 Most commands now load modules from the config file,
12144 though in a few (such as version) this isn't done
12145 because it couldn't be used for anything.
12147 In the case of ca and req the config file used is
12148 the same as the utility itself: that is the -config
12149 command line option can be used to specify an
12154 * Move default behaviour from OPENSSL_config(). If appname is NULL
12155 use "openssl_conf" if filename is NULL use default openssl config file.
12159 * Add an argument to OPENSSL_config() to allow the use of an alternative
12160 config section name. Add a new flag to tolerate a missing config file
12161 and move code to CONF_modules_load_file().
12165 * Support for crypto accelerator cards from Accelerated Encryption
12166 Processing, www.aep.ie. (Use engine 'aep')
12167 The support was copied from 0.9.6c [engine] and adapted/corrected
12168 to work with the new engine framework.
12170 *AEP Inc. and Richard Levitte*
12172 * Support for SureWare crypto accelerator cards from Baltimore
12173 Technologies. (Use engine 'sureware')
12174 The support was copied from 0.9.6c [engine] and adapted
12175 to work with the new engine framework.
12179 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12180 make the newer ENGINE framework commands for the CHIL engine work.
12182 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12184 * Make it possible to produce shared libraries on ReliantUNIX.
12186 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12188 * Add the configuration target debug-linux-ppro.
12189 Make 'openssl rsa' use the general key loading routines
12190 implemented in `apps.c`, and make those routines able to
12191 handle the key format FORMAT_NETSCAPE and the variant
12194 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12196 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12198 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12200 * Add -keyform to rsautl, and document -engine.
12202 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12204 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12205 BIO_R_NO_SUCH_FILE error code rather than the generic
12206 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12210 * Add new functions
12211 ERR_peek_last_error
12212 ERR_peek_last_error_line
12213 ERR_peek_last_error_line_data.
12214 These are similar to
12216 ERR_peek_error_line
12217 ERR_peek_error_line_data,
12218 but report on the latest error recorded rather than the first one
12219 still in the error queue.
12221 *Ben Laurie, Bodo Moeller*
12223 * default_algorithms option in ENGINE config module. This allows things
12225 default_algorithms = ALL
12226 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12230 * Preliminary ENGINE config module.
12234 * New experimental application configuration code.
12238 * Change the AES code to follow the same name structure as all other
12239 symmetric ciphers, and behave the same way. Move everything to
12240 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12242 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12244 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12246 *Ben Laurie and Theo de Raadt*
12248 * Add option to output public keys in req command.
12250 *Massimiliano Pala madwolf@openca.org*
12252 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12253 (up to about 10% better than before for P-192 and P-224).
12257 * New functions/macros
12259 SSL_CTX_set_msg_callback(ctx, cb)
12260 SSL_CTX_set_msg_callback_arg(ctx, arg)
12261 SSL_set_msg_callback(ssl, cb)
12262 SSL_set_msg_callback_arg(ssl, arg)
12264 to request calling a callback function
12266 void cb(int write_p, int version, int content_type,
12267 const void *buf, size_t len, SSL *ssl, void *arg)
12269 whenever a protocol message has been completely received
12270 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12271 protocol version according to which the SSL library interprets
12272 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12273 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12274 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12275 specification (change_cipher_spec(20), alert(21), handshake(22)).
12276 'buf' and 'len' point to the actual message, 'ssl' to the
12277 SSL object, and 'arg' is the application-defined value set by
12278 SSL[_CTX]_set_msg_callback_arg().
12280 'openssl s_client' and 'openssl s_server' have new '-msg' options
12281 to enable a callback that displays all protocol messages.
12285 * Change the shared library support so shared libraries are built as
12286 soon as the corresponding static library is finished, and thereby get
12287 openssl and the test programs linked against the shared library.
12288 This still only happens when the keyword "shard" has been given to
12289 the configuration scripts.
12291 NOTE: shared library support is still an experimental thing, and
12292 backward binary compatibility is still not guaranteed.
12294 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12296 * Add support for Subject Information Access extension.
12298 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12300 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12301 additional bytes when new memory had to be allocated, not just
12302 when reusing an existing buffer.
12306 * New command line and configuration option 'utf8' for the req command.
12307 This allows field values to be specified as UTF8 strings.
12311 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12312 runs for the former and machine-readable output for the latter.
12316 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12317 of the e-mail address in the DN (i.e., it will go into a certificate
12318 extension only). The new configuration file option 'email_in_dn = no'
12319 has the same effect.
12321 *Massimiliano Pala madwolf@openca.org*
12323 * Change all functions with names starting with `des_` to be starting
12324 with `DES_` instead. Add wrappers that are compatible with libdes,
12325 but are named `_ossl_old_des_*`. Finally, add macros that map the
12326 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12327 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12328 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12331 Since we provide two compatibility mappings, the user needs to
12332 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12333 compatibility is desired. The default (i.e., when that macro
12334 isn't defined) is OpenSSL 0.9.6c compatibility.
12336 There are also macros that enable and disable the support of old
12337 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12338 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12339 are defined, the default will apply: to support the old des routines.
12341 In either case, one must include openssl/des.h to get the correct
12342 definitions. Do not try to just include openssl/des_old.h, that
12345 NOTE: This is a major break of an old API into a new one. Software
12346 authors are encouraged to switch to the `DES_` style functions. Some
12347 time in the future, des_old.h and the libdes compatibility functions
12348 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12349 default), and then completely removed.
12353 * Test for certificates which contain unsupported critical extensions.
12354 If such a certificate is found during a verify operation it is
12355 rejected by default: this behaviour can be overridden by either
12356 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12357 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12358 X509_supported_extension() has also been added which returns 1 if a
12359 particular extension is supported.
12363 * Modify the behaviour of EVP cipher functions in similar way to digests
12364 to retain compatibility with existing code.
12368 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12369 compatibility with existing code. In particular the 'ctx' parameter does
12370 not have to be to be initialized before the call to EVP_DigestInit() and
12371 it is tidied up after a call to EVP_DigestFinal(). New function
12372 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12373 EVP_MD_CTX_copy() changed to not require the destination to be
12374 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12375 requires the destination to be valid.
12377 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12378 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12382 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12383 so that complete 'Handshake' protocol structures are kept in memory
12384 instead of overwriting 'msg_type' and 'length' with 'body' data.
12388 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12390 *Massimo Santin via Richard Levitte*
12392 * Major restructuring to the underlying ENGINE code. This includes
12393 reduction of linker bloat, separation of pure "ENGINE" manipulation
12394 (initialisation, etc) from functionality dealing with implementations
12395 of specific crypto interfaces. This change also introduces integrated
12396 support for symmetric ciphers and digest implementations - so ENGINEs
12397 can now accelerate these by providing EVP_CIPHER and EVP_MD
12398 implementations of their own. This is detailed in
12399 [crypto/engine/README.md](crypto/engine/README.md)
12400 as it couldn't be adequately described here. However, there are a few
12401 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12402 were changed in the original introduction of ENGINE code have now
12403 reverted back - the hooking from this code to ENGINE is now a good
12404 deal more passive and at run-time, operations deal directly with
12405 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12406 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12407 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12408 they were not being used by the framework as there is no concept of a
12409 BIGNUM_METHOD and they could not be generalised to the new
12410 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12411 ENGINE_cpy() has been removed as it cannot be consistently defined in
12416 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12420 * Change mkdef.pl to sort symbols that get the same entry number,
12421 and make sure the automatically generated functions `ERR_load_*`
12422 become part of libeay.num as well.
12426 * New function SSL_renegotiate_pending(). This returns true once
12427 renegotiation has been requested (either SSL_renegotiate() call
12428 or HelloRequest/ClientHello received from the peer) and becomes
12429 false once a handshake has been completed.
12430 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12431 sends a HelloRequest, but does not ensure that a handshake takes
12432 place. SSL_renegotiate_pending() is useful for checking if the
12433 client has followed the request.)
12437 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12438 By default, clients may request session resumption even during
12439 renegotiation (if session ID contexts permit); with this option,
12440 session resumption is possible only in the first handshake.
12442 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12443 more bits available for options that should not be part of
12444 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12448 * Add some demos for certificate and certificate request creation.
12452 * Make maximum certificate chain size accepted from the peer application
12453 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12454 "Douglas E. Engert" <deengert@anl.gov>.
12458 * Add support for shared libraries for Unixware-7
12459 (Boyd Lynn Gerber <gerberb@zenez.com>).
12463 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12464 be done prior to destruction. Use this to unload error strings from
12465 ENGINEs that load their own error strings. NB: This adds two new API
12466 functions to "get" and "set" this destroy handler in an ENGINE.
12470 * Alter all existing ENGINE implementations (except "openssl" and
12471 "openbsd") to dynamically instantiate their own error strings. This
12472 makes them more flexible to be built both as statically-linked ENGINEs
12473 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12474 Also, add stub code to each that makes building them as self-contained
12475 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12479 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12480 implementations into applications that are completely implemented in
12481 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12482 commands that can be used to configure what shared-library to load and
12483 to control aspects of the way it is handled. Also, made an update to
12484 the [README-Engine.md](README-Engine.md) file
12485 that brings its information up-to-date and
12486 provides some information and instructions on the "dynamic" ENGINE
12487 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12491 * Make it possible to unload ranges of ERR strings with a new
12492 "ERR_unload_strings" function.
12496 * Add a copy() function to EVP_MD.
12500 * Make EVP_MD routines take a context pointer instead of just the
12501 md_data void pointer.
12505 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12506 that the digest can only process a single chunk of data
12507 (typically because it is provided by a piece of
12508 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12509 is only going to provide a single chunk of data, and hence the
12510 framework needn't accumulate the data for oneshot drivers.
12514 * As with "ERR", make it possible to replace the underlying "ex_data"
12515 functions. This change also alters the storage and management of global
12516 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12517 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12518 index counters. The API functions that use this state have been changed
12519 to take a "class_index" rather than pointers to the class's local STACK
12520 and counter, and there is now an API function to dynamically create new
12521 classes. This centralisation allows us to (a) plug a lot of the
12522 thread-safety problems that existed, and (b) makes it possible to clean
12523 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12524 such data would previously have always leaked in application code and
12525 workarounds were in place to make the memory debugging turn a blind eye
12526 to it. Application code that doesn't use this new function will still
12527 leak as before, but their memory debugging output will announce it now
12528 rather than letting it slide.
12530 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12531 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12532 has a return value to indicate success or failure.
12536 * Make it possible to replace the underlying "ERR" functions such that the
12537 global state (2 LHASH tables and 2 locks) is only used by the "default"
12538 implementation. This change also adds two functions to "get" and "set"
12539 the implementation prior to it being automatically set the first time
12540 any other ERR function takes place. Ie. an application can call "get",
12541 pass the return value to a module it has just loaded, and that module
12542 can call its own "set" function using that value. This means the
12543 module's "ERR" operations will use (and modify) the error state in the
12544 application and not in its own statically linked copy of OpenSSL code.
12548 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12549 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12550 the operation, and provides a more encapsulated way for external code
12551 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12552 to use these functions rather than manually incrementing the counts.
12554 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12558 * Add EVP test program.
12562 * Add symmetric cipher support to ENGINE. Expect the API to change!
12566 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12567 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12568 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12569 These allow a CRL to be built without having to access X509_CRL fields
12570 directly. Modify 'ca' application to use new functions.
12574 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12575 bug workarounds. Rollback attack detection is a security feature.
12576 The problem will only arise on OpenSSL servers when TLSv1 is not
12577 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12578 Software authors not wanting to support TLSv1 will have special reasons
12579 for their choice and can explicitly enable this option.
12581 *Bodo Moeller, Lutz Jaenicke*
12583 * Rationalise EVP so it can be extended: don't include a union of
12584 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12585 (similar to those existing for EVP_CIPHER_CTX).
12590 EVP_MD_CTX_init(&md); /* new function call */
12591 EVP_DigestInit(&md, EVP_sha1());
12592 EVP_DigestUpdate(&md, in, len);
12593 EVP_DigestFinal(&md, out, NULL);
12594 EVP_MD_CTX_cleanup(&md); /* new function call */
12598 * Make DES key schedule conform to the usual scheme, as well as
12599 correcting its structure. This means that calls to DES functions
12600 now have to pass a pointer to a des_key_schedule instead of a
12601 plain des_key_schedule (which was actually always a pointer
12604 des_key_schedule ks;
12606 des_set_key_checked(..., &ks);
12607 des_ncbc_encrypt(..., &ks, ...);
12609 (Note that a later change renames 'des_...' into 'DES_...'.)
12613 * Initial reduction of linker bloat: the use of some functions, such as
12614 PEM causes large amounts of unused functions to be linked in due to
12615 poor organisation. For example pem_all.c contains every PEM function
12616 which has a knock on effect of linking in large amounts of (unused)
12617 ASN1 code. Grouping together similar functions and splitting unrelated
12618 functions prevents this.
12622 * Cleanup of EVP macros.
12626 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12627 correct `_ecb suffix`.
12631 * Add initial OCSP responder support to ocsp application. The
12632 revocation information is handled using the text based index
12633 use by the ca application. The responder can either handle
12634 requests generated internally, supplied in files (for example
12635 via a CGI script) or using an internal minimal server.
12639 * Add configuration choices to get zlib compression for TLS.
12643 * Changes to Kerberos SSL for RFC 2712 compliance:
12644 1. Implemented real KerberosWrapper, instead of just using
12645 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12646 2. Implemented optional authenticator field of KerberosWrapper.
12648 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12649 and authenticator structs; see crypto/krb5/.
12651 Generalized Kerberos calls to support multiple Kerberos libraries.
12652 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12653 via Richard Levitte*
12655 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12656 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12657 values for each of the key sizes rather than having just
12658 parameters (and 'speed' generating keys each time).
12662 * Speed up EVP routines.
12665 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12666 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12667 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12668 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12670 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12671 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12672 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12675 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12677 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12681 * Added the OS2-EMX target.
12683 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12685 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12686 New functions to support `NCONF `routines in extension code.
12687 New function `CONF_set_nconf()`
12688 to allow functions which take an `NCONF` to also handle the old `LHASH`
12689 structure: this means that the old `CONF` compatible routines can be
12690 retained (in particular w.rt. extensions) without having to duplicate the
12691 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12695 * Enhance the general user interface with mechanisms for inner control
12696 and with possibilities to have yes/no kind of prompts.
12700 * Change all calls to low level digest routines in the library and
12701 applications to use EVP. Add missing calls to HMAC_cleanup() and
12702 don't assume HMAC_CTX can be copied using memcpy().
12704 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12706 * Add the possibility to control engines through control names but with
12707 arbitrary arguments instead of just a string.
12708 Change the key loaders to take a UI_METHOD instead of a callback
12709 function pointer. NOTE: this breaks binary compatibility with earlier
12710 versions of OpenSSL [engine].
12711 Adapt the nCipher code for these new conditions and add a card insertion
12716 * Enhance the general user interface with mechanisms to better support
12717 dialog box interfaces, application-defined prompts, the possibility
12718 to use defaults (for example default passwords from somewhere else)
12719 and interrupts/cancellations.
12723 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12724 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12728 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12729 tidy up some unnecessarily weird code in 'sk_new()').
12731 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12733 * Change the key loading routines for ENGINEs to use the same kind
12734 callback (pem_password_cb) as all other routines that need this
12739 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12740 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12741 than this minimum value is recommended.
12745 * New random seeder for OpenVMS, using the system process statistics
12746 that are easily reachable.
12750 * Windows apparently can't transparently handle global
12751 variables defined in DLLs. Initialisations such as:
12753 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12755 won't compile. This is used by the any applications that need to
12756 declare their own ASN1 modules. This was fixed by adding the option
12757 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12758 needed for static libraries under Win32.
12762 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12763 setting of purpose and trust fields. New X509_STORE trust and
12764 purpose functions and tidy up setting in other SSL functions.
12768 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12769 structure. These are inherited by X509_STORE_CTX when it is
12770 initialised. This allows various defaults to be set in the
12771 X509_STORE structure (such as flags for CRL checking and custom
12772 purpose or trust settings) for functions which only use X509_STORE_CTX
12773 internally such as S/MIME.
12775 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12776 trust settings if they are not set in X509_STORE. This allows X509_STORE
12777 purposes and trust (in S/MIME for example) to override any set by default.
12779 Add command line options for CRL checking to smime, s_client and s_server
12784 * Initial CRL based revocation checking. If the CRL checking flag(s)
12785 are set then the CRL is looked up in the X509_STORE structure and
12786 its validity and signature checked, then if the certificate is found
12787 in the CRL the verify fails with a revoked error.
12789 Various new CRL related callbacks added to X509_STORE_CTX structure.
12791 Command line options added to 'verify' application to support this.
12793 This needs some additional work, such as being able to handle multiple
12794 CRLs with different times, extension based lookup (rather than just
12795 by subject name) and ultimately more complete V2 CRL extension
12800 * Add a general user interface API (crypto/ui/). This is designed
12801 to replace things like des_read_password and friends (backward
12802 compatibility functions using this new API are provided).
12803 The purpose is to remove prompting functions from the DES code
12804 section as well as provide for prompting through dialog boxes in
12805 a window system and the like.
12809 * Add "ex_data" support to ENGINE so implementations can add state at a
12810 per-structure level rather than having to store it globally.
12814 * Make it possible for ENGINE structures to be copied when retrieved by
12815 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12816 This causes the "original" ENGINE structure to act like a template,
12817 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12818 operational state can be localised to each ENGINE structure, despite the
12819 fact they all share the same "methods". New ENGINE structures returned in
12820 this case have no functional references and the return value is the single
12821 structural reference. This matches the single structural reference returned
12822 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12827 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12828 needs to match any other type at all we need to manually clear the
12833 * Changes to the "openssl engine" utility to include;
12834 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12835 about an ENGINE's available control commands.
12836 - executing control commands from command line arguments using the
12837 '-pre' and '-post' switches. '-post' is only used if '-t' is
12838 specified and the ENGINE is successfully initialised. The syntax for
12839 the individual commands are colon-separated, for example;
12840 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12844 * New dynamic control command support for ENGINEs. ENGINEs can now
12845 declare their own commands (numbers), names (strings), descriptions,
12846 and input types for run-time discovery by calling applications. A
12847 subset of these commands are implicitly classed as "executable"
12848 depending on their input type, and only these can be invoked through
12849 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12850 can be based on user input, config files, etc). The distinction is
12851 that "executable" commands cannot return anything other than a boolean
12852 result and can only support numeric or string input, whereas some
12853 discoverable commands may only be for direct use through
12854 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12855 pointers, or other custom uses. The "executable" commands are to
12856 support parameterisations of ENGINE behaviour that can be
12857 unambiguously defined by ENGINEs and used consistently across any
12858 OpenSSL-based application. Commands have been added to all the
12859 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12860 control over shared-library paths without source code alterations.
12864 * Changed all ENGINE implementations to dynamically allocate their
12865 ENGINEs rather than declaring them statically. Apart from this being
12866 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12867 this also allows the implementations to compile without using the
12868 internal engine_int.h header.
12872 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12873 'const' value. Any code that should be able to modify a RAND_METHOD
12874 should already have non-const pointers to it (ie. they should only
12875 modify their own ones).
12879 * Made a variety of little tweaks to the ENGINE code.
12880 - "atalla" and "ubsec" string definitions were moved from header files
12881 to C code. "nuron" string definitions were placed in variables
12882 rather than hard-coded - allowing parameterisation of these values
12883 later on via ctrl() commands.
12884 - Removed unused "#if 0"'d code.
12885 - Fixed engine list iteration code so it uses ENGINE_free() to release
12886 structural references.
12887 - Constified the RAND_METHOD element of ENGINE structures.
12888 - Constified various get/set functions as appropriate and added
12889 missing functions (including a catch-all ENGINE_cpy that duplicates
12890 all ENGINE values onto a new ENGINE except reference counts/state).
12891 - Removed NULL parameter checks in get/set functions. Setting a method
12892 or function to NULL is a way of cancelling out a previously set
12893 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12894 and doesn't justify the extra error symbols and code.
12895 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12896 flags from engine_int.h to engine.h.
12897 - Changed prototypes for ENGINE handler functions (init(), finish(),
12898 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12902 * Implement binary inversion algorithm for BN_mod_inverse in addition
12903 to the algorithm using long division. The binary algorithm can be
12904 used only if the modulus is odd. On 32-bit systems, it is faster
12905 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12906 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12907 up to 450 bits. In 64-bit environments, the binary algorithm
12908 appears to be advantageous for much longer moduli; here we use it
12909 for moduli up to 2048 bits.
12913 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12914 could not support the combine flag in choice fields.
12918 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12919 extensions from a certificate request to the certificate.
12923 * Allow multiple 'certopt' and 'nameopt' options to be separated
12924 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12925 file: this allows the display of the certificate about to be
12926 signed to be customised, to allow certain fields to be included
12927 or excluded and extension details. The old system didn't display
12928 multicharacter strings properly, omitted fields not in the policy
12929 and couldn't display additional details such as extensions.
12933 * Function EC_POINTs_mul for multiple scalar multiplication
12934 of an arbitrary number of elliptic curve points
12935 \sum scalars[i]*points[i],
12936 optionally including the generator defined for the EC_GROUP:
12937 scalar*generator + \sum scalars[i]*points[i].
12939 EC_POINT_mul is a simple wrapper function for the typical case
12940 that the point list has just one item (besides the optional
12945 * First EC_METHODs for curves over GF(p):
12947 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12948 operations and provides various method functions that can also
12949 operate with faster implementations of modular arithmetic.
12951 EC_GFp_mont_method() reuses most functions that are part of
12952 EC_GFp_simple_method, but uses Montgomery arithmetic.
12954 *Bodo Moeller; point addition and point doubling
12955 implementation directly derived from source code provided by
12956 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12958 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12959 crypto/ec/ec_lib.c):
12961 Curves are EC_GROUP objects (with an optional group generator)
12962 based on EC_METHODs that are built into the library.
12964 Points are EC_POINT objects based on EC_GROUP objects.
12966 Most of the framework would be able to handle curves over arbitrary
12967 finite fields, but as there are no obvious types for fields other
12968 than GF(p), some functions are limited to that for now.
12972 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12973 that the file contains a complete HTTP response.
12977 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12978 change the def and num file printf format specifier from "%-40sXXX"
12979 to "%-39s XXX". The latter will always guarantee a space after the
12980 field while the former will cause them to run together if the field
12981 is 40 of more characters long.
12985 * Constify the cipher and digest 'method' functions and structures
12986 and modify related functions to take constant EVP_MD and EVP_CIPHER
12991 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12992 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12996 * Modify `EVP_Digest*()` routines so they now return values. Although the
12997 internal software routines can never fail additional hardware versions
13002 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13004 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13005 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13008 ERR_R_NESTED_ASN1_ERROR
13010 ERR_R_MISSING_ASN1_EOS
13011 were 4 .. 9, conflicting with
13012 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13014 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13015 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13017 Add new error code 'ERR_R_INTERNAL_ERROR'.
13021 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13026 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13027 sets the subject name for a new request or supersedes the
13028 subject name in a given request. Formats that can be parsed are
13029 'CN=Some Name, OU=myOU, C=IT'
13031 'CN=Some Name/OU=myOU/C=IT'.
13033 Add options '-batch' and '-verbose' to 'openssl req'.
13035 *Massimiliano Pala <madwolf@hackmasters.net>*
13037 * Introduce the possibility to access global variables through
13038 functions on platform were that's the best way to handle exporting
13039 global variables in shared libraries. To enable this functionality,
13040 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13041 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13042 is normally done by Configure or something similar).
13044 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13045 in the source file (foo.c) like this:
13047 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13048 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13050 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13051 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13053 OPENSSL_DECLARE_GLOBAL(int,foo);
13054 #define foo OPENSSL_GLOBAL_REF(foo)
13055 OPENSSL_DECLARE_GLOBAL(double,bar);
13056 #define bar OPENSSL_GLOBAL_REF(bar)
13058 The #defines are very important, and therefore so is including the
13059 header file everywhere where the defined globals are used.
13061 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13062 of ASN.1 items, but that structure is a bit different.
13064 The largest change is in util/mkdef.pl which has been enhanced with
13065 better and easier to understand logic to choose which symbols should
13066 go into the Windows .def files as well as a number of fixes and code
13067 cleanup (among others, algorithm keywords are now sorted
13068 lexicographically to avoid constant rewrites).
13072 * In BN_div() keep a copy of the sign of 'num' before writing the
13073 result to 'rm' because if rm==num the value will be overwritten
13074 and produce the wrong result if 'num' is negative: this caused
13075 problems with BN_mod() and BN_nnmod().
13079 * Function OCSP_request_verify(). This checks the signature on an
13080 OCSP request and verifies the signer certificate. The signer
13081 certificate is just checked for a generic purpose and OCSP request
13086 * Add OCSP_check_validity() function to check the validity of OCSP
13087 responses. OCSP responses are prepared in real time and may only
13088 be a few seconds old. Simply checking that the current time lies
13089 between thisUpdate and nextUpdate max reject otherwise valid responses
13090 caused by either OCSP responder or client clock inaccuracy. Instead
13091 we allow thisUpdate and nextUpdate to fall within a certain period of
13092 the current time. The age of the response can also optionally be
13093 checked. Two new options -validity_period and -status_age added to
13098 * If signature or public key algorithm is unrecognized print out its
13099 OID rather that just UNKNOWN.
13103 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13104 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13105 ID to be generated from the issuer certificate alone which can then be
13106 passed to OCSP_id_issuer_cmp().
13110 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13111 ASN1 modules to export functions returning ASN1_ITEM pointers
13112 instead of the ASN1_ITEM structures themselves. This adds several
13113 new macros which allow the underlying ASN1 function/structure to
13114 be accessed transparently. As a result code should not use ASN1_ITEM
13115 references directly (such as &X509_it) but instead use the relevant
13116 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13117 use of the new ASN1 code on platforms where exporting structures
13118 is problematical (for example in shared libraries) but exporting
13119 functions returning pointers to structures is not.
13123 * Add support for overriding the generation of SSL/TLS session IDs.
13124 These callbacks can be registered either in an SSL_CTX or per SSL.
13125 The purpose of this is to allow applications to control, if they wish,
13126 the arbitrary values chosen for use as session IDs, particularly as it
13127 can be useful for session caching in multiple-server environments. A
13128 command-line switch for testing this (and any client code that wishes
13129 to use such a feature) has been added to "s_server".
13131 *Geoff Thorpe, Lutz Jaenicke*
13133 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13134 of the form `#if defined(...) || defined(...) || ...` and
13135 `#if !defined(...) && !defined(...) && ...`. This also avoids
13136 the growing number of special cases it was previously handling.
13140 * Make all configuration macros available for application by making
13141 sure they are available in opensslconf.h, by giving them names starting
13142 with `OPENSSL_` to avoid conflicts with other packages and by making
13143 sure e_os2.h will cover all platform-specific cases together with
13145 Additionally, it is now possible to define configuration/platform-
13146 specific names (called "system identities"). In the C code, these
13147 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13148 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13149 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13154 * New option -set_serial to 'req' and 'x509' this allows the serial
13155 number to use to be specified on the command line. Previously self
13156 signed certificates were hard coded with serial number 0 and the
13157 CA options of 'x509' had to use a serial number in a file which was
13162 * New options to 'ca' utility to support V2 CRL entry extensions.
13163 Currently CRL reason, invalidity date and hold instruction are
13164 supported. Add new CRL extensions to V3 code and some new objects.
13168 * New function EVP_CIPHER_CTX_set_padding() this is used to
13169 disable standard block padding (aka PKCS#5 padding) in the EVP
13170 API, which was previously mandatory. This means that the data is
13171 not padded in any way and so the total length much be a multiple
13172 of the block size, otherwise an error occurs.
13176 * Initial (incomplete) OCSP SSL support.
13180 * New function OCSP_parse_url(). This splits up a URL into its host,
13181 port and path components: primarily to parse OCSP URLs. New -url
13182 option to ocsp utility.
13186 * New nonce behavior. The return value of OCSP_check_nonce() now
13187 reflects the various checks performed. Applications can decide
13188 whether to tolerate certain situations such as an absent nonce
13189 in a response when one was present in a request: the ocsp application
13190 just prints out a warning. New function OCSP_add1_basic_nonce()
13191 this is to allow responders to include a nonce in a response even if
13192 the request is nonce-less.
13196 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13197 skipped when using openssl x509 multiple times on a single input file,
13198 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13202 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13203 set string type: to handle setting ASN1_TIME structures. Fix ca
13204 utility to correctly initialize revocation date of CRLs.
13208 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13209 the clients preferred ciphersuites and rather use its own preferences.
13210 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13211 Internet Explorer by ensuring unchanged hash method during stepup.
13212 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13216 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13217 to aes and add a new 'exist' option to print out symbols that don't
13222 * Additional options to ocsp utility to allow flags to be set and
13223 additional certificates supplied.
13227 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13228 OCSP client a number of certificate to only verify the response
13233 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13234 handle the new API. Currently only ECB, CBC modes supported. Add new
13237 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13238 Encryption Standard (AES) Ciphersuites for Transport Layer
13239 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13240 not enabled by default and were not part of the "ALL" ciphersuite
13241 alias because they were not yet official; they could be
13242 explicitly requested by specifying the "AESdraft" ciphersuite
13243 group alias. In the final release of OpenSSL 0.9.7, the group
13244 alias is called "AES" and is part of "ALL".)
13246 *Ben Laurie, Steve Henson, Bodo Moeller*
13248 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13249 request to response.
13253 * Functions for OCSP responders. OCSP_request_onereq_count(),
13254 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13255 extract information from a certificate request. OCSP_response_create()
13256 creates a response and optionally adds a basic response structure.
13257 OCSP_basic_add1_status() adds a complete single response to a basic
13258 response and returns the OCSP_SINGLERESP structure just added (to allow
13259 extensions to be included for example). OCSP_basic_add1_cert() adds a
13260 certificate to a basic response and OCSP_basic_sign() signs a basic
13261 response with various flags. New helper functions ASN1_TIME_check()
13262 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13263 (converts ASN1_TIME to GeneralizedTime).
13267 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13268 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13269 structure from a certificate. X509_pubkey_digest() digests the public_key
13270 contents: this is used in various key identifiers.
13274 * Make sk_sort() tolerate a NULL argument.
13276 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13278 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13279 passed by the function are trusted implicitly. If any of them signed the
13280 response then it is assumed to be valid and is not verified.
13284 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13285 to data. This was previously part of the PKCS7 ASN1 code. This
13286 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13287 *Steve Henson, reported by Kenneth R. Robinette
13288 <support@securenetterm.com>*
13290 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13291 routines: without these tracing memory leaks is very painful.
13292 Fix leaks in PKCS12 and PKCS7 routines.
13296 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13297 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13298 effectively meant GeneralizedTime would never be used. Now it
13299 is initialised to -1 but X509_time_adj() now has to check the value
13300 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13301 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13302 *Steve Henson, reported by Kenneth R. Robinette
13303 <support@securenetterm.com>*
13305 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13306 result in a zero length in the ASN1_INTEGER structure which was
13307 not consistent with the structure when d2i_ASN1_INTEGER() was used
13308 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13309 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13310 where it did not print out a minus for negative ASN1_INTEGER.
13314 * Add summary printout to ocsp utility. The various functions which
13315 convert status values to strings have been renamed to:
13316 OCSP_response_status_str(), OCSP_cert_status_str() and
13317 OCSP_crl_reason_str() and are no longer static. New options
13318 to verify nonce values and to disable verification. OCSP response
13319 printout format cleaned up.
13323 * Add additional OCSP certificate checks. These are those specified
13324 in RFC2560. This consists of two separate checks: the CA of the
13325 certificate being checked must either be the OCSP signer certificate
13326 or the issuer of the OCSP signer certificate. In the latter case the
13327 OCSP signer certificate must contain the OCSP signing extended key
13328 usage. This check is performed by attempting to match the OCSP
13329 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13330 in the OCSP_CERTID structures of the response.
13334 * Initial OCSP certificate verification added to OCSP_basic_verify()
13335 and related routines. This uses the standard OpenSSL certificate
13336 verify routines to perform initial checks (just CA validity) and
13337 to obtain the certificate chain. Then additional checks will be
13338 performed on the chain. Currently the root CA is checked to see
13339 if it is explicitly trusted for OCSP signing. This is used to set
13340 a root CA as a global signing root: that is any certificate that
13341 chains to that CA is an acceptable OCSP signing certificate.
13345 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13346 extensions from a separate configuration file.
13347 As when reading extensions from the main configuration file,
13348 the '-extensions ...' option may be used for specifying the
13351 *Massimiliano Pala <madwolf@comune.modena.it>*
13353 * New OCSP utility. Allows OCSP requests to be generated or
13354 read. The request can be sent to a responder and the output
13355 parsed, outputted or printed in text form. Not complete yet:
13356 still needs to check the OCSP response validity.
13360 * New subcommands for 'openssl ca':
13361 `openssl ca -status <serial>` prints the status of the cert with
13362 the given serial number (according to the index file).
13363 `openssl ca -updatedb` updates the expiry status of certificates
13366 *Massimiliano Pala <madwolf@comune.modena.it>*
13368 * New '-newreq-nodes' command option to CA.pl. This is like
13369 '-newreq', but calls 'openssl req' with the '-nodes' option
13370 so that the resulting key is not encrypted.
13372 *Damien Miller <djm@mindrot.org>*
13374 * New configuration for the GNU Hurd.
13376 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13378 * Initial code to implement OCSP basic response verify. This
13379 is currently incomplete. Currently just finds the signer's
13380 certificate and verifies the signature on the response.
13384 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13385 value of OPENSSLDIR. This is available via the new '-d' option
13386 to 'openssl version', and is also included in 'openssl version -a'.
13390 * Allowing defining memory allocation callbacks that will be given
13391 file name and line number information in additional arguments
13392 (a `const char*` and an int). The basic functionality remains, as
13393 well as the original possibility to just replace malloc(),
13394 realloc() and free() by functions that do not know about these
13395 additional arguments. To register and find out the current
13396 settings for extended allocation functions, the following
13397 functions are provided:
13399 CRYPTO_set_mem_ex_functions
13400 CRYPTO_set_locked_mem_ex_functions
13401 CRYPTO_get_mem_ex_functions
13402 CRYPTO_get_locked_mem_ex_functions
13404 These work the same way as CRYPTO_set_mem_functions and friends.
13405 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13406 extended allocation function is enabled.
13407 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13408 a conventional allocation function is enabled.
13410 *Richard Levitte, Bodo Moeller*
13412 * Finish off removing the remaining LHASH function pointer casts.
13413 There should no longer be any prototype-casting required when using
13414 the LHASH abstraction, and any casts that remain are "bugs". See
13415 the callback types and macros at the head of lhash.h for details
13416 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13420 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13421 If /dev/[u]random devices are not available or do not return enough
13422 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13424 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13425 /etc/entropy will be queried once each in this sequence, querying stops
13426 when enough entropy was collected without querying more sockets.
13430 * Change the Unix RAND_poll() variant to be able to poll several
13431 random devices, as specified by DEVRANDOM, until a sufficient amount
13432 of data has been collected. We spend at most 10 ms on each file
13433 (select timeout) and read in non-blocking mode. DEVRANDOM now
13434 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13435 (previously it was just the string "/dev/urandom"), so on typical
13436 platforms the 10 ms delay will never occur.
13437 Also separate out the Unix variant to its own file, rand_unix.c.
13438 For VMS, there's a currently-empty rand_vms.c.
13442 * Move OCSP client related routines to ocsp_cl.c. These
13443 provide utility functions which an application needing
13444 to issue a request to an OCSP responder and analyse the
13445 response will typically need: as opposed to those which an
13446 OCSP responder itself would need which will be added later.
13448 OCSP_request_sign() signs an OCSP request with an API similar
13449 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13450 response. OCSP_response_get1_basic() extracts basic response
13451 from response. OCSP_resp_find_status(): finds and extracts status
13452 information from an OCSP_CERTID structure (which will be created
13453 when the request structure is built). These are built from lower
13454 level functions which work on OCSP_SINGLERESP structures but
13455 won't normally be used unless the application wishes to examine
13456 extensions in the OCSP response for example.
13458 Replace nonce routines with a pair of functions.
13459 OCSP_request_add1_nonce() adds a nonce value and optionally
13460 generates a random value. OCSP_check_nonce() checks the
13461 validity of the nonce in an OCSP response.
13465 * Change function OCSP_request_add() to OCSP_request_add0_id().
13466 This doesn't copy the supplied OCSP_CERTID and avoids the
13467 need to free up the newly created id. Change return type
13468 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13469 This can then be used to add extensions to the request.
13470 Deleted OCSP_request_new(), since most of its functionality
13471 is now in OCSP_REQUEST_new() (and the case insensitive name
13472 clash) apart from the ability to set the request name which
13473 will be added elsewhere.
13477 * Update OCSP API. Remove obsolete extensions argument from
13478 various functions. Extensions are now handled using the new
13479 OCSP extension code. New simple OCSP HTTP function which
13480 can be used to send requests and parse the response.
13484 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13485 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13486 uses the special reorder version of SET OF to sort the attributes
13487 and reorder them to match the encoded order. This resolves a long
13488 standing problem: a verify on a PKCS7 structure just after signing
13489 it used to fail because the attribute order did not match the
13490 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13491 it uses the received order. This is necessary to tolerate some broken
13492 software that does not order SET OF. This is handled by encoding
13493 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13494 to produce the required SET OF.
13498 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13499 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13500 files to get correct declarations of the ASN.1 item variables.
13504 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13505 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13506 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13507 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13508 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13509 ASN1_ITEM and no wrapper functions.
13513 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13514 replace the old function pointer based I/O routines. Change most of
13515 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13519 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13520 lines, recognize more "algorithms" that can be deselected, and make
13521 it complain about algorithm deselection that isn't recognised.
13525 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13526 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13527 to use new functions. Add NO_ASN1_OLD which can be set to remove
13528 some old style ASN1 functions: this can be used to determine if old
13529 code will still work when these eventually go away.
13533 * New extension functions for OCSP structures, these follow the
13534 same conventions as certificates and CRLs.
13538 * New function X509V3_add1_i2d(). This automatically encodes and
13539 adds an extension. Its behaviour can be customised with various
13540 flags to append, replace or delete. Various wrappers added for
13541 certificates and CRLs.
13545 * Fix to avoid calling the underlying ASN1 print routine when
13546 an extension cannot be parsed. Correct a typo in the
13547 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13551 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13552 entries for variables.
13556 * Add functionality to `apps/openssl.c` for detecting locking
13557 problems: As the program is single-threaded, all we have
13558 to do is register a locking callback using an array for
13559 storing which locks are currently held by the program.
13563 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13564 SSL_get_ex_data_X509_STORE_idx(), which is used in
13565 ssl_verify_cert_chain() and thus can be called at any time
13566 during TLS/SSL handshakes so that thread-safety is essential.
13567 Unfortunately, the ex_data design is not at all suited
13568 for multi-threaded use, so it probably should be abolished.
13572 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13574 *Broadcom, tweaked and integrated by Geoff Thorpe*
13576 * Move common extension printing code to new function
13577 X509V3_print_extensions(). Reorganise OCSP print routines and
13578 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13582 * New function X509_signature_print() to remove duplication in some
13587 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13588 set (this was treated exactly the same as SET OF previously). This
13589 is used to reorder the STACK representing the structure to match the
13590 encoding. This will be used to get round a problem where a PKCS7
13591 structure which was signed could not be verified because the STACK
13592 order did not reflect the encoded order.
13596 * Reimplement the OCSP ASN1 module using the new code.
13600 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13601 for its ASN1 operations. The old style function pointers still exist
13602 for now but they will eventually go away.
13606 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13607 completely replaces the old ASN1 functionality with a table driven
13608 encoder and decoder which interprets an ASN1_ITEM structure describing
13609 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13610 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13611 has also been converted to the new form.
13615 * Change BN_mod_exp_recp so that negative moduli are tolerated
13616 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13617 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13618 for negative moduli.
13622 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13623 of not touching the result's sign bit.
13627 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13632 * Changed the LHASH code to use prototypes for callbacks, and created
13633 macros to declare and implement thin (optionally static) functions
13634 that provide type-safety and avoid function pointer casting for the
13635 type-specific callbacks.
13639 * Added Kerberos Cipher Suites to be used with TLS, as written in
13641 *Veers Staats <staatsvr@asc.hpc.mil>,
13642 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13644 * Reformat the FAQ so the different questions and answers can be divided
13645 in sections depending on the subject.
13649 * Have the zlib compression code load ZLIB.DLL dynamically under
13654 * New function BN_mod_sqrt for computing square roots modulo a prime
13655 (using the probabilistic Tonelli-Shanks algorithm unless
13656 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13657 be handled deterministically).
13659 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13661 * Make BN_mod_inverse faster by explicitly handling small quotients
13662 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13663 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13667 * New function BN_kronecker.
13671 * Fix BN_gcd so that it works on negative inputs; the result is
13672 positive unless both parameters are zero.
13673 Previously something reasonably close to an infinite loop was
13674 possible because numbers could be growing instead of shrinking
13675 in the implementation of Euclid's algorithm.
13679 * Fix BN_is_word() and BN_is_one() macros to take into account the
13680 sign of the number in question.
13682 Fix BN_is_word(a,w) to work correctly for w == 0.
13684 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13685 because its test if the absolute value of 'a' equals 'w'.
13686 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13687 it exists mostly for use in the implementations of BN_is_zero(),
13688 BN_is_one(), and BN_is_word().
13692 * New function BN_swap.
13696 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13697 the exponentiation functions are more likely to produce reasonable
13698 results on negative inputs.
13702 * Change BN_mod_mul so that the result is always non-negative.
13703 Previously, it could be negative if one of the factors was negative;
13704 I don't think anyone really wanted that behaviour.
13708 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13709 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13710 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13711 and add new functions:
13720 BN_mod_lshift1_quick
13722 BN_mod_lshift_quick
13724 These functions always generate non-negative results.
13726 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13727 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13729 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13730 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13731 be reduced modulo `m`.
13733 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13736 The following entry accidentally appeared in the CHANGES file
13737 distributed with OpenSSL 0.9.7. The modifications described in
13738 it do *not* apply to OpenSSL 0.9.7.
13740 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13741 was actually never needed) and in BN_mul(). The removal in BN_mul()
13742 required a small change in bn_mul_part_recursive() and the addition
13743 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13744 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13745 bn_sub_words() and bn_add_words() except they take arrays with
13751 * In 'openssl passwd', verify passwords read from the terminal
13752 unless the '-salt' option is used (which usually means that
13753 verification would just waste user's time since the resulting
13754 hash is going to be compared with some given password hash)
13755 or the new '-noverify' option is used.
13757 This is an incompatible change, but it does not affect
13758 non-interactive use of 'openssl passwd' (passwords on the command
13759 line, '-stdin' option, '-in ...' option) and thus should not
13760 cause any problems.
13764 * Remove all references to RSAref, since there's no more need for it.
13768 * Make DSO load along a path given through an environment variable
13769 (SHLIB_PATH) with shl_load().
13773 * Constify the ENGINE code as a result of BIGNUM constification.
13774 Also constify the RSA code and most things related to it. In a
13775 few places, most notable in the depth of the ASN.1 code, ugly
13776 casts back to non-const were required (to be solved at a later
13781 * Make it so the openssl application has all engines loaded by default.
13785 * Constify the BIGNUM routines a little more.
13789 * Add the following functions:
13791 ENGINE_load_cswift()
13793 ENGINE_load_atalla()
13794 ENGINE_load_nuron()
13795 ENGINE_load_builtin_engines()
13797 That way, an application can itself choose if external engines that
13798 are built-in in OpenSSL shall ever be used or not. The benefit is
13799 that applications won't have to be linked with libdl or other dso
13800 libraries unless it's really needed.
13802 Changed 'openssl engine' to load all engines on demand.
13803 Changed the engine header files to avoid the duplication of some
13804 declarations (they differed!).
13808 * 'openssl engine' can now list capabilities.
13812 * Better error reporting in 'openssl engine'.
13816 * Never call load_dh_param(NULL) in s_server.
13820 * Add engine application. It can currently list engines by name and
13821 identity, and test if they are actually available.
13825 * Improve RPM specification file by forcing symbolic linking and making
13826 sure the installed documentation is also owned by root.root.
13828 *Damien Miller <djm@mindrot.org>*
13830 * Give the OpenSSL applications more possibilities to make use of
13831 keys (public as well as private) handled by engines.
13835 * Add OCSP code that comes from CertCo.
13839 * Add VMS support for the Rijndael code.
13843 * Added untested support for Nuron crypto accelerator.
13847 * Add support for external cryptographic devices. This code was
13848 previously distributed separately as the "engine" branch.
13850 *Geoff Thorpe, Richard Levitte*
13852 * Rework the filename-translation in the DSO code. It is now possible to
13853 have far greater control over how a "name" is turned into a filename
13854 depending on the operating environment and any oddities about the
13855 different shared library filenames on each system.
13859 * Support threads on FreeBSD-elf in Configure.
13863 * Fix for SHA1 assembly problem with MASM: it produces
13864 warnings about corrupt line number information when assembling
13865 with debugging information. This is caused by the overlapping
13868 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13871 NCONF_get_number() has no error checking at all. As a replacement,
13872 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13873 promoted strongly. The old NCONF_get_number is kept around for
13874 binary backward compatibility.
13875 Make it possible for methods to load from something other than a BIO,
13876 by providing a function pointer that is given a name instead of a BIO.
13877 For example, this could be used to load configuration data from an
13882 * Fix for non blocking accept BIOs. Added new I/O special reason
13883 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13884 with non blocking I/O was not possible because no retry code was
13885 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13890 * Added the beginnings of Rijndael support.
13894 * Fix for bug in DirectoryString mask setting. Add support for
13895 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13896 to allow certificate printing to more controllable, additional
13897 'certopt' option to 'x509' to allow new printing options to be
13902 * Clean old EAY MD5 hack from e_os.h.
13906 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13908 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13909 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13911 *Joe Orton, Steve Henson*
13913 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13915 * Fix additional bug revealed by the NISCC test suite:
13917 Stop bug triggering large recursion when presented with
13918 certain ASN.1 tags ([CVE-2003-0851])
13922 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13924 * Fix various bugs revealed by running the NISCC test suite:
13926 Stop out of bounds reads in the ASN1 code when presented with
13927 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13929 If verify callback ignores invalid public key errors don't try to check
13930 certificate signature with the NULL public key.
13934 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13935 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13940 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13941 extra data after the compression methods not only for TLS 1.0
13942 but also for SSL 3.0 (as required by the specification).
13944 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13946 * Change X509_certificate_type() to mark the key as exported/exportable
13947 when it's 512 *bits* long, not 512 bytes.
13951 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13953 * Countermeasure against the Klima-Pokorny-Rosa extension of
13954 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13955 a protocol version number mismatch like a decryption error
13956 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13960 * Turn on RSA blinding by default in the default implementation
13961 to avoid a timing attack. Applications that don't want it can call
13962 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13963 They would be ill-advised to do so in most cases.
13965 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13967 * Change RSA blinding code so that it works when the PRNG is not
13968 seeded (in this case, the secret RSA exponent is abused as
13969 an unpredictable seed -- if it is not unpredictable, there
13970 is no point in blinding anyway). Make RSA blinding thread-safe
13971 by remembering the creator's thread ID in rsa->blinding and
13972 having all other threads use local one-time blinding factors
13973 (this requires more computation than sharing rsa->blinding, but
13974 avoids excessive locking; and if an RSA object is not shared
13975 between threads, blinding will still be very fast).
13979 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13981 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13982 via timing by performing a MAC computation even if incorrect
13983 block cipher padding has been found. This is a countermeasure
13984 against active attacks where the attacker has to distinguish
13985 between bad padding and a MAC verification error. ([CVE-2003-0078])
13987 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13988 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13989 Martin Vuagnoux (EPFL, Ilion)*
13991 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13993 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13994 memory from its contents. This is done with a counter that will
13995 place alternating values in each byte. This can be used to solve
13996 two issues: 1) the removal of calls to memset() by highly optimizing
13997 compilers, and 2) cleansing with other values than 0, since those can
13998 be read through on certain media, for example a swap space on disk.
14002 * Bugfix: client side session caching did not work with external caching,
14003 because the session->cipher setting was not restored when reloading
14004 from the external cache. This problem was masked, when
14005 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14006 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14010 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14011 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14013 *Zeev Lieber <zeev-l@yahoo.com>*
14015 * Undo an undocumented change introduced in 0.9.6e which caused
14016 repeated calls to OpenSSL_add_all_ciphers() and
14017 OpenSSL_add_all_digests() to be ignored, even after calling
14022 * Change the default configuration reader to deal with last line not
14023 being properly terminated.
14027 * Change X509_NAME_cmp() so it applies the special rules on handling
14028 DN values that are of type PrintableString, as well as RDNs of type
14029 emailAddress where the value has the type ia5String.
14031 *stefank@valicert.com via Richard Levitte*
14033 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14034 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14035 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14036 the bitwise-OR of the two for use by the majority of applications
14037 wanting this behaviour, and update the docs. The documented
14038 behaviour and actual behaviour were inconsistent and had been
14039 changing anyway, so this is more a bug-fix than a behavioural
14042 *Geoff Thorpe, diagnosed by Nadav Har'El*
14044 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14045 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14049 * Fix initialization code race conditions in
14050 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14051 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14052 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14053 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14054 ssl2_get_cipher_by_char(),
14055 ssl3_get_cipher_by_char().
14057 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14059 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14060 the cached sessions are flushed, as the remove_cb() might use ex_data
14061 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14062 (see [openssl.org #212]).
14064 *Geoff Thorpe, Lutz Jaenicke*
14066 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14067 length, instead of the encoding length to d2i_ASN1_OBJECT.
14071 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14073 * [In 0.9.6g-engine release:]
14074 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14076 *Lynn Gazis <lgazis@rainbow.com>*
14078 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14080 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14081 and get fix the header length calculation.
14082 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14083 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14085 * Use proper error handling instead of 'assertions' in buffer
14086 overflow checks added in 0.9.6e. This prevents DoS (the
14087 assertions could call abort()).
14089 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14091 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14093 * Add various sanity checks to asn1_get_length() to reject
14094 the ASN1 length bytes if they exceed sizeof(long), will appear
14095 negative or the content length exceeds the length of the
14098 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14100 * Fix cipher selection routines: ciphers without encryption had no flags
14101 for the cipher strength set and where therefore not handled correctly
14102 by the selection routines (PR #130).
14106 * Fix EVP_dsa_sha macro.
14111 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14112 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14113 that was added in OpenSSL 0.9.6d.
14115 As the countermeasure turned out to be incompatible with some
14116 broken SSL implementations, the new option is part of SSL_OP_ALL.
14117 SSL_OP_ALL is usually employed when compatibility with weird SSL
14118 implementations is desired (e.g. '-bugs' option to 's_client' and
14119 's_server'), so the new option is automatically set in many
14124 * Changes in security patch:
14126 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14127 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14128 Air Force Materiel Command, USAF, under agreement number
14131 * Add various sanity checks to asn1_get_length() to reject
14132 the ASN1 length bytes if they exceed sizeof(long), will appear
14133 negative or the content length exceeds the length of the
14134 supplied buffer. ([CVE-2002-0659])
14136 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14138 * Assertions for various potential buffer overflows, not known to
14139 happen in practice.
14141 *Ben Laurie (CHATS)*
14143 * Various temporary buffers to hold ASCII versions of integers were
14144 too small for 64 bit platforms. ([CVE-2002-0655])
14145 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14147 * Remote buffer overflow in SSL3 protocol - an attacker could
14148 supply an oversized session ID to a client. ([CVE-2002-0656])
14150 *Ben Laurie (CHATS)*
14152 * Remote buffer overflow in SSL2 protocol - an attacker could
14153 supply an oversized client master key. ([CVE-2002-0656])
14155 *Ben Laurie (CHATS)*
14157 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14159 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14160 encoded as NULL) with id-dsa-with-sha1.
14162 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14164 * Check various `X509_...()` return values in `apps/req.c`.
14166 *Nils Larsch <nla@trustcenter.de>*
14168 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14169 an end-of-file condition would erroneously be flagged, when the CRLF
14170 was just at the end of a processed block. The bug was discovered when
14171 processing data through a buffering memory BIO handing the data to a
14172 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14173 <ptsekov@syntrex.com> and Nedelcho Stanev.
14177 * Implement a countermeasure against a vulnerability recently found
14178 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14179 before application data chunks to avoid the use of known IVs
14180 with data potentially chosen by the attacker.
14184 * Fix length checks in ssl3_get_client_hello().
14188 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14189 to prevent ssl3_read_internal() from incorrectly assuming that
14190 ssl3_read_bytes() found application data while handshake
14191 processing was enabled when in fact s->s3->in_read_app_data was
14192 merely automatically cleared during the initial handshake.
14194 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14196 * Fix object definitions for Private and Enterprise: they were not
14197 recognized in their shortname (=lowercase) representation. Extend
14198 obj_dat.pl to issue an error when using undefined keywords instead
14199 of silently ignoring the problem (Svenning Sorensen
14200 <sss@sss.dnsalias.net>).
14204 * Fix DH_generate_parameters() so that it works for 'non-standard'
14205 generators, i.e. generators other than 2 and 5. (Previously, the
14206 code did not properly initialise the 'add' and 'rem' values to
14207 BN_generate_prime().)
14209 In the new general case, we do not insist that 'generator' is
14210 actually a primitive root: This requirement is rather pointless;
14211 a generator of the order-q subgroup is just as good, if not
14216 * Map new X509 verification errors to alerts. Discovered and submitted by
14217 Tom Wu <tom@arcot.com>.
14221 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14222 returning non-zero before the data has been completely received
14223 when using non-blocking I/O.
14225 *Bodo Moeller; problem pointed out by John Hughes*
14227 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14229 *Ben Laurie, Lutz Jaenicke*
14231 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14232 Yoram Zahavi <YoramZ@gilian.com>).
14236 * Add information about CygWin 1.3 and on, and preserve proper
14237 configuration for the versions before that.
14239 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14241 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14242 check whether we deal with a copy of a session and do not delete from
14243 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14244 <izhar@checkpoint.com>.
14248 * Do not store session data into the internal session cache, if it
14249 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14250 flag is set). Proposed by Aslam <aslam@funk.com>.
14254 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14259 * [In 0.9.6d-engine release:]
14260 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14262 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14264 * Add the configuration target linux-s390x.
14266 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14268 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14269 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14270 variable as an indication that a ClientHello message has been
14271 received. As the flag value will be lost between multiple
14272 invocations of ssl3_accept when using non-blocking I/O, the
14273 function may not be aware that a handshake has actually taken
14274 place, thus preventing a new session from being added to the
14277 To avoid this problem, we now set s->new_session to 2 instead of
14278 using a local variable.
14280 *Lutz Jaenicke, Bodo Moeller*
14282 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14283 if the SSL_R_LENGTH_MISMATCH error is detected.
14285 *Geoff Thorpe, Bodo Moeller*
14287 * New 'shared_ldflag' column in Configure platform table.
14291 * Fix EVP_CIPHER_mode macro.
14293 *"Dan S. Camper" <dan@bti.net>*
14295 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14296 type, we must throw them away by setting rr->length to 0.
14298 *D P Chang <dpc@qualys.com>*
14300 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14302 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14303 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14304 worked incorrectly for those cases where range = `10..._2` and
14305 `3*range` is two bits longer than range.)
14309 * Only add signing time to PKCS7 structures if it is not already
14314 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14315 OBJ_ld_ce should be OBJ_id_ce.
14316 Also some ip-pda OIDs in crypto/objects/objects.txt were
14317 incorrect (cf. RFC 3039).
14319 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14321 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14322 returns early because it has nothing to do.
14324 *Andy Schneider <andy.schneider@bjss.co.uk>*
14326 * [In 0.9.6c-engine release:]
14327 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14329 *Andy Schneider <andy.schneider@bjss.co.uk>*
14331 * [In 0.9.6c-engine release:]
14332 Add support for Cryptographic Appliance's keyserver technology.
14333 (Use engine 'keyclient')
14335 *Cryptographic Appliances and Geoff Thorpe*
14337 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14338 is called via tools/c89.sh because arguments have to be
14339 rearranged (all '-L' options must appear before the first object
14342 *Richard Shapiro <rshapiro@abinitio.com>*
14344 * [In 0.9.6c-engine release:]
14345 Add support for Broadcom crypto accelerator cards, backported
14348 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14350 * [In 0.9.6c-engine release:]
14351 Add support for SureWare crypto accelerator cards from
14352 Baltimore Technologies. (Use engine 'sureware')
14354 *Baltimore Technologies and Mark Cox*
14356 * [In 0.9.6c-engine release:]
14357 Add support for crypto accelerator cards from Accelerated
14358 Encryption Processing, www.aep.ie. (Use engine 'aep')
14360 *AEP Inc. and Mark Cox*
14362 * Add a configuration entry for gcc on UnixWare.
14364 *Gary Benson <gbenson@redhat.com>*
14366 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14367 messages are stored in a single piece (fixed-length part and
14368 variable-length part combined) and fix various bugs found on the way.
14372 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14373 instead. BIO_gethostbyname() does not know what timeouts are
14374 appropriate, so entries would stay in cache even when they have
14376 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14378 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14379 faced with a pathologically small ClientHello fragment that does
14380 not contain client_version: Instead of aborting with an error,
14381 simply choose the highest available protocol version (i.e.,
14382 TLS 1.0 unless it is disabled). In practice, ClientHello
14383 messages are never sent like this, but this change gives us
14384 strictly correct behaviour at least for TLS.
14388 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14389 never resets s->method to s->ctx->method when called from within
14390 one of the SSL handshake functions.
14392 *Bodo Moeller; problem pointed out by Niko Baric*
14394 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14395 (sent using the client's version number) if client_version is
14396 smaller than the protocol version in use. Also change
14397 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14398 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14399 the client will at least see that alert.
14403 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14408 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14409 client receives HelloRequest while in a handshake.
14411 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14413 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14414 should end in 'break', not 'goto end' which circumvents various
14415 cleanups done in state SSL_ST_OK. But session related stuff
14416 must be disabled for SSL_ST_OK in the case that we just sent a
14419 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14420 before just sending a HelloRequest.
14422 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14424 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14425 reveal whether illegal block cipher padding was found or a MAC
14426 verification error occurred. (Neither SSLerr() codes nor alerts
14427 are directly visible to potential attackers, but the information
14428 may leak via logfiles.)
14430 Similar changes are not required for the SSL 2.0 implementation
14431 because the number of padding bytes is sent in clear for SSL 2.0,
14432 and the extra bytes are just ignored. However ssl/s2_pkt.c
14433 failed to verify that the purported number of padding bytes is in
14438 * Add OpenUNIX-8 support including shared libraries
14439 (Boyd Lynn Gerber <gerberb@zenez.com>).
14443 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14444 'wristwatch attack' using huge encoding parameters (cf.
14445 James H. Manger's CRYPTO 2001 paper). Note that the
14446 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14447 encoding parameters and hence was not vulnerable.
14451 * BN_sqr() bug fix.
14453 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14455 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14456 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14457 followed by modular reduction.
14459 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14461 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14462 equivalent based on BN_pseudo_rand() instead of BN_rand().
14466 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14467 This function was broken, as the check for a new client hello message
14468 to handle SGC did not allow these large messages.
14469 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14473 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14477 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14478 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14482 * Rework the configuration and shared library support for Tru64 Unix.
14483 The configuration part makes use of modern compiler features and
14484 still retains old compiler behavior for those that run older versions
14485 of the OS. The shared library support part includes a variant that
14486 uses the RPATH feature, and is available through the special
14487 configuration target "alpha-cc-rpath", which will never be selected
14490 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14492 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14493 with the same message size as in ssl3_get_certificate_request().
14494 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14495 messages might inadvertently be reject as too long.
14497 *Petr Lampa <lampa@fee.vutbr.cz>*
14499 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14503 * Modified SSL library such that the verify_callback that has been set
14504 specifically for an SSL object with SSL_set_verify() is actually being
14505 used. Before the change, a verify_callback set with this function was
14506 ignored and the verify_callback() set in the SSL_CTX at the time of
14507 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14508 to allow the necessary settings.
14512 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14513 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14514 done automatically (in contradiction to the requirements of the C
14515 standard). This made problems when used from OpenSSH.
14519 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14520 dh->length and always used
14522 BN_rand_range(priv_key, dh->p).
14524 BN_rand_range() is not necessary for Diffie-Hellman, and this
14525 specific range makes Diffie-Hellman unnecessarily inefficient if
14526 dh->length (recommended exponent length) is much smaller than the
14527 length of dh->p. We could use BN_rand_range() if the order of
14528 the subgroup was stored in the DH structure, but we only have
14533 BN_rand(priv_key, l, ...)
14535 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14542 RSA_eay_public_encrypt
14543 RSA_eay_private_decrypt
14544 RSA_eay_private_encrypt (signing)
14545 RSA_eay_public_decrypt (signature verification)
14547 (default implementations for RSA_public_encrypt,
14548 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14549 always reject numbers >= n.
14553 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14554 to synchronize access to 'locking_thread'. This is necessary on
14555 systems where access to 'locking_thread' (an 'unsigned long'
14556 variable) is not atomic.
14560 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14561 *before* setting the 'crypto_lock_rand' flag. The previous code had
14562 a race condition if 0 is a valid thread ID.
14564 *Travis Vitek <vitek@roguewave.com>*
14566 * Add support for shared libraries under Irix.
14568 *Albert Chin-A-Young <china@thewrittenword.com>*
14570 * Add configuration option to build on Linux on both big-endian and
14571 little-endian MIPS.
14573 *Ralf Baechle <ralf@uni-koblenz.de>*
14575 * Add the possibility to create shared libraries on HP-UX.
14579 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14581 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14582 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14583 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14584 PRNG state recovery was possible based on the output of
14585 one PRNG request appropriately sized to gain knowledge on
14586 'md' followed by enough consecutive 1-byte PRNG requests
14587 to traverse all of 'state'.
14589 1. When updating 'md_local' (the current thread's copy of 'md')
14590 during PRNG output generation, hash all of the previous
14591 'md_local' value, not just the half used for PRNG output.
14593 2. Make the number of bytes from 'state' included into the hash
14594 independent from the number of PRNG bytes requested.
14596 The first measure alone would be sufficient to avoid
14597 Markku-Juhani's attack. (Actually it had never occurred
14598 to me that the half of 'md_local' used for chaining was the
14599 half from which PRNG output bytes were taken -- I had always
14600 assumed that the secret half would be used.) The second
14601 measure makes sure that additional data from 'state' is never
14602 mixed into 'md_local' in small portions; this heuristically
14603 further strengthens the PRNG.
14607 * Fix crypto/bn/asm/mips3.s.
14611 * When only the key is given to "enc", the IV is undefined. Print out
14612 an error message in this case.
14616 * Handle special case when X509_NAME is empty in X509 printing routines.
14620 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14621 positive and less than q.
14625 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14626 used: it isn't thread safe and the add_lock_callback should handle
14629 *Paul Rose <Paul.Rose@bridge.com>*
14631 * Verify that incoming data obeys the block size in
14632 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14638 *Ulf Möller, Bodo Möller*
14640 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14641 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14642 when fixing the server behaviour for backwards-compatible 'client
14643 hello' messages. (Note that the attack is impractical against
14644 SSL 3.0 and TLS 1.0 anyway because length and version checking
14645 means that the probability of guessing a valid ciphertext is
14646 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14649 Before 0.9.5, the countermeasure (hide the error by generating a
14650 random 'decryption result') did not work properly because
14651 ERR_clear_error() was missing, meaning that SSL_get_error() would
14652 detect the supposedly ignored error.
14654 Both problems are now fixed.
14658 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14659 (previously it was 1024).
14663 * Fix for compatibility mode trust settings: ignore trust settings
14664 unless some valid trust or reject settings are present.
14668 * Fix for blowfish EVP: its a variable length cipher.
14672 * Fix various bugs related to DSA S/MIME verification. Handle missing
14673 parameters in DSA public key structures and return an error in the
14674 DSA routines if parameters are absent.
14678 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14679 in the current directory if neither $RANDFILE nor $HOME was set.
14680 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14681 caused some confusion to Windows users who haven't defined $HOME.
14682 Thus RAND_file_name() is changed again: e_os.h can define a
14683 DEFAULT_HOME, which will be used if $HOME is not set.
14684 For Windows, we use "C:"; on other platforms, we still require
14685 environment variables.
14687 * Move 'if (!initialized) RAND_poll()' into regions protected by
14688 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14689 having multiple threads call RAND_poll() concurrently.
14693 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14694 combination of a flag and a thread ID variable.
14695 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14696 flag), *other* threads can enter ssleay_add_bytes without obeying
14697 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14698 that they do not hold after the first thread unsets add_do_not_lock).
14702 * Change bctest again: '-x' expressions are not available in all
14703 versions of 'test'.
14707 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14709 * Fix a couple of memory leaks in PKCS7_dataDecode()
14711 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14713 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14714 the default extension for executables, if any. Also, make the perl
14715 scripts that use symlink() to test if it really exists and use "cp"
14716 if it doesn't. All this made OpenSSL compilable and installable in
14721 * Fix for asn1_GetSequence() for indefinite length constructed data.
14722 If SEQUENCE is length is indefinite just set c->slen to the total
14723 amount of data available.
14725 *Steve Henson, reported by shige@FreeBSD.org*
14727 *This change does not apply to 0.9.7.*
14729 * Change bctest to avoid here-documents inside command substitution
14730 (workaround for FreeBSD /bin/sh bug).
14731 For compatibility with Ultrix, avoid shell functions (introduced
14732 in the bctest version that searches along $PATH).
14736 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14737 with des_encrypt() defined on some operating systems, like Solaris
14742 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14743 On the Importance of Eliminating Errors in Cryptographic
14744 Computations, J. Cryptology 14 (2001) 2, 101-119,
14745 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14749 * MIPS assembler BIGNUM division bug fix.
14753 * Disabled incorrect Alpha assembler code.
14757 * Fix PKCS#7 decode routines so they correctly update the length
14758 after reading an EOC for the EXPLICIT tag.
14762 *This change does not apply to 0.9.7.*
14764 * Fix bug in PKCS#12 key generation routines. This was triggered
14765 if a 3DES key was generated with a 0 initial byte. Include
14766 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14767 (but broken) behaviour.
14771 * Enhance bctest to search for a working bc along $PATH and print
14774 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14776 * Fix memory leaks in err.c: free err_data string if necessary;
14777 don't write to the wrong index in ERR_set_error_data.
14781 * Implement ssl23_peek (analogous to ssl23_read), which previously
14786 * Replace rdtsc with `_emit` statements for VC++ version 5.
14788 *Jeremy Cooper <jeremy@baymoo.org>*
14790 * Make it possible to reuse SSLv2 sessions.
14794 * In copy_email() check for >= 0 as a return value for
14795 X509_NAME_get_index_by_NID() since 0 is a valid index.
14797 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14799 * Avoid coredump with unsupported or invalid public keys by checking if
14800 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14801 PKCS7_verify() fails with non detached data.
14805 * Don't use getenv in library functions when run as setuid/setgid.
14806 New function OPENSSL_issetugid().
14810 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14811 due to incorrect handling of multi-threading:
14813 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14815 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14817 3. Count how many times MemCheck_off() has been called so that
14818 nested use can be treated correctly. This also avoids
14819 inband-signalling in the previous code (which relied on the
14820 assumption that thread ID 0 is impossible).
14824 * Add "-rand" option also to s_client and s_server.
14828 * Fix CPU detection on Irix 6.x.
14829 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14830 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14832 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14837 *This change does not apply to 0.9.7.*
14839 * Use the cached encoding of an X509_NAME structure rather than
14840 copying it. This is apparently the reason for the libsafe "errors"
14841 but the code is actually correct.
14845 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14846 Bleichenbacher's DSA attack.
14847 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14848 to be set and top=0 forces the highest bit to be set; top=-1 is new
14849 and leaves the highest bit random.
14851 *Ulf Moeller, Bodo Moeller*
14853 * In the `NCONF_...`-based implementations for `CONF_...` queries
14854 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14855 a temporary CONF structure with the data component set to NULL
14856 (which gives segmentation faults in lh_retrieve).
14857 Instead, use NULL for the CONF pointer in CONF_get_string and
14858 CONF_get_number (which may use environment variables) and directly
14859 return NULL from CONF_get_section.
14863 * Fix potential buffer overrun for EBCDIC.
14867 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14868 keyUsage if basicConstraints absent for a CA.
14872 * Make SMIME_write_PKCS7() write mail header values with a format that
14873 is more generally accepted (no spaces before the semicolon), since
14874 some programs can't parse those values properly otherwise. Also make
14875 sure BIO's that break lines after each write do not create invalid
14880 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14881 macros previously used would not encode an empty SEQUENCE OF
14882 and break the signature.
14886 *This change does not apply to 0.9.7.*
14888 * Zero the premaster secret after deriving the master secret in
14893 * Add some EVP_add_digest_alias registrations (as found in
14894 OpenSSL_add_all_digests()) to SSL_library_init()
14895 aka OpenSSL_add_ssl_algorithms(). This provides improved
14896 compatibility with peers using X.509 certificates
14897 with unconventional AlgorithmIdentifier OIDs.
14901 * Fix for Irix with NO_ASM.
14903 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14905 * ./config script fixes.
14907 *Ulf Moeller, Richard Levitte*
14909 * Fix 'openssl passwd -1'.
14913 * Change PKCS12_key_gen_asc() so it can cope with non null
14914 terminated strings whose length is passed in the passlen
14915 parameter, for example from PEM callbacks. This was done
14916 by adding an extra length parameter to asc2uni().
14918 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14920 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14921 call failed, free the DSA structure.
14925 * Fix to uni2asc() to cope with zero length Unicode strings.
14926 These are present in some PKCS#12 files.
14930 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14931 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14932 when writing a 32767 byte record.
14934 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14936 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14937 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14939 (RSA objects have a reference count access to which is protected
14940 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14941 so they are meant to be shared between threads.)
14942 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14943 "Reddie, Steven" <Steven.Reddie@ca.com>*
14945 * Fix a deadlock in CRYPTO_mem_leaks().
14949 * Use better test patterns in bntest.
14953 * rand_win.c fix for Borland C.
14957 * BN_rshift bugfix for n == 0.
14961 * Add a 'bctest' script that checks for some known 'bc' bugs
14962 so that 'make test' does not abort just because 'bc' is broken.
14966 * Store verify_result within SSL_SESSION also for client side to
14967 avoid potential security hole. (Re-used sessions on the client side
14968 always resulted in verify_result==X509_V_OK, not using the original
14969 result of the server certificate verification.)
14973 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14974 SSL3_RT_APPLICATION_DATA, return 0.
14975 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14980 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14981 releases, have been re-implemented by renaming the previous
14982 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14983 and ssl3_read_internal, respectively, and adding 'peek' parameters
14984 to them. The new ssl[23]_{read,peek} functions are calls to
14985 ssl[23]_read_internal with the 'peek' flag set appropriately.
14986 A 'peek' parameter has also been added to ssl3_read_bytes, which
14987 does the actual work for ssl3_read_internal.
14991 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14992 the method-specific "init()" handler. Also clean up ex_data after
14993 calling the method-specific "finish()" handler. Previously, this was
14994 happening the other way round.
14998 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14999 The previous value, 12, was not always sufficient for BN_mod_exp().
15003 * Make sure that shared libraries get the internal name engine with
15004 the full version number and not just 0. This should mark the
15005 shared libraries as not backward compatible. Of course, this should
15006 be changed again when we can guarantee backward binary compatibility.
15010 * Fix typo in get_cert_by_subject() in by_dir.c
15012 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15014 * Rework the system to generate shared libraries:
15016 - Make note of the expected extension for the shared libraries and
15017 if there is a need for symbolic links from for example libcrypto.so.0
15018 to libcrypto.so.0.9.7. There is extended info in Configure for
15021 - Make as few rebuilds of the shared libraries as possible.
15023 - Still avoid linking the OpenSSL programs with the shared libraries.
15025 - When installing, install the shared libraries separately from the
15030 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15032 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15033 and not in SSL_clear because the latter is also used by the
15034 accept/connect functions; previously, the settings made by
15035 SSL_set_read_ahead would be lost during the handshake.
15037 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15039 * Correct util/mkdef.pl to be selective about disabled algorithms.
15040 Previously, it would create entries for disabled algorithms no
15045 * Added several new manual pages for SSL_* function.
15049 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15051 * In ssl23_get_client_hello, generate an error message when faced
15052 with an initial SSL 3.0/TLS record that is too small to contain the
15053 first two bytes of the ClientHello message, i.e. client_version.
15054 (Note that this is a pathologic case that probably has never happened
15055 in real life.) The previous approach was to use the version number
15056 from the record header as a substitute; but our protocol choice
15057 should not depend on that one because it is not authenticated
15058 by the Finished messages.
15062 * More robust randomness gathering functions for Windows.
15064 *Jeffrey Altman <jaltman@columbia.edu>*
15066 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15067 not set then we don't setup the error code for issuer check errors
15068 to avoid possibly overwriting other errors which the callback does
15069 handle. If an application does set the flag then we assume it knows
15070 what it is doing and can handle the new informational codes
15075 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15076 a general "ANY" type, as such it should be able to decode anything
15077 including tagged types. However it didn't check the class so it would
15078 wrongly interpret tagged types in the same way as their universal
15079 counterpart and unknown types were just rejected. Changed so that the
15080 tagged and unknown types are handled in the same way as a SEQUENCE:
15081 that is the encoding is stored intact. There is also a new type
15082 "V_ASN1_OTHER" which is used when the class is not universal, in this
15083 case we have no idea what the actual type is so we just lump them all
15088 * On VMS, stdout may very well lead to a file that is written to
15089 in a record-oriented fashion. That means that every write() will
15090 write a separate record, which will be read separately by the
15091 programs trying to read from it. This can be very confusing.
15093 The solution is to put a BIO filter in the way that will buffer
15094 text until a linefeed is reached, and then write everything a
15095 line at a time, so every record written will be an actual line,
15096 not chunks of lines and not (usually doesn't happen, but I've
15097 seen it once) several lines in one record. BIO_f_linebuffer() is
15100 Currently, it's a VMS-only method, because that's where it has
15101 been tested well enough.
15105 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15106 it can return incorrect results.
15107 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15108 but it was in 0.9.6-beta[12].)
15112 * Disable the check for content being present when verifying detached
15113 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15114 include zero length content when signing messages.
15118 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15119 BIO_ctrl (for BIO pairs).
15123 * Add DSO method for VMS.
15127 * Bug fix: Montgomery multiplication could produce results with the
15132 * Add RPM specification openssl.spec and modify it to build three
15133 packages. The default package contains applications, application
15134 documentation and run-time libraries. The devel package contains
15135 include files, static libraries and function documentation. The
15136 doc package contains the contents of the doc directory. The original
15137 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15141 * Add a large number of documentation files for many SSL routines.
15143 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15145 * Add a configuration entry for Sony News 4.
15147 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15149 * Don't set the two most significant bits to one when generating a
15150 random number < q in the DSA library.
15154 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15155 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15156 the underlying transport is blocking) if a handshake took place.
15157 (The default behaviour is needed by applications such as s_client
15158 and s_server that use select() to determine when to use SSL_read;
15159 but for applications that know in advance when to expect data, it
15160 just makes things more complicated.)
15164 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15169 * Add a few more EBCDIC conditionals that make `req` and `x509`
15170 work better on such systems.
15172 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15174 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15175 Update PKCS12_parse() so it copies the friendlyName and the
15176 keyid to the certificates aux info.
15180 * Fix bug in PKCS7_verify() which caused an infinite loop
15181 if there was more than one signature.
15183 *Sven Uszpelkat <su@celocom.de>*
15185 * Major change in util/mkdef.pl to include extra information
15186 about each symbol, as well as presenting variables as well
15187 as functions. This change means that there's n more need
15188 to rebuild the .num files when some algorithms are excluded.
15192 * Allow the verify time to be set by an application,
15193 rather than always using the current time.
15197 * Phase 2 verify code reorganisation. The certificate
15198 verify code now looks up an issuer certificate by a
15199 number of criteria: subject name, authority key id
15200 and key usage. It also verifies self signed certificates
15201 by the same criteria. The main comparison function is
15202 X509_check_issued() which performs these checks.
15204 Lot of changes were necessary in order to support this
15205 without completely rewriting the lookup code.
15207 Authority and subject key identifier are now cached.
15209 The LHASH 'certs' is X509_STORE has now been replaced
15210 by a STACK_OF(X509_OBJECT). This is mainly because an
15211 LHASH can't store or retrieve multiple objects with
15212 the same hash value.
15214 As a result various functions (which were all internal
15215 use only) have changed to handle the new X509_STORE
15216 structure. This will break anything that messed round
15217 with X509_STORE internally.
15219 The functions X509_STORE_add_cert() now checks for an
15220 exact match, rather than just subject name.
15222 The X509_STORE API doesn't directly support the retrieval
15223 of multiple certificates matching a given criteria, however
15224 this can be worked round by performing a lookup first
15225 (which will fill the cache with candidate certificates)
15226 and then examining the cache for matches. This is probably
15227 the best we can do without throwing out X509_LOOKUP
15228 entirely (maybe later...).
15230 The X509_VERIFY_CTX structure has been enhanced considerably.
15232 All certificate lookup operations now go via a get_issuer()
15233 callback. Although this currently uses an X509_STORE it
15234 can be replaced by custom lookups. This is a simple way
15235 to bypass the X509_STORE hackery necessary to make this
15236 work and makes it possible to use more efficient techniques
15237 in future. A very simple version which uses a simple
15238 STACK for its trusted certificate store is also provided
15239 using X509_STORE_CTX_trusted_stack().
15241 The verify_cb() and verify() callbacks now have equivalents
15242 in the X509_STORE_CTX structure.
15244 X509_STORE_CTX also has a 'flags' field which can be used
15245 to customise the verify behaviour.
15249 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15250 excludes S/MIME capabilities.
15254 * When a certificate request is read in keep a copy of the
15255 original encoding of the signed data and use it when outputting
15256 again. Signatures then use the original encoding rather than
15257 a decoded, encoded version which may cause problems if the
15258 request is improperly encoded.
15262 * For consistency with other BIO_puts implementations, call
15263 buffer_write(b, ...) directly in buffer_puts instead of calling
15266 In BIO_puts, increment b->num_write as in BIO_write.
15268 *Peter.Sylvester@EdelWeb.fr*
15270 * Fix BN_mul_word for the case where the word is 0. (We have to use
15271 BN_zero, we may not return a BIGNUM with an array consisting of
15272 words set to zero.)
15276 * Avoid calling abort() from within the library when problems are
15277 detected, except if preprocessor symbols have been defined
15278 (such as REF_CHECK, BN_DEBUG etc.).
15282 * New openssl application 'rsautl'. This utility can be
15283 used for low level RSA operations. DER public key
15284 BIO/fp routines also added.
15288 * New Configure entry and patches for compiling on QNX 4.
15290 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15292 * A demo state-machine implementation was sponsored by
15293 Nuron (<http://www.nuron.com/>) and is now available in
15294 demos/state_machine.
15298 * New options added to the 'dgst' utility for signature
15299 generation and verification.
15303 * Unrecognized PKCS#7 content types are now handled via a
15304 catch all ASN1_TYPE structure. This allows unsupported
15305 types to be stored as a "blob" and an application can
15306 encode and decode it manually.
15310 * Fix various signed/unsigned issues to make a_strex.c
15311 compile under VC++.
15313 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15315 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15316 length if passed a buffer. ASN1_INTEGER_to_BN failed
15317 if passed a NULL BN and its argument was negative.
15319 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15321 * Modification to PKCS#7 encoding routines to output definite
15322 length encoding. Since currently the whole structures are in
15323 memory there's not real point in using indefinite length
15324 constructed encoding. However if OpenSSL is compiled with
15325 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15329 * Added BIO_vprintf() and BIO_vsnprintf().
15333 * Added more prefixes to parse for in the strings written
15334 through a logging bio, to cover all the levels that are available
15335 through syslog. The prefixes are now:
15337 PANIC, EMERG, EMR => LOG_EMERG
15338 ALERT, ALR => LOG_ALERT
15339 CRIT, CRI => LOG_CRIT
15340 ERROR, ERR => LOG_ERR
15341 WARNING, WARN, WAR => LOG_WARNING
15342 NOTICE, NOTE, NOT => LOG_NOTICE
15343 INFO, INF => LOG_INFO
15344 DEBUG, DBG => LOG_DEBUG
15346 and as before, if none of those prefixes are present at the
15347 beginning of the string, LOG_ERR is chosen.
15349 On Win32, the `LOG_*` levels are mapped according to this:
15351 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15352 LOG_WARNING => EVENTLOG_WARNING_TYPE
15353 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15357 * Made it possible to reconfigure with just the configuration
15358 argument "reconf" or "reconfigure". The command line arguments
15359 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15360 and are retrieved from there when reconfiguring.
15366 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15368 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15372 * The obj_dat.pl script was messing up the sorting of object
15373 names. The reason was that it compared the quoted version
15374 of strings as a result "OCSP" > "OCSP Signing" because
15375 " > SPACE. Changed script to store unquoted versions of
15376 names and add quotes on output. It was also omitting some
15377 names from the lookup table if they were given a default
15378 value (that is if SN is missing it is given the same
15379 value as LN and vice versa), these are now added on the
15380 grounds that if an object has a name we should be able to
15381 look it up. Finally added warning output when duplicate
15382 short or long names are found.
15386 * Changes needed for Tandem NSK.
15388 *Scott Uroff <scott@xypro.com>*
15390 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15391 RSA_padding_check_SSLv23(), special padding was never detected
15392 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15393 version rollback attacks was not effective.
15395 In s23_clnt.c, don't use special rollback-attack detection padding
15396 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15397 client; similarly, in s23_srvr.c, don't do the rollback check if
15398 SSL 2.0 is the only protocol enabled in the server.
15402 * Make it possible to get hexdumps of unprintable data with 'openssl
15403 asn1parse'. By implication, the functions ASN1_parse_dump() and
15404 BIO_dump_indent() are added.
15408 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15409 these print out strings and name structures based on various
15410 flags including RFC2253 support and proper handling of
15411 multibyte characters. Added options to the 'x509' utility
15412 to allow the various flags to be set.
15416 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15417 Also change the functions X509_cmp_current_time() and
15418 X509_gmtime_adj() work with an ASN1_TIME structure,
15419 this will enable certificates using GeneralizedTime in validity
15420 dates to be checked.
15424 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15425 negative public key encodings) on by default,
15426 NO_NEG_PUBKEY_BUG can be set to disable it.
15430 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15431 content octets. An i2c_ASN1_OBJECT is unnecessary because
15432 the encoding can be trivially obtained from the structure.
15436 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15437 not read locks (`CRYPTO_r_[un]lock`).
15441 * A first attempt at creating official support for shared
15442 libraries through configuration. I've kept it so the
15443 default is static libraries only, and the OpenSSL programs
15444 are always statically linked for now, but there are
15445 preparations for dynamic linking in place.
15446 This has been tested on Linux and Tru64.
15450 * Randomness polling function for Win9x, as described in:
15451 Peter Gutmann, Software Generation of Practically Strong
15456 * Fix so PRNG is seeded in req if using an already existing
15461 * New options to smime application. -inform and -outform
15462 allow alternative formats for the S/MIME message including
15463 PEM and DER. The -content option allows the content to be
15464 specified separately. This should allow things like Netscape
15465 form signing output easier to verify.
15469 * Fix the ASN1 encoding of tags using the 'long form'.
15473 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15474 STRING types. These convert content octets to and from the
15475 underlying type. The actual tag and length octets are
15476 already assumed to have been read in and checked. These
15477 are needed because all other string types have virtually
15478 identical handling apart from the tag. By having versions
15479 of the ASN1 functions that just operate on content octets
15480 IMPLICIT tagging can be handled properly. It also allows
15481 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15482 and ASN1_INTEGER are identical apart from the tag.
15486 * Change the handling of OID objects as follows:
15488 - New object identifiers are inserted in objects.txt, following
15489 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15490 - objects.pl is used to process obj_mac.num and create a new
15492 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15495 This is currently kind of a hack, and the perl code in objects.pl
15496 isn't very elegant, but it works as I intended. The simplest way
15497 to check that it worked correctly is to look in obj_dat.h and
15498 check the array nid_objs and make sure the objects haven't moved
15499 around (this is important!). Additions are OK, as well as
15500 consistent name changes.
15504 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15508 * Addition of the command line parameter '-rand file' to 'openssl req'.
15509 The given file adds to whatever has already been seeded into the
15510 random pool through the RANDFILE configuration file option or
15511 environment variable, or the default random state file.
15515 * mkstack.pl now sorts each macro group into lexical order.
15516 Previously the output order depended on the order the files
15517 appeared in the directory, resulting in needless rewriting
15522 * Patches to make OpenSSL compile under Win32 again. Mostly
15523 work arounds for the VC++ problem that it treats func() as
15524 func(void). Also stripped out the parts of mkdef.pl that
15525 added extra typesafe functions: these no longer exist.
15529 * Reorganisation of the stack code. The macros are now all
15530 collected in safestack.h . Each macro is defined in terms of
15531 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15532 DEBUG_SAFESTACK is now handled in terms of function casts,
15533 this has the advantage of retaining type safety without the
15534 use of additional functions. If DEBUG_SAFESTACK is not defined
15535 then the non typesafe macros are used instead. Also modified the
15536 mkstack.pl script to handle the new form. Needs testing to see
15537 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15538 the default if no major problems. Similar behaviour for ASN1_SET_OF
15539 and PKCS12_STACK_OF.
15543 * When some versions of IIS use the 'NET' form of private key the
15544 key derivation algorithm is different. Normally MD5(password) is
15545 used as a 128 bit RC4 key. In the modified case
15546 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15547 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15548 as the old Netscape_RSA functions except they have an additional
15549 'sgckey' parameter which uses the modified algorithm. Also added
15550 an -sgckey command line option to the rsa utility. Thanks to
15551 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15552 algorithm to openssl-dev.
15556 * The evp_local.h macros were using 'c.##kname' which resulted in
15557 invalid expansion on some systems (SCO 5.0.5 for example).
15558 Corrected to 'c.kname'.
15560 *Phillip Porch <root@theporch.com>*
15562 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15563 a STACK of email addresses from a certificate or request, these look
15564 in the subject name and the subject alternative name extensions and
15565 omit any duplicate addresses.
15569 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15570 This makes DSA verification about 2 % faster.
15574 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15575 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15576 plus overhead for 1024 bit moduli).
15577 This makes exponentiations about 0.5 % faster for 1024 bit
15578 exponents (as measured by "openssl speed rsa2048").
15582 * Rename memory handling macros to avoid conflicts with other
15584 Malloc => OPENSSL_malloc
15585 Malloc_locked => OPENSSL_malloc_locked
15586 Realloc => OPENSSL_realloc
15587 Free => OPENSSL_free
15591 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15592 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15596 * CygWin32 support.
15598 *John Jarvie <jjarvie@newsguy.com>*
15600 * The type-safe stack code has been rejigged. It is now only compiled
15601 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15602 by default all type-specific stack functions are "#define"d back to
15603 standard stack functions. This results in more streamlined output
15604 but retains the type-safety checking possibilities of the original
15609 * The STACK code has been cleaned up, and certain type declarations
15610 that didn't make a lot of sense have been brought in line. This has
15611 also involved a cleanup of sorts in safestack.h to more correctly
15612 map type-safe stack functions onto their plain stack counterparts.
15613 This work has also resulted in a variety of "const"ifications of
15614 lots of the code, especially `_cmp` operations which should normally
15615 be prototyped with "const" parameters anyway.
15619 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15620 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15621 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15622 where all of 'md' is used each time the PRNG is used, but 'state'
15623 is used only indexed by a cyclic counter. As entropy may not be
15624 well distributed from the beginning, 'md' is important as a
15625 chaining variable. However, the output function chains only half
15626 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15627 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15628 in all of 'state' being rewritten, with the new values depending
15629 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15633 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15634 the handshake is continued after ssl_verify_cert_chain();
15635 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15636 can lead to 'unexplainable' connection aborts later.
15638 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15640 * Major EVP API cipher revision.
15641 Add hooks for extra EVP features. This allows various cipher
15642 parameters to be set in the EVP interface. Support added for variable
15643 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15644 setting of RC2 and RC5 parameters.
15646 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15649 Remove lots of duplicated code from the EVP library. For example *every*
15650 cipher init() function handles the 'iv' in the same way according to the
15651 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15652 for CFB and OFB modes they zero ctx->num.
15654 New functionality allows removal of S/MIME code RC2 hack.
15656 Most of the routines have the same form and so can be declared in terms
15659 By shifting this to the top level EVP_CipherInit() it can be removed from
15660 all individual ciphers. If the cipher wants to handle IVs or keys
15661 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15664 Change lots of functions like EVP_EncryptUpdate() to now return a
15665 value: although software versions of the algorithms cannot fail
15666 any installed hardware versions can.
15670 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15671 this option is set, tolerate broken clients that send the negotiated
15672 protocol version number instead of the requested protocol version
15677 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15678 i.e. non-zero for export ciphersuites, zero otherwise.
15679 Previous versions had this flag inverted, inconsistent with
15680 rsa_tmp_cb (..._TMP_RSA_CB).
15682 *Bodo Moeller; problem reported by Amit Chopra*
15684 * Add missing DSA library text string. Work around for some IIS
15685 key files with invalid SEQUENCE encoding.
15689 * Add a document (doc/standards.txt) that list all kinds of standards
15690 and so on that are implemented in OpenSSL.
15694 * Enhance c_rehash script. Old version would mishandle certificates
15695 with the same subject name hash and wouldn't handle CRLs at all.
15696 Added -fingerprint option to crl utility, to support new c_rehash
15701 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15705 * Fix for SSL server purpose checking. Server checking was
15706 rejecting certificates which had extended key usage present
15707 but no ssl client purpose.
15709 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15711 * Make PKCS#12 code work with no password. The PKCS#12 spec
15712 is a little unclear about how a blank password is handled.
15713 Since the password in encoded as a BMPString with terminating
15714 double NULL a zero length password would end up as just the
15715 double NULL. However no password at all is different and is
15716 handled differently in the PKCS#12 key generation code. NS
15717 treats a blank password as zero length. MSIE treats it as no
15718 password on export: but it will try both on import. We now do
15719 the same: PKCS12_parse() tries zero length and no password if
15720 the password is set to "" or NULL (NULL is now a valid password:
15721 it wasn't before) as does the pkcs12 application.
15725 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15726 perror when PEM_read_bio_X509_REQ fails, the error message must
15727 be obtained from the error queue.
15731 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15732 it in ERR_remove_state if appropriate, and change ERR_get_state
15733 accordingly to avoid race conditions (this is necessary because
15734 thread_hash is no longer constant once set).
15738 * Bugfix for linux-elf makefile.one.
15742 * RSA_get_default_method() will now cause a default
15743 RSA_METHOD to be chosen if one doesn't exist already.
15744 Previously this was only set during a call to RSA_new()
15745 or RSA_new_method(NULL) meaning it was possible for
15746 RSA_get_default_method() to return NULL.
15750 * Added native name translation to the existing DSO code
15751 that will convert (if the flag to do so is set) filenames
15752 that are sufficiently small and have no path information
15753 into a canonical native form. Eg. "blah" converted to
15754 "libblah.so" or "blah.dll" etc.
15758 * New function ERR_error_string_n(e, buf, len) which is like
15759 ERR_error_string(e, buf), but writes at most 'len' bytes
15760 including the 0 terminator. For ERR_error_string_n, 'buf'
15763 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15765 * CONF library reworked to become more general. A new CONF
15766 configuration file reader "class" is implemented as well as a
15767 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15768 old `CONF_*` functions are still there, but are reimplemented to
15769 work in terms of the new functions. Also, a set of functions
15770 to handle the internal storage of the configuration data is
15771 provided to make it easier to write new configuration file
15772 reader "classes" (I can definitely see something reading a
15773 configuration file in XML format, for example), called `_CONF_*`,
15774 or "the configuration storage API"...
15776 The new configuration file reading functions are:
15778 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15779 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15781 NCONF_default, NCONF_WIN32
15783 NCONF_dump_fp, NCONF_dump_bio
15785 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15786 NCONF_new creates a new CONF object. This works in the same way
15787 as other interfaces in OpenSSL, like the BIO interface.
15788 `NCONF_dump_*` dump the internal storage of the configuration file,
15789 which is useful for debugging. All other functions take the same
15790 arguments as the old `CONF_*` functions with the exception of the
15791 first that must be a `CONF *` instead of a `LHASH *`.
15793 To make it easier to use the new classes with the old `CONF_*` functions,
15794 the function CONF_set_default_method is provided.
15798 * Add '-tls1' option to 'openssl ciphers', which was already
15799 mentioned in the documentation but had not been implemented.
15800 (This option is not yet really useful because even the additional
15801 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15805 * Initial DSO code added into libcrypto for letting OpenSSL (and
15806 OpenSSL-based applications) load shared libraries and bind to
15807 them in a portable way.
15809 *Geoff Thorpe, with contributions from Richard Levitte*
15811 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15813 * Make sure _lrotl and _lrotr are only used with MSVC.
15815 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15816 (the default implementation of RAND_status).
15818 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15819 to '-clrext' (= clear extensions), as intended and documented.
15820 *Bodo Moeller; inconsistency pointed out by Michael Attili
15821 <attili@amaxo.com>*
15823 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15824 was larger than the MD block size.
15826 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15828 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15829 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15830 using the passed key: if the passed key was a private key the result
15831 of X509_print(), for example, would be to print out all the private key
15836 * des_quad_cksum() byte order bug fix.
15837 *Ulf Möller, using the problem description in krb4-0.9.7, where
15838 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15840 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15843 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15845 * For easily testing in shell scripts whether some command
15846 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15847 returns with exit code 0 iff no command of the given name is available.
15848 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15849 the output goes to stdout and nothing is printed to stderr.
15850 Additional arguments are always ignored.
15852 Since for each cipher there is a command of the same name,
15853 the 'no-cipher' compilation switches can be tested this way.
15855 ('openssl no-XXX' is not able to detect pseudo-commands such
15856 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15860 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15864 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15865 is set; it will be thrown away anyway because each handshake creates
15867 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15868 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15869 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15870 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15874 * New s_client option -ign_eof: EOF at stdin is ignored, and
15875 'Q' and 'R' lose their special meanings (quit/renegotiate).
15876 This is part of what -quiet does; unlike -quiet, -ign_eof
15877 does not suppress any output.
15881 * Add compatibility options to the purpose and trust code. The
15882 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15883 accepts a certificate or CA, this was the previous behaviour,
15884 with all the associated security issues.
15886 X509_TRUST_COMPAT is the old trust behaviour: only and
15887 automatically trust self signed roots in certificate store. A
15888 new trust setting X509_TRUST_DEFAULT is used to specify that
15889 a purpose has no associated trust setting and it should instead
15890 use the value in the default purpose.
15894 * Fix the PKCS#8 DSA private key code so it decodes keys again
15895 and fix a memory leak.
15899 * In util/mkerr.pl (which implements 'make errors'), preserve
15900 reason strings from the previous version of the .c file, as
15901 the default to have only downcase letters (and digits) in
15902 automatically generated reasons codes is not always appropriate.
15906 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15907 using strerror. Previously, ERR_reason_error_string() returned
15908 library names as reason strings for SYSerr; but SYSerr is a special
15909 case where small numbers are errno values, not library numbers.
15913 * Add '-dsaparam' option to 'openssl dhparam' application. This
15914 converts DSA parameters into DH parameters. (When creating parameters,
15915 DSA_generate_parameters is used.)
15919 * Include 'length' (recommended exponent length) in C code generated
15920 by 'openssl dhparam -C'.
15924 * The second argument to set_label in perlasm was already being used
15925 so couldn't be used as a "file scope" flag. Moved to third argument
15930 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15931 instead of RAND_bytes for encryption IVs and salts.
15935 * Include RAND_status() into RAND_METHOD instead of implementing
15936 it only for md_rand.c Otherwise replacing the PRNG by calling
15937 RAND_set_rand_method would be impossible.
15941 * Don't let DSA_generate_key() enter an infinite loop if the random
15942 number generation fails.
15946 * New 'rand' application for creating pseudo-random output.
15950 * Added configuration support for Linux/IA64
15952 *Rolf Haberrecker <rolf@suse.de>*
15954 * Assembler module support for Mingw32.
15958 * Shared library support for HPUX (in shlib/).
15960 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15962 * Shared library support for Solaris gcc.
15964 *Lutz Behnke <behnke@trustcenter.de>*
15966 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15968 * PKCS7_encrypt() was adding text MIME headers twice because they
15969 were added manually and by SMIME_crlf_copy().
15973 * In bntest.c don't call BN_rand with zero bits argument.
15975 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15977 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15978 case was implemented. This caused BN_div_recp() to fail occasionally.
15982 * Add an optional second argument to the set_label() in the perl
15983 assembly language builder. If this argument exists and is set
15984 to 1 it signals that the assembler should use a symbol whose
15985 scope is the entire file, not just the current function. This
15986 is needed with MASM which uses the format label:: for this scope.
15988 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15990 * Change the ASN1 types so they are typedefs by default. Before
15991 almost all types were #define'd to ASN1_STRING which was causing
15992 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15997 * Change names of new functions to the new get1/get0 naming
15998 convention: After 'get1', the caller owns a reference count
15999 and has to call `..._free`; 'get0' returns a pointer to some
16000 data structure without incrementing reference counters.
16001 (Some of the existing 'get' functions increment a reference
16002 counter, some don't.)
16003 Similarly, 'set1' and 'add1' functions increase reference
16004 counters or duplicate objects.
16008 * Allow for the possibility of temp RSA key generation failure:
16009 the code used to assume it always worked and crashed on failure.
16013 * Fix potential buffer overrun problem in BIO_printf().
16014 *Ulf Möller, using public domain code by Patrick Powell; problem
16015 pointed out by David Sacerdote <das33@cornell.edu>*
16017 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16018 RAND_egd() and RAND_status(). In the command line application,
16019 the EGD socket can be specified like a seed file using RANDFILE
16024 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16025 Some CAs (e.g. Verisign) distribute certificates in this form.
16029 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16030 list to exclude them. This means that no special compilation option
16031 is needed to use anonymous DH: it just needs to be included in the
16036 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16037 EVP_MD_type. The old functionality is available in a new macro called
16038 EVP_MD_md(). Change code that uses it and update docs.
16042 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16043 where the `void *` argument is replaced by a function pointer argument.
16044 Previously `void *` was abused to point to functions, which works on
16045 many platforms, but is not correct. As these functions are usually
16046 called by macros defined in OpenSSL header files, most source code
16047 should work without changes.
16051 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16052 sections with information on -D... compiler switches used for
16053 compiling the library so that applications can see them. To enable
16054 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16055 must be defined. E.g.,
16056 #define OPENSSL_ALGORITHM_DEFINES
16057 #include <openssl/opensslconf.h>
16058 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16060 *Richard Levitte, Ulf and Bodo Möller*
16062 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16067 * Change the 'other' type in certificate aux info to a STACK_OF
16068 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16069 the required ASN1 format: arbitrary types determined by an OID.
16073 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16074 argument to 'req'. This is not because the function is newer or
16075 better than others it just uses the work 'NEW' in the certificate
16076 request header lines. Some software needs this.
16080 * Reorganise password command line arguments: now passwords can be
16081 obtained from various sources. Delete the PEM_cb function and make
16082 it the default behaviour: i.e. if the callback is NULL and the
16083 usrdata argument is not NULL interpret it as a null terminated pass
16084 phrase. If usrdata and the callback are NULL then the pass phrase
16085 is prompted for as usual.
16089 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16090 the support is automatically enabled. The resulting binaries will
16091 autodetect the card and use it if present.
16093 *Ben Laurie and Compaq Inc.*
16095 * Work around for Netscape hang bug. This sends certificate request
16096 and server done in one record. Since this is perfectly legal in the
16097 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16098 the bugs/SSLv3 entry for more info.
16102 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16106 * Add -rand argument to smime and pkcs12 applications and read/write
16111 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16115 * Add command line password options to the remaining applications.
16119 * Bug fix for BN_div_recp() for numerators with an even number of
16124 * More tests in bntest.c, and changed test_bn output.
16128 * ./config recognizes MacOS X now.
16132 * Bug fix for BN_div() when the first words of num and divisor are
16133 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16137 * Add support for various broken PKCS#8 formats, and command line
16138 options to produce them.
16142 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16143 get temporary BIGNUMs from a BN_CTX.
16147 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16152 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16153 include a #define from the old name to the new. The original intent
16154 was that statically linked binaries could for example just call
16155 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16156 link with digests. This never worked because SSLeay_add_all_digests()
16157 and SSLeay_add_all_ciphers() were in the same source file so calling
16158 one would link with the other. They are now in separate source files.
16162 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16166 * Use a less unusual form of the Miller-Rabin primality test (it used
16167 a binary algorithm for exponentiation integrated into the Miller-Rabin
16168 loop, our standard modexp algorithms are faster).
16172 * Support for the EBCDIC character set completed.
16174 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16176 * Source code cleanups: use const where appropriate, eliminate casts,
16177 use `void *` instead of `char *` in lhash.
16181 * Bugfix: ssl3_send_server_key_exchange was not restartable
16182 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16183 this the server could overwrite ephemeral keys that the client
16188 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16189 using 50 iterations of the Rabin-Miller test.
16191 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16192 iterations of the Rabin-Miller test as required by the appendix
16193 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16194 As BN_is_prime_fasttest includes trial division, DSA parameter
16195 generation becomes much faster.
16197 This implies a change for the callback functions in DSA_is_prime
16198 and DSA_generate_parameters: The callback function is called once
16199 for each positive witness in the Rabin-Miller test, not just
16200 occasionally in the inner loop; and the parameters to the
16201 callback function now provide an iteration count for the outer
16202 loop rather than for the current invocation of the inner loop.
16203 DSA_generate_parameters additionally can call the callback
16204 function with an 'iteration count' of -1, meaning that a
16205 candidate has passed the trial division test (when q is generated
16206 from an application-provided seed, trial division is skipped).
16210 * New function BN_is_prime_fasttest that optionally does trial
16211 division before starting the Rabin-Miller test and has
16212 an additional BN_CTX * argument (whereas BN_is_prime always
16213 has to allocate at least one BN_CTX).
16214 'callback(1, -1, cb_arg)' is called when a number has passed the
16215 trial division stage.
16219 * Fix for bug in CRL encoding. The validity dates weren't being handled
16224 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16228 * New function BN_pseudo_rand().
16232 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16233 bignum version of BN_from_montgomery() with the working code from
16234 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16239 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16240 made it impossible to use the same SSL_SESSION data structure in
16241 SSL2 clients in multiple threads.
16245 * The return value of RAND_load_file() no longer counts bytes obtained
16246 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16247 to seed the PRNG (previously an explicit byte count was required).
16249 *Ulf Möller, Bodo Möller*
16251 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16252 used `char *` instead of `void *` and had casts all over the place.
16256 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16260 * Retain source code compatibility for BN_prime_checks macro:
16261 BN_is_prime(..., BN_prime_checks, ...) now uses
16262 BN_prime_checks_for_size to determine the appropriate number of
16263 Rabin-Miller iterations.
16267 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16268 DH_CHECK_P_NOT_SAFE_PRIME.
16269 (Check if this is true? OpenPGP calls them "strong".)
16273 * Merge the functionality of "dh" and "gendh" programs into a new program
16274 "dhparam". The old programs are retained for now but will handle DH keys
16275 (instead of parameters) in future.
16279 * Make the ciphers, s_server and s_client programs check the return values
16280 when a new cipher list is set.
16284 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16285 ciphers. Before when the 56bit ciphers were enabled the sorting was
16288 The syntax for the cipher sorting has been extended to support sorting by
16289 cipher-strength (using the strength_bits hard coded in the tables).
16290 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16292 Fix a bug in the cipher-command parser: when supplying a cipher command
16293 string with an "undefined" symbol (neither command nor alphanumeric
16294 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16295 an error is flagged.
16297 Due to the strength-sorting extension, the code of the
16298 ssl_create_cipher_list() function was completely rearranged. I hope that
16299 the readability was also increased :-)
16301 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16303 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16304 for the first serial number and places 2 in the serial number file. This
16305 avoids problems when the root CA is created with serial number zero and
16306 the first user certificate has the same issuer name and serial number
16311 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16312 the new code. Add documentation for this stuff.
16316 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16317 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16318 structures and behave in an analogous way to the X509v3 functions:
16319 they shouldn't be called directly but wrapper functions should be used
16322 So we also now have some wrapper functions that call the X509at functions
16323 when passed certificate requests. (TO DO: similar things can be done with
16324 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16325 things. Some of these need some d2i or i2d and print functionality
16326 because they handle more complex structures.)
16330 * Add missing #ifndefs that caused missing symbols when building libssl
16331 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16332 NO_RSA in `ssl/s2*.c`.
16334 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16336 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16337 has a return value which indicates the quality of the random data
16338 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16339 error queue. New function RAND_pseudo_bytes() generates output that is
16340 guaranteed to be unique but not unpredictable. RAND_add is like
16341 RAND_seed, but takes an extra argument for an entropy estimate
16342 (RAND_seed always assumes full entropy).
16346 * Do more iterations of Rabin-Miller probable prime test (specifically,
16347 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16348 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16349 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16350 false-positive rate of at most 2^-80 for random input.
16354 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16358 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16359 in the 0.9.5 release), this returns the chain
16360 from an X509_CTX structure with a dup of the stack and all
16361 the X509 reference counts upped: so the stack will exist
16362 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16365 Also make SSL_SESSION_print() print out the verify return
16370 * Add manpage for the pkcs12 command. Also change the default
16371 behaviour so MAC iteration counts are used unless the new
16372 -nomaciter option is used. This improves file security and
16373 only older versions of MSIE (4.0 for example) need it.
16377 * Honor the no-xxx Configure options when creating .DEF files.
16381 * Add PKCS#10 attributes to field table: challengePassword,
16382 unstructuredName and unstructuredAddress. These are taken from
16383 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16384 international characters are used.
16386 More changes to X509_ATTRIBUTE code: allow the setting of types
16387 based on strings. Remove the 'loc' parameter when adding
16388 attributes because these will be a SET OF encoding which is sorted
16393 * Initial changes to the 'req' utility to allow request generation
16394 automation. This will allow an application to just generate a template
16395 file containing all the field values and have req construct the
16398 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16399 used all over the place including certificate requests and PKCS#7
16400 structures. They are currently handled manually where necessary with
16401 some primitive wrappers for PKCS#7. The new functions behave in a
16402 manner analogous to the X509 extension functions: they allow
16403 attributes to be looked up by NID and added.
16405 Later something similar to the X509V3 code would be desirable to
16406 automatically handle the encoding, decoding and printing of the
16407 more complex types. The string types like challengePassword can
16408 be handled by the string table functions.
16410 Also modified the multi byte string table handling. Now there is
16411 a 'global mask' which masks out certain types. The table itself
16412 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16413 is useful when for example there is only one permissible type
16414 (as in countryName) and using the mask might result in no valid
16419 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16420 SSL_get_peer_finished to allow applications to obtain the latest
16421 Finished messages sent to the peer or expected from the peer,
16422 respectively. (SSL_get_peer_finished is usually the Finished message
16423 actually received from the peer, otherwise the protocol will be aborted.)
16425 As the Finished message are message digests of the complete handshake
16426 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16427 be used for external authentication procedures when the authentication
16428 provided by SSL/TLS is not desired or is not enough.
16432 * Enhanced support for Alpha Linux is added. Now ./config checks if
16433 the host supports BWX extension and if Compaq C is present on the
16434 $PATH. Just exploiting of the BWX extension results in 20-30%
16435 performance kick for some algorithms, e.g. DES and RC4 to mention
16436 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16441 * Add support for MS "fast SGC". This is arguably a violation of the
16442 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16443 weak crypto and after checking the certificate is SGC a second one
16444 with strong crypto. MS SGC stops the first handshake after receiving
16445 the server certificate message and sends a second client hello. Since
16446 a server will typically do all the time consuming operations before
16447 expecting any further messages from the client (server key exchange
16448 is the most expensive) there is little difference between the two.
16450 To get OpenSSL to support MS SGC we have to permit a second client
16451 hello message after we have sent server done. In addition we have to
16452 reset the MAC if we do get this second client hello.
16456 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16457 if a DER encoded private key is RSA or DSA traditional format. Changed
16458 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16459 format DER encoded private key. Newer code should use PKCS#8 format which
16460 has the key type encoded in the ASN1 structure. Added DER private key
16461 support to pkcs8 application.
16465 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16466 ciphersuites has been selected (as required by the SSL 3/TLS 1
16467 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16468 is set, we interpret this as a request to violate the specification
16469 (the worst that can happen is a handshake failure, and 'correct'
16470 behaviour would result in a handshake failure anyway).
16474 * In SSL_CTX_add_session, take into account that there might be multiple
16475 SSL_SESSION structures with the same session ID (e.g. when two threads
16476 concurrently obtain them from an external cache).
16477 The internal cache can handle only one SSL_SESSION with a given ID,
16478 so if there's a conflict, we now throw out the old one to achieve
16483 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16484 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16485 some routines that use cipher OIDs: some ciphers do not have OIDs
16486 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16491 * Simplify the trust setting structure and code. Now we just have
16492 two sequences of OIDs for trusted and rejected settings. These will
16493 typically have values the same as the extended key usage extension
16494 and any application specific purposes.
16496 The trust checking code now has a default behaviour: it will just
16497 check for an object with the same NID as the passed id. Functions can
16498 be provided to override either the default behaviour or the behaviour
16499 for a given id. SSL client, server and email already have functions
16500 in place for compatibility: they check the NID and also return "trusted"
16501 if the certificate is self signed.
16505 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16506 traditional format into an EVP_PKEY structure.
16510 * Add a password callback function PEM_cb() which either prompts for
16511 a password if usr_data is NULL or otherwise assumes it is a null
16512 terminated password. Allow passwords to be passed on command line
16513 environment or config files in a few more utilities.
16517 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16518 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16519 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16520 Update documentation.
16524 * Support for ASN1 "NULL" type. This could be handled before by using
16525 ASN1_TYPE but there wasn't any function that would try to read a NULL
16526 and produce an error if it couldn't. For compatibility we also have
16527 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16528 don't allocate anything because they don't need to.
16532 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16535 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16537 * Rebuild of the memory allocation routines used by OpenSSL code and
16538 possibly others as well. The purpose is to make an interface that
16539 provide hooks so anyone can build a separate set of allocation and
16540 deallocation routines to be used by OpenSSL, for example memory
16541 pool implementations, or something else, which was previously hard
16542 since Malloc(), Realloc() and Free() were defined as macros having
16543 the values malloc, realloc and free, respectively (except for Win32
16544 compilations). The same is provided for memory debugging code.
16545 OpenSSL already comes with functionality to find memory leaks, but
16546 this gives people a chance to debug other memory problems.
16548 With these changes, a new set of functions and macros have appeared:
16550 CRYPTO_set_mem_debug_functions() [F]
16551 CRYPTO_get_mem_debug_functions() [F]
16552 CRYPTO_dbg_set_options() [F]
16553 CRYPTO_dbg_get_options() [F]
16554 CRYPTO_malloc_debug_init() [M]
16556 The memory debug functions are NULL by default, unless the library
16557 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16558 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16559 gives the standard debugging functions that come with OpenSSL) or
16560 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16561 provided by the library user) must be used. When the standard
16562 debugging functions are used, CRYPTO_dbg_set_options can be used to
16563 request additional information:
16564 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16565 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16567 Also, things like CRYPTO_set_mem_functions will always give the
16568 expected result (the new set of functions is used for allocation
16569 and deallocation) at all times, regardless of platform and compiler
16572 To finish it up, some functions that were never use in any other
16573 way than through macros have a new API and new semantic:
16575 CRYPTO_dbg_malloc()
16576 CRYPTO_dbg_realloc()
16579 All macros of value have retained their old syntax.
16581 *Richard Levitte and Bodo Moeller*
16583 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16584 ordering of SMIMECapabilities wasn't in "strength order" and there
16585 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16590 * Some ASN1 types with illegal zero length encoding (INTEGER,
16591 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16593 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16595 * Merge in my S/MIME library for OpenSSL. This provides a simple
16596 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16597 functionality to handle multipart/signed properly) and a utility
16598 called 'smime' to call all this stuff. This is based on code I
16599 originally wrote for Celo who have kindly allowed it to be
16600 included in OpenSSL.
16604 * Add variants des_set_key_checked and des_set_key_unchecked of
16605 des_set_key (aka des_key_sched). Global variable des_check_key
16606 decides which of these is called by des_set_key; this way
16607 des_check_key behaves as it always did, but applications and
16608 the library itself, which was buggy for des_check_key == 1,
16609 have a cleaner way to pick the version they need.
16613 * New function PKCS12_newpass() which changes the password of a
16618 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16619 dynamic mix. In both cases the ids can be used as an index into the
16620 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16621 functions so they accept a list of the field values and the
16622 application doesn't need to directly manipulate the X509_TRUST
16627 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16632 * Modify the way the V3 extension code looks up extensions. This now
16633 works in a similar way to the object code: we have some "standard"
16634 extensions in a static table which is searched with OBJ_bsearch()
16635 and the application can add dynamic ones if needed. The file
16636 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16637 updated whenever a new extension is added to the core code and kept
16638 in ext_nid order. There is a simple program 'tabtest.c' which checks
16639 this. New extensions are not added too often so this file can readily
16640 be maintained manually.
16642 There are two big advantages in doing things this way. The extensions
16643 can be looked up immediately and no longer need to be "added" using
16644 X509V3_add_standard_extensions(): this function now does nothing.
16645 Side note: I get *lots* of email saying the extension code doesn't
16646 work because people forget to call this function.
16647 Also no dynamic allocation is done unless new extensions are added:
16648 so if we don't add custom extensions there is no need to call
16649 X509V3_EXT_cleanup().
16653 * Modify enc utility's salting as follows: make salting the default. Add a
16654 magic header, so unsalted files fail gracefully instead of just decrypting
16655 to garbage. This is because not salting is a big security hole, so people
16656 should be discouraged from doing it.
16660 * Fixes and enhancements to the 'x509' utility. It allowed a message
16661 digest to be passed on the command line but it only used this
16662 parameter when signing a certificate. Modified so all relevant
16663 operations are affected by the digest parameter including the
16664 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16665 DSA key was used because it didn't fix the digest.
16669 * Initial certificate chain verify code. Currently tests the untrusted
16670 certificates for consistency with the verify purpose (which is set
16671 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16673 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16674 this is because it will reject chains with invalid extensions whereas
16675 every previous version of OpenSSL and SSLeay made no checks at all.
16677 Trust code: checks the root CA for the relevant trust settings. Trust
16678 settings have an initial value consistent with the verify purpose: e.g.
16679 if the verify purpose is for SSL client use it expects the CA to be
16680 trusted for SSL client use. However the default value can be changed to
16681 permit custom trust settings: one example of this would be to only trust
16682 certificates from a specific "secure" set of CAs.
16684 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16685 which should be used for version portability: especially since the
16686 verify structure is likely to change more often now.
16688 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16689 to set them. If not set then assume SSL clients will verify SSL servers
16692 Two new options to the verify program: -untrusted allows a set of
16693 untrusted certificates to be passed in and -purpose which sets the
16694 intended purpose of the certificate. If a purpose is set then the
16695 new chain verify code is used to check extension consistency.
16699 * Support for the authority information access extension.
16703 * Modify RSA and DSA PEM read routines to transparently handle
16704 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16705 public keys in a format compatible with certificate
16706 SubjectPublicKeyInfo structures. Unfortunately there were already
16707 functions called *_PublicKey_* which used various odd formats so
16708 these are retained for compatibility: however the DSA variants were
16709 never in a public release so they have been deleted. Changed dsa/rsa
16710 utilities to handle the new format: note no releases ever handled public
16711 keys so we should be OK.
16713 The primary motivation for this change is to avoid the same fiasco
16714 that dogs private keys: there are several incompatible private key
16715 formats some of which are standard and some OpenSSL specific and
16716 require various evil hacks to allow partial transparent handling and
16717 even then it doesn't work with DER formats. Given the option anything
16718 other than PKCS#8 should be dumped: but the other formats have to
16719 stay in the name of compatibility.
16721 With public keys and the benefit of hindsight one standard format
16722 is used which works with EVP_PKEY, RSA or DSA structures: though
16723 it clearly returns an error if you try to read the wrong kind of key.
16725 Added a -pubkey option to the 'x509' utility to output the public key.
16726 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16727 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16728 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16729 that do the same as the `EVP_PKEY_assign_*()` except they up the
16730 reference count of the added key (they don't "swallow" the
16735 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16736 CRLs would fail if the file contained no certificates or no CRLs:
16737 added a new function to read in both types and return the number
16738 read: this means that if none are read it will be an error. The
16739 DER versions of the certificate and CRL reader would always fail
16740 because it isn't possible to mix certificates and CRLs in DER format
16741 without choking one or the other routine. Changed this to just read
16742 a certificate: this is the best we can do. Also modified the code
16743 in `apps/verify.c` to take notice of return codes: it was previously
16744 attempting to read in certificates from NULL pointers and ignoring
16745 any errors: this is one reason why the cert and CRL reader seemed
16746 to work. It doesn't check return codes from the default certificate
16747 routines: these may well fail if the certificates aren't installed.
16751 * Code to support otherName option in GeneralName.
16755 * First update to verify code. Change the verify utility
16756 so it warns if it is passed a self signed certificate:
16757 for consistency with the normal behaviour. X509_verify
16758 has been modified to it will now verify a self signed
16759 certificate if *exactly* the same certificate appears
16760 in the store: it was previously impossible to trust a
16761 single self signed certificate. This means that:
16762 openssl verify ss.pem
16763 now gives a warning about a self signed certificate but
16764 openssl verify -CAfile ss.pem ss.pem
16769 * For servers, store verify_result in SSL_SESSION data structure
16770 (and add it to external session representation).
16771 This is needed when client certificate verifications fails,
16772 but an application-provided verification callback (set by
16773 SSL_CTX_set_cert_verify_callback) allows accepting the session
16774 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16775 but returns 1): When the session is reused, we have to set
16776 ssl->verify_result to the appropriate error code to avoid
16779 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16781 * Fix a bug in the new PKCS#7 code: it didn't consider the
16782 case in PKCS7_dataInit() where the signed PKCS7 structure
16783 didn't contain any existing data because it was being created.
16785 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16787 * Add a salt to the key derivation routines in enc.c. This
16788 forms the first 8 bytes of the encrypted file. Also add a
16789 -S option to allow a salt to be input on the command line.
16793 * New function X509_cmp(). Oddly enough there wasn't a function
16794 to compare two certificates. We do this by working out the SHA1
16795 hash and comparing that. X509_cmp() will be needed by the trust
16800 * SSL_get1_session() is like SSL_get_session(), but increments
16801 the reference count in the SSL_SESSION returned.
16803 *Geoff Thorpe <geoff@eu.c2.net>*
16805 * Fix for 'req': it was adding a null to request attributes.
16806 Also change the X509_LOOKUP and X509_INFO code to handle
16807 certificate auxiliary information.
16811 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16816 * Add the possibility to add extra information to the memory leak
16817 detecting output, to form tracebacks, showing from where each
16818 allocation was originated: CRYPTO_push_info("constant string") adds
16819 the string plus current file name and line number to a per-thread
16820 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16821 is like calling CYRPTO_pop_info() until the stack is empty.
16822 Also updated memory leak detection code to be multi-thread-safe.
16826 * Add options -text and -noout to pkcs7 utility and delete the
16827 encryption options which never did anything. Update docs.
16831 * Add options to some of the utilities to allow the pass phrase
16832 to be included on either the command line (not recommended on
16833 OSes like Unix) or read from the environment. Update the
16834 manpages and fix a few bugs.
16838 * Add a few manpages for some of the openssl commands.
16842 * Fix the -revoke option in ca. It was freeing up memory twice,
16843 leaking and not finding already revoked certificates.
16847 * Extensive changes to support certificate auxiliary information.
16848 This involves the use of X509_CERT_AUX structure and X509_AUX
16849 functions. An X509_AUX function such as PEM_read_X509_AUX()
16850 can still read in a certificate file in the usual way but it
16851 will also read in any additional "auxiliary information". By
16852 doing things this way a fair degree of compatibility can be
16853 retained: existing certificates can have this information added
16854 using the new 'x509' options.
16856 Current auxiliary information includes an "alias" and some trust
16857 settings. The trust settings will ultimately be used in enhanced
16858 certificate chain verification routines: currently a certificate
16859 can only be trusted if it is self signed and then it is trusted
16864 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16865 The problem was that one of the replacement routines had not been working
16866 since SSLeay releases. For now the offending routine has been replaced
16867 with non-optimised assembler. Even so, this now gives around 95%
16868 performance improvement for 1024 bit RSA signs.
16872 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16873 handling. Most clients have the effective key size in bits equal to
16874 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16875 A few however don't do this and instead use the size of the decrypted key
16876 to determine the RC2 key length and the AlgorithmIdentifier to determine
16877 the effective key length. In this case the effective key length can still
16878 be 40 bits but the key length can be 168 bits for example. This is fixed
16879 by manually forcing an RC2 key into the EVP_PKEY structure because the
16880 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16881 the key length and effective key length are equal.
16885 * Add a bunch of functions that should simplify the creation of
16886 X509_NAME structures. Now you should be able to do:
16887 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16888 and have it automatically work out the correct field type and fill in
16889 the structures. The more adventurous can try:
16890 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16891 and it will (hopefully) work out the correct multibyte encoding.
16895 * Change the 'req' utility to use the new field handling and multibyte
16896 copy routines. Before the DN field creation was handled in an ad hoc
16897 way in req, ca, and x509 which was rather broken and didn't support
16898 BMPStrings or UTF8Strings. Since some software doesn't implement
16899 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16900 using the dirstring_type option. See the new comment in the default
16901 openssl.cnf for more info.
16905 * Make crypto/rand/md_rand.c more robust:
16906 - Assure unique random numbers after fork().
16907 - Make sure that concurrent threads access the global counter and
16908 md serializably so that we never lose entropy in them
16909 or use exactly the same state in multiple threads.
16910 Access to the large state is not always serializable because
16911 the additional locking could be a performance killer, and
16912 md should be large enough anyway.
16916 * New file `apps/app_rand.c` with commonly needed functionality
16917 for handling the random seed file.
16919 Use the random seed file in some applications that previously did not:
16921 dsaparam -genkey (which also ignored its '-rand' option),
16924 x509 (when signing).
16925 Except on systems with /dev/urandom, it is crucial to have a random
16926 seed file at least for key creation, DSA signing, and for DH exchanges;
16927 for RSA signatures we could do without one.
16929 gendh and gendsa (unlike genrsa) used to read only the first byte
16930 of each file listed in the '-rand' option. The function as previously
16931 found in genrsa is now in app_rand.c and is used by all programs
16932 that support '-rand'.
16936 * In RAND_write_file, use mode 0600 for creating files;
16937 don't just chmod when it may be too late.
16941 * Report an error from X509_STORE_load_locations
16942 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16946 * New function ASN1_mbstring_copy() this copies a string in either
16947 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16948 into an ASN1_STRING type. A mask of permissible types is passed
16949 and it chooses the "minimal" type to use or an error if not type
16954 * Add function equivalents to the various macros in asn1.h. The old
16955 macros are retained with an `M_` prefix. Code inside the library can
16956 use the `M_` macros. External code (including the openssl utility)
16957 should *NOT* in order to be "shared library friendly".
16961 * Add various functions that can check a certificate's extensions
16962 to see if it usable for various purposes such as SSL client,
16963 server or S/MIME and CAs of these types. This is currently
16964 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16965 verification. Also added a -purpose flag to x509 utility to
16966 print out all the purposes.
16970 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16975 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16976 for, obtain and decode and extension and obtain its critical flag.
16977 This allows all the necessary extension code to be handled in a
16978 single function call.
16982 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16983 platforms. See crypto/rc4/rc4_enc.c for further details.
16987 * New -noout option to asn1parse. This causes no output to be produced
16988 its main use is when combined with -strparse and -out to extract data
16989 from a file (which may not be in ASN.1 format).
16993 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16994 when producing the local key id.
16996 *Richard Levitte <levitte@stacken.kth.se>*
16998 * New option -dhparam in s_server. This allows a DH parameter file to be
16999 stated explicitly. If it is not stated then it tries the first server
17000 certificate file. The previous behaviour hard coded the filename
17005 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17006 a public key to be input or output. For example:
17007 openssl rsa -in key.pem -pubout -out pubkey.pem
17008 Also added necessary DSA public key functions to handle this.
17012 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17013 in the message. This was handled by allowing
17014 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17016 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17018 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17019 to the end of the strings whereas this didn't. This would cause problems
17020 if strings read with d2i_ASN1_bytes() were later modified.
17022 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17024 * Fix for base64 decode bug. When a base64 bio reads only one line of
17025 data and it contains EOF it will end up returning an error. This is
17026 caused by input 46 bytes long. The cause is due to the way base64
17027 BIOs find the start of base64 encoded data. They do this by trying a
17028 trial decode on each line until they find one that works. When they
17029 do a flag is set and it starts again knowing it can pass all the
17030 data directly through the decoder. Unfortunately it doesn't reset
17031 the context it uses. This means that if EOF is reached an attempt
17032 is made to pass two EOFs through the context and this causes the
17033 resulting error. This can also cause other problems as well. As is
17034 usual with these problems it takes *ages* to find and the fix is
17035 trivial: move one line.
17037 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17039 * Ugly workaround to get s_client and s_server working under Windows. The
17040 old code wouldn't work because it needed to select() on sockets and the
17041 tty (for keypresses and to see if data could be written). Win32 only
17042 supports select() on sockets so we select() with a 1s timeout on the
17043 sockets and then see if any characters are waiting to be read, if none
17044 are present then we retry, we also assume we can always write data to
17045 the tty. This isn't nice because the code then blocks until we've
17046 received a complete line of data and it is effectively polling the
17047 keyboard at 1s intervals: however it's quite a bit better than not
17048 working at all :-) A dedicated Windows application might handle this
17049 with an event loop for example.
17053 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17054 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17055 will be called when RSA_sign() and RSA_verify() are used. This is useful
17056 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17057 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17058 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17059 This necessitated the support of an extra signature type NID_md5_sha1
17060 for SSL signatures and modifications to the SSL library to use it instead
17061 of calling RSA_public_decrypt() and RSA_private_encrypt().
17065 * Add new -verify -CAfile and -CApath options to the crl program, these
17066 will lookup a CRL issuers certificate and verify the signature in a
17067 similar way to the verify program. Tidy up the crl program so it
17068 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17069 less strict. It will now permit CRL extensions even if it is not
17070 a V2 CRL: this will allow it to tolerate some broken CRLs.
17074 * Initialize all non-automatic variables each time one of the openssl
17075 sub-programs is started (this is necessary as they may be started
17076 multiple times from the "OpenSSL>" prompt).
17078 *Lennart Bang, Bodo Moeller*
17080 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17081 removing all other RSA functionality (this is what NO_RSA does). This
17082 is so (for example) those in the US can disable those operations covered
17083 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17088 * Non-copying interface to BIO pairs.
17089 (still largely untested)
17093 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17094 ASCII string. This was handled independently in various places before.
17098 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17099 UTF8 strings a character at a time.
17103 * Use client_version from client hello to select the protocol
17104 (s23_srvr.c) and for RSA client key exchange verification
17105 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17109 * Add various utility functions to handle SPKACs, these were previously
17110 handled by poking round in the structure internals. Added new function
17111 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17112 print, verify and generate SPKACs. Based on an original idea from
17113 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17117 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17121 * Allow the config file extension section to be overwritten on the
17122 command line. Based on an original idea from Massimiliano Pala
17123 <madwolf@comune.modena.it>. The new option is called -extensions
17124 and can be applied to ca, req and x509. Also -reqexts to override
17125 the request extensions in req and -crlexts to override the crl extensions
17130 * Add new feature to the SPKAC handling in ca. Now you can include
17131 the same field multiple times by preceding it by "XXXX." for example:
17134 this is the same syntax as used in the req config file.
17138 * Allow certificate extensions to be added to certificate requests. These
17139 are specified in a 'req_extensions' option of the req section of the
17140 config file. They can be printed out with the -text option to req but
17141 are otherwise ignored at present.
17145 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17146 data read consists of only the final block it would not decrypted because
17147 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17148 A misplaced 'break' also meant the decrypted final block might not be
17149 copied until the next read.
17153 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17154 a few extra parameters to the DH structure: these will be useful if
17155 for example we want the value of 'q' or implement X9.42 DH.
17159 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17160 provides hooks that allow the default DSA functions or functions on a
17161 "per key" basis to be replaced. This allows hardware acceleration and
17162 hardware key storage to be handled without major modification to the
17163 library. Also added low level modexp hooks and CRYPTO_EX structure and
17164 associated functions.
17168 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17169 as "read only": it can't be written to and the buffer it points to will
17170 not be freed. Reading from a read only BIO is much more efficient than
17171 a normal memory BIO. This was added because there are several times when
17172 an area of memory needs to be read from a BIO. The previous method was
17173 to create a memory BIO and write the data to it, this results in two
17174 copies of the data and an O(n^2) reading algorithm. There is a new
17175 function BIO_new_mem_buf() which creates a read only memory BIO from
17176 an area of memory. Also modified the PKCS#7 routines to use read only
17181 * Bugfix: ssl23_get_client_hello did not work properly when called in
17182 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17183 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17184 but a retry condition occurred while trying to read the rest.
17188 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17189 NID_pkcs7_encrypted by default: this was wrong since this should almost
17190 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17191 the encrypted data type: this is a more sensible place to put it and it
17192 allows the PKCS#12 code to be tidied up that duplicated this
17197 * Changed obj_dat.pl script so it takes its input and output files on
17198 the command line. This should avoid shell escape redirection problems
17203 * Initial support for certificate extension requests, these are included
17204 in things like Xenroll certificate requests. Included functions to allow
17205 extensions to be obtained and added.
17209 * -crlf option to s_client and s_server for sending newlines as
17210 CRLF (as required by many protocols).
17214 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17216 * Install libRSAglue.a when OpenSSL is built with RSAref.
17218 *Ralf S. Engelschall*
17220 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17222 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17224 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17229 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17230 DH parameters/keys (q is lost during that conversion, but the resulting
17231 DH parameters contain its length).
17233 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17234 much faster than DH_generate_parameters (which creates parameters
17235 where `p = 2*q + 1`), and also the smaller q makes DH computations
17236 much more efficient (160-bit exponentiation instead of 1024-bit
17237 exponentiation); so this provides a convenient way to support DHE
17238 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17239 utter importance to use
17240 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17242 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17243 when such DH parameters are used, because otherwise small subgroup
17244 attacks may become possible!
17248 * Avoid memory leak in i2d_DHparams.
17252 * Allow the -k option to be used more than once in the enc program:
17253 this allows the same encrypted message to be read by multiple recipients.
17257 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17258 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17259 it will always use the numerical form of the OID, even if it has a short
17264 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17265 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17266 otherwise bn_mod_exp was called. In the case of hardware keys for example
17267 no private key components need be present and it might store extra data
17268 in the RSA structure, which cannot be accessed from bn_mod_exp.
17269 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17270 private key operations.
17274 * Added support for SPARC Linux.
17278 * pem_password_cb function type incompatibly changed from
17279 typedef int pem_password_cb(char *buf, int size, int rwflag);
17281 ....(char *buf, int size, int rwflag, void *userdata);
17282 so that applications can pass data to their callbacks:
17283 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17284 additional void * argument, which is just handed through whenever
17285 the password callback is called.
17287 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17289 New function SSL_CTX_set_default_passwd_cb_userdata.
17291 Compatibility note: As many C implementations push function arguments
17292 onto the stack in reverse order, the new library version is likely to
17293 interoperate with programs that have been compiled with the old
17294 pem_password_cb definition (PEM_whatever takes some data that
17295 happens to be on the stack as its last argument, and the callback
17296 just ignores this garbage); but there is no guarantee whatsoever that
17299 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17300 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17301 problems not only on Windows, but also on some Unix platforms.
17302 To avoid problematic command lines, these definitions are now in an
17303 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17304 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17308 * MIPS III/IV assembler module is reimplemented.
17312 * More DES library cleanups: remove references to srand/rand and
17313 delete an unused file.
17317 * Add support for the free Netwide assembler (NASM) under Win32,
17318 since not many people have MASM (ml) and it can be hard to obtain.
17319 This is currently experimental but it seems to work OK and pass all
17320 the tests. Check out INSTALL.W32 for info.
17324 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17325 without temporary keys kept an extra copy of the server key,
17326 and connections with temporary keys did not free everything in case
17331 * New function RSA_check_key and new openssl rsa option -check
17332 for verifying the consistency of RSA keys.
17334 *Ulf Moeller, Bodo Moeller*
17336 * Various changes to make Win32 compile work:
17337 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17338 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17339 comparison" warnings.
17340 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17344 * Add a debugging option to PKCS#5 v2 key generation function: when
17345 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17346 derived keys are printed to stderr.
17350 * Copy the flags in ASN1_STRING_dup().
17352 *Roman E. Pavlov <pre@mo.msk.ru>*
17354 * The x509 application mishandled signing requests containing DSA
17355 keys when the signing key was also DSA and the parameters didn't match.
17357 It was supposed to omit the parameters when they matched the signing key:
17358 the verifying software was then supposed to automatically use the CA's
17359 parameters if they were absent from the end user certificate.
17361 Omitting parameters is no longer recommended. The test was also
17362 the wrong way round! This was probably due to unusual behaviour in
17363 EVP_cmp_parameters() which returns 1 if the parameters match.
17364 This meant that parameters were omitted when they *didn't* match and
17365 the certificate was useless. Certificates signed with 'ca' didn't have
17368 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17370 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17371 The interface is as follows:
17372 Applications can use
17373 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17374 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17375 "off" is now the default.
17376 The library internally uses
17377 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17378 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17379 to disable memory-checking temporarily.
17381 Some inconsistent states that previously were possible (and were
17382 even the default) are now avoided.
17384 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17385 with each memory chunk allocated; this is occasionally more helpful
17386 than just having a counter.
17388 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17390 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17395 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17396 which largely parallels "options", but is for changing API behaviour,
17397 whereas "options" are about protocol behaviour.
17398 Initial "mode" flags are:
17400 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17401 a single record has been written.
17402 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17403 retries use the same buffer location.
17404 (But all of the contents must be
17409 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17412 * Fix problems with no-hmac etc.
17414 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17416 * New functions RSA_get_default_method(), RSA_set_method() and
17417 RSA_get_method(). These allows replacement of RSA_METHODs without having
17418 to mess around with the internals of an RSA structure.
17422 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17423 Also really enable memory leak checks in openssl.c and in some
17426 *Chad C. Mulligan, Bodo Moeller*
17428 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17429 up the length of negative integers. This has now been simplified to just
17430 store the length when it is first determined and use it later, rather
17431 than trying to keep track of where data is copied and updating it to
17433 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17435 * Add a new function PKCS7_signatureVerify. This allows the verification
17436 of a PKCS#7 signature but with the signing certificate passed to the
17437 function itself. This contrasts with PKCS7_dataVerify which assumes the
17438 certificate is present in the PKCS#7 structure. This isn't always the
17439 case: certificates can be omitted from a PKCS#7 structure and be
17440 distributed by "out of band" means (such as a certificate database).
17444 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17445 function prototypes in pem.h, also change util/mkdef.pl to add the
17446 necessary function names.
17450 * mk1mf.pl (used by Windows builds) did not properly read the
17451 options set by Configure in the top level Makefile, and Configure
17452 was not even able to write more than one option correctly.
17453 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17457 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17458 file to be loaded from a BIO or FILE pointer. The BIO version will
17459 for example allow memory BIOs to contain config info.
17463 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17464 Whoever hopes to achieve shared-library compatibility across versions
17465 must use this, not the compile-time macro.
17466 (Exercise 0.9.4: Which is the minimum library version required by
17468 Note: All this applies only to multi-threaded programs, others don't
17473 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17474 through a BIO pair triggered the default case, i.e.
17475 SSLerr(...,SSL_R_UNKNOWN_STATE).
17479 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17480 can use the SSL library even if none of the specific BIOs is
17485 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17486 for the encoded length.
17488 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17490 * Add initial documentation of the X509V3 functions.
17494 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17495 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17496 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17497 secure PKCS#8 private key format with a high iteration count.
17501 * Fix determination of Perl interpreter: A perl or perl5
17502 *directory* in $PATH was also accepted as the interpreter.
17504 *Ralf S. Engelschall*
17506 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17507 wrong with it but it was very old and did things like calling
17508 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17509 unusual formatting.
17513 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17514 to use the new extension code.
17518 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17519 with macros. This should make it easier to change their form, add extra
17520 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17525 * Add to configuration table a new entry that can specify an alternative
17526 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17527 according to Mark Crispin <MRC@Panda.COM>.
17531 * DES CBC did not update the IV. Weird.
17535 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17536 Changing the behaviour of the former might break existing programs --
17537 where IV updating is needed, des_ncbc_encrypt can be used.
17540 * When bntest is run from "make test" it drives bc to check its
17541 calculations, as well as internally checking them. If an internal check
17542 fails, it needs to cause bc to give a non-zero result or make test carries
17543 on without noticing the failure. Fixed.
17547 * DES library cleanups.
17551 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17552 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17553 ciphers. NOTE: although the key derivation function has been verified
17554 against some published test vectors it has not been extensively tested
17555 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17560 * Instead of "mkdir -p", which is not fully portable, use new
17561 Perl script "util/mkdir-p.pl".
17565 * Rewrite the way password based encryption (PBE) is handled. It used to
17566 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17567 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17568 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17569 the 'parameter' field of the AlgorithmIdentifier is passed to the
17570 underlying key generation function so it must do its own ASN1 parsing.
17571 This has also changed the EVP_PBE_CipherInit() function which now has a
17572 'parameter' argument instead of literal salt and iteration count values
17573 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17577 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17578 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17579 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17580 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17581 value was just used as a "magic string" and not used directly its
17582 value doesn't matter.
17586 * Introduce some semblance of const correctness to BN. Shame C doesn't
17591 * "linux-sparc64" configuration (ultrapenguin).
17593 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17594 "linux-sparc" configuration.
17596 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17598 * config now generates no-xxx options for missing ciphers.
17602 * Support the EBCDIC character set (work in progress).
17603 File ebcdic.c not yet included because it has a different license.
17605 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17607 * Support BS2000/OSD-POSIX.
17609 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17611 * Make callbacks for key generation use `void *` instead of `char *`.
17615 * Make S/MIME samples compile (not yet tested).
17619 * Additional typesafe stacks.
17623 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17627 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17629 * New configuration variant "sco5-gcc".
17631 * Updated some demos.
17633 *Sean O Riordain, Wade Scholine*
17635 * Add missing BIO_free at exit of pkcs12 application.
17639 * Fix memory leak in conf.c.
17643 * Updates for Win32 to assembler version of MD5.
17647 * Set #! path to perl in `apps/der_chop` to where we found it
17648 instead of using a fixed path.
17652 * SHA library changes for irix64-mips4-cc.
17656 * Improvements for VMS support.
17660 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17662 * Bignum library bug fix. IRIX 6 passes "make test" now!
17663 This also avoids the problems with SC4.2 and unpatched SC5.
17665 *Andy Polyakov <appro@fy.chalmers.se>*
17667 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17668 These are required because of the typesafe stack would otherwise break
17669 existing code. If old code used a structure member which used to be STACK
17670 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17671 sk_num or sk_value it would produce an error because the num, data members
17672 are not present in STACK_OF. Now it just produces a warning. sk_set
17673 replaces the old method of assigning a value to sk_value
17674 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17675 that does this will no longer work (and should use sk_set instead) but
17676 this could be regarded as a "questionable" behaviour anyway.
17680 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17681 correctly handle encrypted S/MIME data.
17685 * Change type of various DES function arguments from des_cblock
17686 (which means, in function argument declarations, pointer to char)
17687 to des_cblock * (meaning pointer to array with 8 char elements),
17688 which allows the compiler to do more typechecking; it was like
17689 that back in SSLeay, but with lots of ugly casts.
17691 Introduce new type const_des_cblock.
17695 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17696 problems: find RecipientInfo structure that matches recipient certificate
17697 and initialise the ASN1 structures properly based on passed cipher.
17701 * Belatedly make the BN tests actually check the results.
17705 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17706 to and from BNs: it was completely broken. New compilation option
17707 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17708 key elements as negative integers.
17712 * Reorganize and speed up MD5.
17714 *Andy Polyakov <appro@fy.chalmers.se>*
17718 *Richard Levitte <richard@levitte.org>*
17720 * New option -out to asn1parse to allow the parsed structure to be
17721 output to a file. This is most useful when combined with the -strparse
17722 option to examine the output of things like OCTET STRINGS.
17726 * Make SSL library a little more fool-proof by not requiring any longer
17727 that `SSL_set_{accept,connect}_state` be called before
17728 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17729 in many applications because usually everything *appeared* to work as
17730 intended anyway -- now it really works as intended).
17734 * Move openssl.cnf out of lib/.
17738 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17739 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17740 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17742 *Ralf S. Engelschall*
17744 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17745 handle PKCS#7 enveloped data properly.
17747 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17749 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17750 copying pointers. The cert_st handling is changed by this in
17751 various ways (and thus what used to be known as ctx->default_cert
17752 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17753 any longer when s->cert does not give us what we need).
17754 ssl_cert_instantiate becomes obsolete by this change.
17755 As soon as we've got the new code right (possibly it already is?),
17756 we have solved a couple of bugs of the earlier code where s->cert
17757 was used as if it could not have been shared with other SSL structures.
17759 Note that using the SSL API in certain dirty ways now will result
17760 in different behaviour than observed with earlier library versions:
17761 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17762 does not influence s as it used to.
17764 In order to clean up things more thoroughly, inside SSL_SESSION
17765 we don't use CERT any longer, but a new structure SESS_CERT
17766 that holds per-session data (if available); currently, this is
17767 the peer's certificate chain and, for clients, the server's certificate
17768 and temporary key. CERT holds only those values that can have
17769 meaningful defaults in an SSL_CTX.
17773 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17774 from the internal representation. Various PKCS#7 fixes: remove some
17775 evil casts and set the enc_dig_alg field properly based on the signing
17780 * Allow PKCS#12 password to be set from the command line or the
17781 environment. Let 'ca' get its config file name from the environment
17782 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17787 * Allow certificate policies extension to use an IA5STRING for the
17788 organization field. This is contrary to the PKIX definition but
17789 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17794 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17795 without disallowing inline assembler and the like for non-pedantic builds.
17799 * Support Borland C++ builder.
17801 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17807 * SHA-1 cleanups and performance enhancements.
17809 *Andy Polyakov <appro@fy.chalmers.se>*
17811 * Sparc v8plus assembler for the bignum library.
17813 *Andy Polyakov <appro@fy.chalmers.se>*
17815 * Accept any -xxx and +xxx compiler options in Configure.
17819 * Update HPUX configuration.
17823 * Add missing `sk_<type>_unshift()` function to safestack.h
17825 *Ralf S. Engelschall*
17827 * New function SSL_CTX_use_certificate_chain_file that sets the
17828 "extra_cert"s in addition to the certificate. (This makes sense
17829 only for "PEM" format files, as chains as a whole are not
17834 * Support verify_depth from the SSL API.
17835 x509_vfy.c had what can be considered an off-by-one-error:
17836 Its depth (which was not part of the external interface)
17837 was actually counting the number of certificates in a chain;
17838 now it really counts the depth.
17842 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17843 instead of X509err, which often resulted in confusing error
17844 messages since the error codes are not globally unique
17845 (e.g. an alleged error in ssl3_accept when a certificate
17846 didn't match the private key).
17848 * New function SSL_CTX_set_session_id_context that allows to set a default
17849 value (so that you don't need SSL_set_session_id_context for each
17850 connection using the SSL_CTX).
17854 * OAEP decoding bug fix.
17858 * Support INSTALL_PREFIX for package builders, as proposed by
17863 * New Configure options "threads" and "no-threads". For systems
17864 where the proper compiler options are known (currently Solaris
17865 and Linux), "threads" is the default.
17869 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17873 * Install various scripts to $(OPENSSLDIR)/misc, not to
17874 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17875 such as /usr/local/bin.
17879 * "make linux-shared" to build shared libraries.
17881 *Niels Poppe <niels@netbox.org>*
17883 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17887 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17888 extension adding in x509 utility.
17892 * Remove NOPROTO sections and error code comments.
17896 * Partial rewrite of the DEF file generator to now parse the ANSI
17901 * New Configure options --prefix=DIR and --openssldir=DIR.
17905 * Complete rewrite of the error code script(s). It is all now handled
17906 by one script at the top level which handles error code gathering,
17907 header rewriting and C source file generation. It should be much better
17908 than the old method: it now uses a modified version of Ulf's parser to
17909 read the ANSI prototypes in all header files (thus the old K&R definitions
17910 aren't needed for error creation any more) and do a better job of
17911 translating function codes into names. The old 'ASN1 error code embedded
17912 in a comment' is no longer necessary and it doesn't use .err files which
17913 have now been deleted. Also the error code call doesn't have to appear all
17914 on one line (which resulted in some large lines...).
17918 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17922 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17923 0 (which usually indicates a closed connection), but continue reading.
17927 * Fix some race conditions.
17931 * Add support for CRL distribution points extension. Add Certificate
17932 Policies and CRL distribution points documentation.
17936 * Move the autogenerated header file parts to crypto/opensslconf.h.
17940 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17941 8 of keying material. Merlin has also confirmed interop with this fix
17942 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17944 *Merlin Hughes <merlin@baltimore.ie>*
17946 * Fix lots of warnings.
17948 *Richard Levitte <levitte@stacken.kth.se>*
17950 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17951 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17953 *Richard Levitte <levitte@stacken.kth.se>*
17955 * Fix problems with sizeof(long) == 8.
17957 *Andy Polyakov <appro@fy.chalmers.se>*
17959 * Change functions to ANSI C.
17963 * Fix typos in error codes.
17965 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17967 * Remove defunct assembler files from Configure.
17971 * SPARC v8 assembler BIGNUM implementation.
17973 *Andy Polyakov <appro@fy.chalmers.se>*
17975 * Support for Certificate Policies extension: both print and set.
17976 Various additions to support the r2i method this uses.
17980 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17981 return a const string when you are expecting an allocated buffer.
17985 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17986 types DirectoryString and DisplayText.
17990 * Add code to allow r2i extensions to access the configuration database,
17991 add an LHASH database driver and add several ctx helper functions.
17995 * Fix an evil bug in bn_expand2() which caused various BN functions to
17996 fail when they extended the size of a BIGNUM.
18000 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18001 support typesafe stack.
18005 * Fix typo in SSL_[gs]et_options().
18007 *Nils Frostberg <nils@medcom.se>*
18009 * Delete various functions and files that belonged to the (now obsolete)
18010 old X509V3 handling code.
18014 * New Configure option "rsaref".
18018 * Don't auto-generate pem.h.
18022 * Introduce type-safe ASN.1 SETs.
18026 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18028 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18030 * Introduce type-safe STACKs. This will almost certainly break lots of code
18031 that links with OpenSSL (well at least cause lots of warnings), but fear
18032 not: the conversion is trivial, and it eliminates loads of evil casts. A
18033 few STACKed things have been converted already. Feel free to convert more.
18034 In the fullness of time, I'll do away with the STACK type altogether.
18038 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18039 specified in `<certfile>` by updating the entry in the index.txt file.
18040 This way one no longer has to edit the index.txt file manually for
18041 revoking a certificate. The -revoke option does the gory details now.
18043 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18045 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18046 `-text` option at all and this way the `-noout -text` combination was
18047 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18049 *Ralf S. Engelschall*
18051 * Make sure a corresponding plain text error message exists for the
18052 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18053 verify callback function determined that a certificate was revoked.
18055 *Ralf S. Engelschall*
18057 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18058 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18059 all available ciphers including rc5, which was forgotten until now.
18060 In order to let the testing shell script know which algorithms
18061 are available, a new (up to now undocumented) command
18062 `openssl list-cipher-commands` is used.
18066 * Bugfix: s_client occasionally would sleep in select() when
18067 it should have checked SSL_pending() first.
18071 * New functions DSA_do_sign and DSA_do_verify to provide access to
18072 the raw DSA values prior to ASN.1 encoding.
18076 * Tweaks to Configure
18078 *Niels Poppe <niels@netbox.org>*
18080 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18085 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18089 * New config option to avoid instructions that are illegal on the 80386.
18090 The default code is faster, but requires at least a 486.
18094 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18095 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18096 same as SSL2_VERSION anyway.
18100 * New "-showcerts" option for s_client.
18104 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18105 application. Various cleanups and fixes.
18109 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18110 modify error routines to work internally. Add error codes and PBE init
18111 to library startup routines.
18115 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18116 packing functions to asn1 and evp. Changed function names and error
18117 codes along the way.
18121 * PKCS12 integration: and so it begins... First of several patches to
18122 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18123 objects to objects.h
18127 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18128 and display support for Thawte strong extranet extension.
18132 * Add LinuxPPC support.
18134 *Jeff Dubrule <igor@pobox.org>*
18136 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18137 bn_div_words in alpha.s.
18139 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18141 * Make sure the RSA OAEP test is skipped under -DRSAref because
18142 OAEP isn't supported when OpenSSL is built with RSAref.
18144 *Ulf Moeller <ulf@fitug.de>*
18146 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18147 so they no longer are missing under -DNOPROTO.
18149 *Soren S. Jorvang <soren@t.dk>*
18151 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18153 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18154 doesn't work when the session is reused. Coming soon!
18158 * Fix a security hole, that allows sessions to be reused in the wrong
18159 context thus bypassing client cert protection! All software that uses
18160 client certs and session caches in multiple contexts NEEDS PATCHING to
18161 allow session reuse! A fuller solution is in the works.
18163 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18165 * Some more source tree cleanups (removed obsolete files
18166 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18167 permission on "config" script to be executable) and a fix for the INSTALL
18170 *Ulf Moeller <ulf@fitug.de>*
18172 * Remove some legacy and erroneous uses of malloc, free instead of
18175 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18177 * Make rsa_oaep_test return non-zero on error.
18179 *Ulf Moeller <ulf@fitug.de>*
18181 * Add support for native Solaris shared libraries. Configure
18182 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18183 if someone would make that last step automatic.
18185 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18187 * ctx_size was not built with the right compiler during "make links". Fixed.
18191 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18192 except NULL ciphers". This means the default cipher list will no longer
18193 enable NULL ciphers. They need to be specifically enabled e.g. with
18194 the string "DEFAULT:eNULL".
18198 * Fix to RSA private encryption routines: if p < q then it would
18199 occasionally produce an invalid result. This will only happen with
18200 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18204 * Be less restrictive and allow also `perl util/perlpath.pl
18205 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18206 because this way one can also use an interpreter named `perl5` (which is
18207 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18208 installed as `perl`).
18210 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18212 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18214 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18216 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18217 advapi32.lib to Win32 build and change the pem test comparison
18218 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18219 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18220 and crypto/des/ede_cbcm_enc.c.
18224 * DES quad checksum was broken on big-endian architectures. Fixed.
18228 * Comment out two functions in bio.h that aren't implemented. Fix up the
18229 Win32 test batch file so it (might) work again. The Win32 test batch file
18230 is horrible: I feel ill....
18234 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18235 in e_os.h. Audit of header files to check ANSI and non ANSI
18236 sections: 10 functions were absent from non ANSI section and not exported
18237 from Windows DLLs. Fixed up libeay.num for new functions.
18241 * Make `openssl version` output lines consistent.
18243 *Ralf S. Engelschall*
18245 * Fix Win32 symbol export lists for BIO functions: Added
18246 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18247 to ms/libeay{16,32}.def.
18249 *Ralf S. Engelschall*
18251 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18252 fine under Unix and passes some trivial tests I've now added. But the
18253 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18254 added to make sure no one expects that this stuff really works in the
18255 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18256 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18259 *Ralf S. Engelschall*
18261 * Fix the generation of two part addresses in perl.
18263 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18265 * Add config entry for Linux on MIPS.
18267 *John Tobey <jtobey@channel1.com>*
18269 * Make links whenever Configure is run, unless we are on Windoze.
18273 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18274 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18279 * Add a useful kludge to allow package maintainers to specify compiler and
18280 other platforms details on the command line without having to patch the
18281 Configure script every time: One now can use
18282 `perl Configure <id>:<details>`,
18283 i.e. platform ids are allowed to have details appended
18284 to them (separated by colons). This is treated as there would be a static
18285 pre-configured entry in Configure's %table under key `<id>` with value
18286 `<details>` and `perl Configure <id>` is called. So, when you want to
18287 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18288 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18289 now, which overrides the FreeBSD-elf entry on-the-fly.
18291 *Ralf S. Engelschall*
18293 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18297 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18298 on the `perl Configure ...` command line. This way one can compile
18299 OpenSSL libraries with Position Independent Code (PIC) which is needed
18300 for linking it into DSOs.
18302 *Ralf S. Engelschall*
18304 * Remarkably, export ciphers were totally broken and no-one had noticed!
18309 * Cleaned up the LICENSE document: The official contact for any license
18310 questions now is the OpenSSL core team under openssl-core@openssl.org.
18311 And add a paragraph about the dual-license situation to make sure people
18312 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18313 to the OpenSSL toolkit.
18315 *Ralf S. Engelschall*
18317 * General source tree makefile cleanups: Made `making xxx in yyy...`
18318 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18319 Additionally cleaned up the `make links` target: Remove unnecessary
18320 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18321 to speed processing and no longer clutter the display with confusing
18322 stuff. Instead only the actually done links are displayed.
18324 *Ralf S. Engelschall*
18326 * Permit null encryption ciphersuites, used for authentication only. It used
18327 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18328 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18333 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18334 signed attributes when verifying signatures (this would break them),
18335 the detached data encoding was wrong and public keys obtained using
18336 X509_get_pubkey() weren't freed.
18340 * Add text documentation for the BUFFER functions. Also added a work around
18341 to a Win95 console bug. This was triggered by the password read stuff: the
18342 last character typed gets carried over to the next fread(). If you were
18343 generating a new cert request using 'req' for example then the last
18344 character of the passphrase would be CR which would then enter the first
18349 * Added the new 'Includes OpenSSL Cryptography Software' button as
18350 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18351 button and can be used by applications based on OpenSSL to show the
18352 relationship to the OpenSSL project.
18354 *Ralf S. Engelschall*
18356 * Remove confusing variables in function signatures in files
18357 ssl/ssl_lib.c and ssl/ssl.h.
18359 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18361 * Don't install bss_file.c under PREFIX/include/
18363 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18365 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18366 functions that return function pointers and has support for NT specific
18367 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18368 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18369 unsigned to signed types: this was killing the Win32 compile.
18373 * Add new certificate file to stack functions,
18374 SSL_add_dir_cert_subjects_to_stack() and
18375 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18376 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18377 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18378 This means that Apache-SSL and similar packages don't have to mess around
18379 to add as many CAs as they want to the preferred list.
18383 * Experiment with doxygen documentation. Currently only partially applied to
18385 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18386 openssl.doxy as the configuration file.
18390 * Get rid of remaining C++-style comments which strict C compilers hate.
18392 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18394 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18395 compiled in by default: it has problems with large keys.
18399 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18400 DH private keys and/or callback functions which directly correspond to
18401 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18402 is needed for applications which have to configure certificates on a
18403 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18405 For the RSA certificate situation is makes no difference, but
18406 for the DSA certificate situation this fixes the "no shared cipher"
18407 problem where the OpenSSL cipher selection procedure failed because the
18408 temporary keys were not overtaken from the context and the API provided
18409 no way to reconfigure them.
18410 The new functions now let applications reconfigure the stuff and they
18411 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18412 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18413 non-public-API function ssl_cert_instantiate() is used as a helper
18414 function and also to reduce code redundancy inside ssl_rsa.c.
18416 *Ralf S. Engelschall*
18418 * Move s_server -dcert and -dkey options out of the undocumented feature
18419 area because they are useful for the DSA situation and should be
18420 recognized by the users.
18422 *Ralf S. Engelschall*
18424 * Fix the cipher decision scheme for export ciphers: the export bits are
18425 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18426 SSL_EXP_MASK. So, the original variable has to be used instead of the
18427 already masked variable.
18429 *Richard Levitte <levitte@stacken.kth.se>*
18431 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18433 *Richard Levitte <levitte@stacken.kth.se>*
18435 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18436 from `int` to `unsigned int` because it is a length and initialized by
18437 EVP_DigestFinal() which expects an `unsigned int *`.
18439 *Richard Levitte <levitte@stacken.kth.se>*
18441 * Don't hard-code path to Perl interpreter on shebang line of Configure
18442 script. Instead use the usual Shell->Perl transition trick.
18444 *Ralf S. Engelschall*
18446 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18447 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18448 -noout -modulus` as it's already the case for `openssl rsa -noout
18449 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18450 currently the public key is printed (a decision which was already done by
18451 `openssl dsa -modulus` in the past) which serves a similar purpose.
18452 Additionally the NO_RSA no longer completely removes the whole -modulus
18453 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18456 *Ralf S. Engelschall*
18458 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18459 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18461 *Arne Ansper <arne@ats.cyber.ee>*
18463 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18464 to be added. Now both 'req' and 'ca' can use new objects defined in the
18469 * Add cool BIO that does syslog (or event log on NT).
18471 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18473 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18474 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18475 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18476 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18480 * Add preliminary config info for new extension code.
18484 * Make RSA_NO_PADDING really use no padding.
18486 *Ulf Moeller <ulf@fitug.de>*
18488 * Generate errors when private/public key check is done.
18492 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18493 for some CRL extensions and new objects added.
18497 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18498 key usage extension and fuller support for authority key id.
18502 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18503 padding method for RSA, which is recommended for new applications in PKCS
18504 #1 v2.0 (RFC 2437, October 1998).
18505 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18506 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18507 against Bleichbacher's attack on RSA.
18508 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18511 * Updates to the new SSL compression code
18513 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18515 * Fix so that the version number in the master secret, when passed
18516 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18517 (because the server will not accept higher), that the version number
18518 is 0x03,0x01, not 0x03,0x00
18520 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18522 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18523 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18524 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18528 * Support for RAW extensions where an arbitrary extension can be
18529 created by including its DER encoding. See `apps/openssl.cnf` for
18534 * Make sure latest Perl versions don't interpret some generated C array
18535 code as Perl array code in the crypto/err/err_genc.pl script.
18537 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18539 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18540 not many people have the assembler. Various Win32 compilation fixes and
18541 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18542 build instructions.
18546 * Modify configure script 'Configure' to automatically create crypto/date.h
18547 file under Win32 and also build pem.h from pem.org. New script
18548 util/mkfiles.pl to create the MINFO file on environments that can't do a
18549 'make files': perl util/mkfiles.pl >MINFO should work.
18553 * Major rework of DES function declarations, in the pursuit of correctness
18554 and purity. As a result, many evil casts evaporated, and some weirdness,
18555 too. You may find this causes warnings in your code. Zapping your evil
18556 casts will probably fix them. Mostly.
18560 * Fix for a typo in asn1.h. Bug fix to object creation script
18561 obj_dat.pl. It considered a zero in an object definition to mean
18562 "end of object": none of the objects in objects.h have any zeros
18563 so it wasn't spotted.
18565 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18567 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18568 Masking (CBCM). In the absence of test vectors, the best I have been able
18569 to do is check that the decrypt undoes the encrypt, so far. Send me test
18570 vectors if you have them.
18574 * Correct calculation of key length for export ciphers (too much space was
18575 allocated for null ciphers). This has not been tested!
18579 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18580 message is now correct (it understands "crypto" and "ssl" on its
18581 command line). There is also now an "update" option. This will update
18582 the util/ssleay.num and util/libeay.num files with any new functions.
18584 perl util/mkdef.pl crypto ssl update
18585 it will update them.
18589 * Overhauled the Perl interface:
18590 - ported BN stuff to OpenSSL's different BN library
18591 - made the perl/ source tree CVS-aware
18592 - renamed the package from SSLeay to OpenSSL (the files still contain
18593 their history because I've copied them in the repository)
18594 - removed obsolete files (the test scripts will be replaced
18595 by better Test::Harness variants in the future)
18597 *Ralf S. Engelschall*
18599 * First cut for a very conservative source tree cleanup:
18600 1. merge various obsolete readme texts into doc/ssleay.txt
18601 where we collect the old documents and readme texts.
18602 2. remove the first part of files where I'm already sure that we no
18603 longer need them because of three reasons: either they are just temporary
18604 files which were left by Eric or they are preserved original files where
18605 I've verified that the diff is also available in the CVS via "cvs diff
18606 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18607 the crypto/md/ stuff).
18609 *Ralf S. Engelschall*
18611 * More extension code. Incomplete support for subject and issuer alt
18612 name, issuer and authority key id. Change the i2v function parameters
18613 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18614 what that's for :-) Fix to ASN1 macro which messed up
18615 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18619 * Preliminary support for ENUMERATED type. This is largely copied from the
18624 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18626 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18628 * Make sure `make rehash` target really finds the `openssl` program.
18630 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18632 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18633 like to hear about it if this slows down other processors.
18637 * Add CygWin32 platform information to Configure script.
18639 *Alan Batie <batie@aahz.jf.intel.com>*
18641 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18643 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18645 * New program nseq to manipulate netscape certificate sequences
18649 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18654 * Fixes to BN code. Previously the default was to define BN_RECURSION
18655 but the BN code had some problems that would cause failures when
18656 doing certificate verification and some other functions.
18658 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18660 * Add ASN1 and PEM code to support netscape certificate sequences.
18664 * Add ASN1 and PEM code to support netscape certificate sequences.
18668 * Add several PKIX and private extended key usage OIDs.
18672 * Modify the 'ca' program to handle the new extension code. Modify
18673 openssl.cnf for new extension format, add comments.
18677 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18678 and add a sample to openssl.cnf so req -x509 now adds appropriate
18683 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18684 error code, add initial support to X509_print() and x509 application.
18688 * Takes a deep breath and start adding X509 V3 extension support code. Add
18689 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18690 stuff is currently isolated and isn't even compiled yet.
18694 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18695 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18696 Removed the versions check from X509 routines when loading extensions:
18697 this allows certain broken certificates that don't set the version
18698 properly to be processed.
18702 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18703 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18704 can still be regenerated with "make depend".
18708 * Spelling mistake in C version of CAST-128.
18710 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18712 * Changes to the error generation code. The perl script err-code.pl
18713 now reads in the old error codes and retains the old numbers, only
18714 adding new ones if necessary. It also only changes the .err files if new
18715 codes are added. The makefiles have been modified to only insert errors
18716 when needed (to avoid needlessly modifying header files). This is done
18717 by only inserting errors if the .err file is newer than the auto generated
18718 C file. To rebuild all the error codes from scratch (the old behaviour)
18719 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18720 or delete all the .err files.
18724 * CAST-128 was incorrectly implemented for short keys. The C version has
18725 been fixed, but is untested. The assembler versions are also fixed, but
18726 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18727 to regenerate it if needed.
18728 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18729 Hagino <itojun@kame.net>*
18731 * File was opened incorrectly in randfile.c.
18733 *Ulf Möller <ulf@fitug.de>*
18735 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18736 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18737 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18738 al: it's just almost always a UTCTime. Note this patch adds new error
18739 codes so do a "make errors" if there are problems.
18743 * Correct Linux 1 recognition in config.
18745 *Ulf Möller <ulf@fitug.de>*
18747 * Remove pointless MD5 hash when using DSA keys in ca.
18749 *Anonymous <nobody@replay.com>*
18751 * Generate an error if given an empty string as a cert directory. Also
18752 generate an error if handed NULL (previously returned 0 to indicate an
18753 error, but didn't set one).
18755 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18757 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18761 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18762 parameters. This was causing a warning which killed off the Win32 compile.
18766 * Remove C++ style comments from crypto/bn/bn_local.h.
18768 *Neil Costigan <neil.costigan@celocom.com>*
18770 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18771 based on a text string, looking up short and long names and finally
18772 "dot" format. The "dot" format stuff didn't work. Added new function
18773 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18774 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18775 OID is not part of the table.
18779 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18780 X509_LOOKUP_by_alias().
18784 * Sort openssl functions by name.
18788 * Get the `gendsa` command working and add it to the `list` command. Remove
18789 encryption from sample DSA keys (in case anyone is interested the password
18794 * Make *all* `*_free` functions accept a NULL pointer.
18796 *Frans Heymans <fheymans@isaserver.be>*
18798 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18801 *Anonymous <nobody@replay.com>*
18803 * s_server should send the CAfile as acceptable CAs, not its own cert.
18805 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18807 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18809 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18811 * Temp key "for export" tests were wrong in s3_srvr.c.
18813 *Anonymous <nobody@replay.com>*
18815 * Add prototype for temp key callback functions
18816 SSL_CTX_set_tmp_{rsa,dh}_callback().
18820 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18821 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18825 * X509_name_add_entry() freed the wrong thing after an error.
18827 *Arne Ansper <arne@ats.cyber.ee>*
18829 * rsa_eay.c would attempt to free a NULL context.
18831 *Arne Ansper <arne@ats.cyber.ee>*
18833 * BIO_s_socket() had a broken should_retry() on Windoze.
18835 *Arne Ansper <arne@ats.cyber.ee>*
18837 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18839 *Arne Ansper <arne@ats.cyber.ee>*
18841 * Make sure the already existing X509_STORE->depth variable is initialized
18842 in X509_STORE_new(), but document the fact that this variable is still
18843 unused in the certificate verification process.
18845 *Ralf S. Engelschall*
18847 * Fix the various library and `apps/` files to free up pkeys obtained from
18848 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18852 * Fix reference counting in X509_PUBKEY_get(). This makes
18853 demos/maurice/example2.c work, amongst others, probably.
18855 *Steve Henson and Ben Laurie*
18857 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18858 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18859 are no longer created. This way we have a single and consistent command
18860 line interface `openssl <command>`, similar to `cvs <command>`.
18862 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18864 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18865 BIT STRING wrapper always have zero unused bits.
18869 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18873 * Make the top-level INSTALL documentation easier to understand.
18877 * Makefiles updated to exit if an error occurs in a sub-directory
18878 make (including if user presses ^C) [Paul Sutton]
18880 * Make Montgomery context stuff explicit in RSA data structure.
18884 * Fix build order of pem and err to allow for generated pem.h.
18888 * Fix renumbering bug in X509_NAME_delete_entry().
18892 * Enhanced the err-ins.pl script so it makes the error library number
18893 global and can add a library name. This is needed for external ASN1 and
18894 other error libraries.
18898 * Fixed sk_insert which never worked properly.
18902 * Fix ASN1 macros so they can handle indefinite length constructed
18903 EXPLICIT tags. Some non standard certificates use these: they can now
18908 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18909 into a single doc/ssleay.txt bundle. This way the information is still
18910 preserved but no longer messes up this directory. Now it's new room for
18911 the new set of documentation files.
18913 *Ralf S. Engelschall*
18915 * SETs were incorrectly DER encoded. This was a major pain, because they
18916 shared code with SEQUENCEs, which aren't coded the same. This means that
18917 almost everything to do with SETs or SEQUENCEs has either changed name or
18918 number of arguments.
18920 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18922 * Fix test data to work with the above.
18926 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18927 was already fixed by Eric for 0.9.1 it seems.
18929 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18931 * Autodetect FreeBSD3.
18935 * Fix various bugs in Configure. This affects the following platforms:
18939 unixware-2.0-pentium
18944 * Eliminate generated files from CVS. Reorder tests to regenerate files
18945 before they are needed.
18949 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18953 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18955 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18956 changed SSLeay to OpenSSL in version strings.
18958 *Ralf S. Engelschall*
18960 * Some fixups to the top-level documents.
18964 * Fixed the nasty bug where rsaref.h was not found under compile-time
18965 because the symlink to include/ was missing.
18967 *Ralf S. Engelschall*
18969 * Incorporated the popular no-RSA/DSA-only patches
18970 which allow to compile a RSA-free SSLeay.
18972 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18974 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18975 when "ssleay" is still not found.
18977 *Ralf S. Engelschall*
18979 * Added more platforms to Configure: Cray T3E, HPUX 11,
18981 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18983 * Updated the README file.
18985 *Ralf S. Engelschall*
18987 * Added various .cvsignore files in the CVS repository subdirs
18988 to make a "cvs update" really silent.
18990 *Ralf S. Engelschall*
18992 * Recompiled the error-definition header files and added
18993 missing symbols to the Win32 linker tables.
18995 *Ralf S. Engelschall*
18997 * Cleaned up the top-level documents;
18998 o new files: CHANGES and LICENSE
18999 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19000 o merged COPYRIGHT into LICENSE
19001 o removed obsolete TODO file
19002 o renamed MICROSOFT to INSTALL.W32
19004 *Ralf S. Engelschall*
19006 * Removed dummy files from the 0.9.1b source tree:
19008 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19009 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19010 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19011 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19012 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19015 *Ralf S. Engelschall*
19017 * Added various platform portability fixes.
19021 * The Genesis of the OpenSSL rpject:
19022 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19023 Young and Tim J. Hudson created while they were working for C2Net until
19026 *The OpenSSL Project*
19028 ### Changes between 0.9.0b and 0.9.1b [not released]
19030 * Updated a few CA certificates under certs/
19034 * Changed some BIGNUM api stuff.
19038 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19039 DGUX x86, Linux Alpha, etc.
19043 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19044 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19049 * Add -strparse option to asn1pars program which parses nested
19052 *Dr Stephen Henson <shenson@bigfoot.com>*
19054 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19058 * DSA fix for "ca" program.
19062 * Added "-genkey" option to "dsaparam" program.
19066 * Added RIPE MD160 (rmd160) message digest.
19070 * Added -a (all) option to "ssleay version" command.
19074 * Added PLATFORM define which is the id given to Configure.
19078 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19082 * Extended the ASN.1 parser routines.
19086 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19090 * Added a BN_CTX to the BN library.
19094 * Fixed the weak key values in DES library
19098 * Changed API in EVP library for cipher aliases.
19102 * Added support for RC2/64bit cipher.
19106 * Converted the lhash library to the crypto/mem.c functions.
19110 * Added more recognized ASN.1 object ids.
19114 * Added more RSA padding checks for SSL/TLS.
19118 * Added BIO proxy/filter functionality.
19122 * Added extra_certs to SSL_CTX which can be used
19123 send extra CA certificates to the client in the CA cert chain sending
19124 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19128 * Now Fortezza is denied in the authentication phase because
19129 this is key exchange mechanism is not supported by SSLeay at all.
19133 * Additional PKCS1 checks.
19137 * Support the string "TLSv1" for all TLS v1 ciphers.
19141 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19142 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19146 * Fixed a few memory leaks.
19150 * Fixed various code and comment typos.
19154 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19155 bytes sent in the client random.
19157 *Edward Bishop <ebishop@spyglass.com>*
19161 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19162 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19163 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19164 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19165 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19166 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19167 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19168 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19169 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19170 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19171 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19172 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19173 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19174 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19175 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19176 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19177 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19178 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19179 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19180 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19181 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19182 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19183 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19184 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19185 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19186 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19187 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19188 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19189 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19190 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19191 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19192 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19193 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19194 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19195 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19196 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19197 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19198 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19199 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19200 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19201 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19202 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19203 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19204 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19205 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19206 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19207 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19208 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19209 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19210 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19211 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19212 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19213 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19214 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19215 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19216 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19217 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19218 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19219 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19220 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19221 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19222 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19223 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19224 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19225 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19226 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19227 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19228 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19229 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19230 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19231 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19232 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19233 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19234 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19235 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19236 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19237 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19238 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19239 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19240 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19241 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19242 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19243 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19244 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19245 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19246 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19247 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19248 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19249 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19250 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19251 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19252 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19253 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19254 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19255 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19256 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19257 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19258 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19259 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19260 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19261 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19262 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19263 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19264 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19265 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19266 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19267 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19268 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19269 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19270 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19271 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19272 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19273 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19274 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19275 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19276 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19277 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19278 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19279 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19280 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19281 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19282 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19283 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19284 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19285 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19286 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19287 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19288 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19289 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19290 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19291 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19292 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19293 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19294 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19295 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19296 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19297 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19298 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19299 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19300 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19301 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19302 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19303 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19304 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19305 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19306 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19307 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19308 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19309 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19310 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19311 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19312 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19313 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19314 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19315 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19316 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19317 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19318 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19319 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19320 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19321 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19322 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655